Dun & Bradstreet, Inc. appeals from a judgment entered on a $3,847,000 general jury verdict against it in a business defamation action. It also appeals from the denial of post-trial motions for judgment n.o.v. and for a new trial.
The affiliated plaintiff corporations, Sun-ward Corporation, Wedg-Cor, Inc., and Marvel Brute Steel Buildings, Inc. (collectively “Sunward”) sued Dun & Bradstreet over the issuance of at least 340 inaccurate reports sent, on request, to various subscribers over a period of approximately two years, beginning in the fall of 1979 and ending in October, 1981, when Dun & Bradstreet was advised of the error and notified subscribers. The reports stated nothing explicitly defamatory about Sunward on their face, but they grossly understated the size of Sunward’s business. Sunward charged that if recipients of the reports had prior information as to Sunward’s true size, they would have interpreted the credit reports to mean that Sunward’s business was in a steep decline, thus imputing to Sunward financial distress and incompetence in the conduct of its business. Sunward further charged that a report of small size in and of itself is defamatory in the competitive business in which it is engaged, No recipients of the reports were produced to testify that they interpreted the reports to have the defamatory meanings charged by Sunward. Federal jurisdiction is based on diversity of citizenship. Colorado law governs.
On appeal Dun & Bradstreet contends: that there was no proof of defamation sufficient to allow the case to go to the jury; that the jury instruction defining “reckless disregard” (which Sunward had to prove to overcome Dun & Bradstreet’s qualified privilege) was erroneous, and evidence on the subject was insufficient; that the district court erred in permitting Sunward to recover presumed damages; and that the evidence on damage was improper. We agree with certain of those arguments, as discussed in this opinion. Therefore, we affirm in part, reverse in part, and remand for a new trial.
Sunward has filed a cross-appeal from the district court’s denial of pre-judgment interest on the jury award. Our decision to reverse and remand this case for a new trial renders that appeal moot.
BACKGROUND
The business of Dun & Bradstreet is well known and has been discussed extensively in decided cases, including the published opinion of the district court in this case. Sunward Corp. v. Dun & Bradstreet, Inc., 568 F.Supp. 602 (D.Colo.1983). In general, Dun & Bradstreet is a credit reporting agency which “provides subscribers with financial and related information about businesses. All the information is confidential; under the terms of the subscription agreement the subscribers may not reveal it to anyone else.” Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 105 S.Ct. 2939, 2941, 86 L.Ed.2d 593 (1985).
Sunward is in the business of manufacturing and selling steel buildings. R. Vol. IV at 73-74. Although it sells nationwide, Sunward’s buildings are sold primarily to agricultural and commercial concerns in the north-central states as barns, warehouses and storage bins. R. Vol. IV at 37, 73, 78.
In 1975 and 1976 Dun & Bradstreet issued basically accurate reports on segments of the Sunward corporate structure, stating, on December 22, 1975, that WedgCor Inc., (a subsidiary of Marvel Steel, Inc.) had annual “sales in the $7 million range.” Ex. 26, R. Vol. IV at 101. Because of Dun & Bradstreet’s records destruction policy it is not known what, if any, reports relating to Sunward were issued between 1976 and 1979. R. Vol. V at 196-97.
*514Over a two-year period, beginning with two reports dated August 15, 1979 (the first recorded issuance of which was on November 12, 1979), and ending in mid-October, 1981, Dun & Bradstreet, in response to requests from subscribers, issued approximately 340 reports on corporations in the Sunward group. Those reports grossly understated the size of Sunward’s business. The 1979 report on Sunward Corporation estimated annual sales at $200,000 with three to five employees, and stated that Sunward had $25,000-$50,000 in fixtures and equipment. Sunward was described as a real estate investment company. Ex. 29, R. Vol. VI at 262-64. The 1979 report on “Marvel Brute Steel Buildings/Marvel Steel Wedg-Cor/Sunward Purchasing Div.” stated that the company had three to five employees, with estimated annual sales of $500,000~$750,000.
In fact, in the 1979 fiscal year which ended February 28, 1980, the Sunward group of corporations had combined gross sales amounting to $29.5 million. Also in 1979, those corporations employed just over 300 salaried personnel, of whom 169 were employed by Marvel Brute Steel and 100 by Sunward. They also had approximately 250 commissioned sales personnel and 100 dealers. Approximately 80% of the total 1979 business was done by the commissioned sales personnel. R. Vol. IV at 80-81.1 In March, 1979, the beginning .of fiscal year inventory amounted to $4.6 million, and plant and equipment, after depreciation, was valued at $5.8 million. The backlog of orders stood at $6.7 million. Ex. 64, R. Vol. IV at 83-86.
Although some differences appeared, the Dun & Bradstreet reports on the Sunward group of corporations continued in the same vein as the 1979 reports until October, 1981 when Danton Wirth, owner and a principal in these corporations, saw one of the reports. Exs. 23, 24, R. Vol. IV at 93-95. Each of the credit reports stated on its face that Sunward had declined to be interviewed and declined all information; therefore, sales figures were the reporter’s estimates. Mr. Wirth testified that from and after 1975 he consistently refused to provide financial information about his companies to Dun & Bradstreet, and instructed his employees not to cooperate. R. Vol. IV at 91-92; R. Vol. V at 176. Requests by the Dun & Bradstreet representative for information in 1979 were rebuffed. R. Vol. VI at 265. The parties are in disagreement as to whether or not Sun-ward was aware of the erroneous credit reports between 1979 and October of 1981, but Mr. Wirth denied any knowledge. R. Vol. V at 174-76.
In the period generally corresponding with that in which the erroneous credit reports were being issued, Sunward’s gross sales declined 56.5% (PLEx. 66, par. 4); its sales force declined to just six active commissioned salespeople;2 and its backlog of orders decreased during the 1979 fiscal year from $6.7 million to $2.3 million. Ex. 64, R. Vol. IV at 83-84, 86-87.
The causal relationship of these and associated business difficulties of Sunward to the Dun & Bradstreet reports is a highly disputed and central issue in this case. Not one of the recipients of the erroneous credit reports was called to testify, nor was any witness called who had heard any reference to the reports. Rather, through three witnesses, Sunward established only *515that there were rumors in the industry that the company was in financial difficulty.
Steven L. Kalicki, a former employee, testified that while employed by Sunward in 1979 he heard rumors within the company, from company employees, that the business was not doing well. R. Vol. IV at 59-60, 63. And, in late 1980 and in 1981, after he had left the company, various contractors in Colorado commented to him that Sunward was doing poorly and about to go out of business. R. Vol. IV at 60-61. He failed to name any of the individuals who made such comments or any other parties to the conversation. Nor did he identify the time or place of the conversations.
Jerry Bishop, who worked for a competitor, testified that he had heard similar rumors on two occasions: in November, 1980, during the cocktail hour at the 25th anniversary meeting of the Metal Building Manufacturers’ Association (he could not remember who made the remark); and in a telephone conversation with a district manager for a competitor, Chief Industries, sometime between December 15, 1982 and January, 1983 (more than a year after Dun & Bradstreet ceased distribution of the erroneous reports). R. Vol. VI at 291-94.
Mr. Wirth testified to the existence of rumors and consequent difficulties with his sales force and dealers. He also admitted that he could not identify any customer or specific business loss directly traceable to the Dun & Bradstreet reports. R. Vol. IV at 103; R. Vol. V at 181. And, he acknowledged the onset of a bad economic climate in the early 1980s, resulting in a general decline in the steel building business. R. Vol. IV at 135, 140-41.
A Sunward employee, Jerry Johnston, testified in general that warranties and engineering certifications are important to the sale of buildings. R. Vol. IV at 45-46. But Sunward made no serious attempt to establish the resulting inference that small companies are not patronized because they are unable to provide or stand behind their warranties or certifications. That very general and sketchy testimony, along with figures showing its decline in business, constituted Sunward's evidence that the credit reports were considered to be defamatory.
The credit reports in question were prepared by Dun & Bradstreet employee Larry Thompson. Mr. Thompson admitted that after Sunward refused to supply any information he prepared the reports using estimates, assumptions and “guesstimates.” R. Vol. VI at 263-66. He used standard Dun & Bradstreet techniques to arrive at some estimates, which were apparently based on erroneous figures from some prior report. Id. However, Mr. Thompson failed to follow prescribed procedures in a number of important respects. He did not check public filings with state agencies. He did not check on Sunward offices and plants in other locations for which he had and reported the address— particularly the large plant in Jamestown, North Dakota. R. Vol. VI at 265-66. He did not even take into account in his report of three to five employees the obvious disproportion of that estimate to the size of the Sunward physical facilities which he visited in an attempt to obtain information in the first instance. R. Vol. VI at 265, 279. The building contained 10,000 square feet of office space according to Mr. Thompson’s own report. R. Vol. VI at 279. He was encouraged in his methods by Dun & Bradstreet’s compensation system, which gives bonus points for sales information and certain other data. R. Vol. VI at 267-69, 271. He was also under time pressure to produce a volume of reports acceptable to his superiors. R. Vol. VI at 274-78.
Dun & Bradstreet defends its procedures, but did not seriously contend at trial that Mr. Thompson's preparation of the reports on Sunward was proper. In his closing argument, counsel for Dun & Bradstreet told the jury:
I’m not, you know, not going to pin the rose on anybody. I wish he hadn’t done it. I’m proud of the fact that Dun & Bradstreet records document step by step by step everything we did. We honestly wrote down everything everybody said and I honestly stand before you and I say to you, I wish that he hadn’t taken *516a sensible procedure and taken it a step too far.
As you heard, we've taken steps to make sure that never again will anybody make that kind of mistake because in these circumstances we’ve stopped it, so it’s not a situation where you’re going to say, Mr. Bartlit, fine and good, what’s going to happen in the future. That kind of estimate is dead and gone in Dun & Bradstreet, as I assume you would expect.
R. Vol. XI at 691.
In its complaint and early in its suit Sunward asserted that the erroneous credit reports impaired its credit. It dropped that assertion at trial. Damage evidence consisted mainly of the decline in sales and projections of estimated profits. The evidence was presented primarily through a report and testimony by Richard D. Clark, an accountant and former part-time employee of Sunward. His testimony, based on mathematical comparisons, assumptions and projections, suggested that Sunward’s damages were between $7,637,119 and $11,294,524. The jury returned a general damage verdict of $3,847,000. Post-trial motions, respectively, for a remittitur and for pre-judgment interest were denied, along with Dun & Bradstreet’s motions for judgment n.o.v. and for a new trial.
The usual standards of review with respect to jury trials and awards apply in this case. See Dixson v. Newsweek, Inc., 562 F.2d 626, 631 (10th Cir.1977). With respect to constitutional issues raised by Dun & Bradstreet, we are not called upon to review fact findings by the jury, see Bose Corp. v. Consumers Union of the United States, Inc., 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984), so much as we are required to review the legal (including constitutional) propriety of standards given by the court in its instructions to the jury. To the extent it has been necessary to review the facts in order to determine if constitutional protections apply because matters of public interest are involved, no determination by the jury is implicated. See Green-moss, 472 U.S. 749, 105 S.Ct. 2939, 86 L.Ed.2d 593 (1985).
Prior to trial, the court granted partial summary judgment to Dun & Bradstreet on the ground that the Dun & Bradstreet reports in question were not libelous per se under Colorado law, and that Dun & Bradstreet, as a credit reporting agency, enjoyed a qualified privilege or conditional immunity for its reports. 568 F.Supp. at 605-608. The court also ruled that the nature of the publication entitled Sunward to a presumption of damages and that, therefore, Sunward was not required to plead and prove special damages. Id. at 606. At trial Sunward’s claim for punitive damages was dismissed. R. Yol. X, at 630-32. The other rulings were the foundation for the court’s instructions to the jury. Of those rulings, only those pertaining to presumed damages have been appealed, based upon the jury instruction on that subject. The other rulings continue undisturbed and are declared to be the law of this case.
I.
PROOF OF DEFAMATION
Dun & Bradstreet contends that this case should not have gone to the jury because Sunward failed to prove that anyone considered the credit reports to be defamatory. Sunward counters by arguing that it is entitled to rely solely on inferences drawn from its circumstantial evidence of a decline in sales, and from rumors that it was in financial difficulty.
We consider proof of defamation to be an important issue in this case since there is no case of recent vintage where a credit reporting agency has been charged with defamation because it reported a business to be smaller than it really was. We must assume that underestimating the size of a business is not unusual in credit reports, and most of the arguments raised in this case could be raised in other instances.
Sunward does not contend that the credit reports in question are defamatory on their face (libel per se). Rather, Sun-*517ward’s case proceeded on the theory that the reports became defamatory through inferences drawn from extrinsic circumstances (libel per quod).3
Quoting controlling Colorado precedent, the district court aptly characterized Sun-ward’s action as follows:
For language to be libelous per se, the facts or reports must be deplorable, derogatory, or disgraceful, and the element of disgraceful imputation must be clearly expressed on the face of the writing. Bernstein v. Dun & Bradstreet, 149 Colo. 150, 368 P.2d 780, 784 (1962). In evaluating a statement or article alleged to be libelous per se,
the court must interpret alone, without aid of inducements, colloquialisms, innuendos, and explanatory circumstances. To be libelous per se, the publication must contain defamatory words specifically directed at the person claiming injury, which words must, on their face, and without the aid of intrinsic proof be unmistakably recognized as injurious. (Citations omitted.)
Inter-State Detective Bureau v. The Denver Post, 29 Colo.App. 313, 484 P.2d 131, 133 (1971). Words which require an innuendo are not libelous per se. In determining whether words are libelous, they are to be given their ordinary and popular meaning. Knapp v. Post Printing and Publishing Co., 111 Colo. 492, 144 P.2d 981 (1943).
I agree with D & B that these inaccuracies and misstatements are not libelous per se. The language in the reports is neutral in content, neither scurrilous nor inflammatory. Only by going beyond the face of the report would the defamation become clear. As such, I find the issue is whether the report is libel per quod.
568 F.Supp. at 605-06.
The essential extrinsic fact in this case is that in relative terms Sunward was large when Dun & Bradstreet reported it to be very small. The record contains no evidence showing the extent of awareness in the marketplace of the size of Sunward’s business, but with sales exceeding $29 million in the 1979 fiscal year its visibility and presence could not have been insignificant. Sunward did establish that in 1975 Dun & Bradstreet had prepared a report on one of the Sunward corporations, as the corporate structure existed at that time. The report correctly reported sales at the seven million dollar level, with forty-five employees. Ex. 26, R. Vol. IV at 99-101. However, because Dun & Bradstreet regularly purges its files it is impossible to know how many subscribers had requested a copy of the 1975 report. Only one recipient was identified. R. Vol. V at 196-97. The fact of relative size difference was augmented by extrinsic evidence concerning the nature, scope and competitiveness of the steel building industry.
The additional external facts relied upon by Sunward are that its sales declined dramatically, and that disparaging rumors existed about its financial condition during the time that the inaccurate Dun & Bradstreet reports were being issued. Sun-ward’s sales force also declined, although the precise argument is that a planned phaseout of the sales force due to a conversion from commissioned sales personnel to sales through dealers was accelerated during the time the reports were extant.
In its complaint Sunward did not identify those facts or, more importantly, the alleged defamatory inferences with any spec*518ificity.4 At trial Sunward distilled its position to a general argument that the Dun & Bradstreet reports, in light of the extrinsic facts, must have been interpreted to mean that Sunward was in financial distress and a state of business failure. In his opening statement, Sunward’s counsel stated that “rumors started that this company ... was going out of business, it was broke, it lacked substance, it was a paper company.” R. Yol. IV at 10. “The reports indicated that, for a steel building manufacturer, Sunward was in deep financial trouble and without the resources to stand behind its product. Rumors arose in the field during the time these reports were distributed that plaintiffs were in financial straits.” Brief of Appellees at 19-20.
There was a second defamatory innuendo, not extensively developed, that a small operator in the metal building industry would not do well because of its size alone. “You don’t even have to have the two [1975 and 1979 reports] side by side to have concerns that an outfit that purports to be a whizbang manufacturer of steel buildings can’t amount to very much if that’s all that it has going for it.” R. Vol. XI at 653. “[T]he reports were defamatory. The reports grossly underestimated plaintiffs’ size and financial situation. The evidence revealed the importance in the competitive steel building industry of adequate financial capacity — ” Brief of Appellees at 16.
As indicated Sunward further implied that it was perceived as unable to stand behind its warranties and engineer's certifications. “Sunward was in deep financial trouble and without the resources to stand behind its product.” Brief of Appellees at 19-20; R. Vol. XI at 653.
Sunward relied entirely upon circumstantial evidence and inferences drawn from that evidence to prove the defamatory imputations just described. It offered no evidence directly linking the Dun & Bradstreet reports to the alleged defamatory meanings. No recipient or other person exposed to one of the reports testified.5 Mr. Wirth testified that he had no knowledge of any customer or sale which was affected by the reports. R. Vol. IV at 103; R. Vol. V at 181. Sunward explains its position as follows:
In this ease, D & B distributed reports to, among others, Sunward’s competitors. The reports indicated that, for a steel building manufacturer, Sunward was in deep financial trouble and without the resources required to stand behind its product. Rumors arose in the field during the time these reports were distributed that plaintiffs were in financial straits. It is a reasonable inference, given the competitive nature of the industry, that competitors of Sunward, which had received the D & B reports, were responsible for these rumors. Circumstantial evidence supported Sun-ward’s theory. No more was required.
Brief of Appellees at 19-20 (emphasis added).
The plaintiffs’ burden of proof in a defamation case is set out in § 613 of the Restatement (Second) of Torts (1977)6 as follows:
(1) In an action for defamation the plaintiff has the burden of proving, when the issue is properly raised,
(a) the defamatory character of the communication,
(b) its publication by the defendant,
(c) its application to the plaintiff,
(d) the recipient’s understanding of its defamatory meaning,
(e) the recipient’s understanding of it as intended to be applied to the plaintiff,
*519(f) special harm resulting to the plaintiff from its publication,
(g) the defendant’s negligence, reckless disregard or knowledge regarding the truth or falsity and the defamatory character of the communication, and
(h) the abuse of a conditional privilege. ******
(emphasis added). Comment c to § 613 further explains the requirement of proving the recipient’s understanding as follows: "Thus, when the defamatory character of the communication depends upon extrinsic circumstances, the plaintiff must prove both their existence and knowledge of them by the recipient of the communication ” (emphasis added). In § 563 of the Restatement the American Law Institute states that “[t]he question to be determined is whether the communication is reasonably understood in a defamatory sense by the recipient____ It is not enough that the language used is reasonably capable of a defamatory interpretation if the recipient did not in fact so understand it.” Restatement § 563 comment c (emphasis added).
The district court in Colorado and the Colorado Supreme Court have recognized this requirement. Rocky Mountain News Printing Co. v. Fridborn, 46 Colo. 440, 104 P. 956, 959 (1909) (“Or they [the defamatory words] may not be defamatory on their face, in which case the action cannot be maintained, unless the plaintiff can and does show that they were, under the particular circumstances, fairly capable of a special meaning rendering them defamatory, and that they were so understood.”) (emphasis added); see also Walters v. Linhof, 559 F.Supp. 1231, 1236 (D.Colo.1983); Knapp v. Post Printing & Publishing Co., 111 Colo. 492, 144 P.2d 981, 984 (1943) (in order for the defendant’s words to be defamatory, their defamatory meaning must be understood under the circumstances).
This court has also referred to this requirement: “No one seems to dispute the well established law, both generally and in Oklahoma, that in libel and slander actions [if a publication is libelous per quod] [t]he jury must then determine whether, under all of the existing circumstances, the recipients of the writing attributed to the language used the defamatory imputations set forth in the complaint____” M.F. Patterson Dental Supply Co. v. Wadley, 401 F.2d 167, 169 (10th Cir.1968).7 See generally Electric Furnace Corp. v. Deering Milliken Research Corp., 383 F.2d 352, 354 (6th Cir.1967) (“The fatal deficiency in plaintiff’s case resides in its failure to show that any recipient of the accused letter gave it the meaning which Electric asserts is fairly inferable from a reading of it”), cert. denied, 390 U.S. 949, 88 S.Ct. 1040, 19 L.Ed.2d 1141 (1968); Rudin v. Dow Jones & Co., 557 F.Supp. 535, 543-45 (S.D.N.Y. 1983); Reichman v. Bureau of Affirmative Action, 536 F.Supp. 1149, 1180 (M.D. Pa.1982); Keddie v. Pennsylvania State University, 412 F.Supp. 1264 (M.D.Pa. 1976); Altoona Clay Products, Inc. v. Dun & Bradstreet, Inc., 286 F.Supp. 899 (W.D.Pa.1968);
As indicated, Sunward asserts that it carried its burden on this critical point by reasonable inferences from circumstantial evidence. We disagree. It relied, and the jury was left to rely, more on speculation than evidence and reasonable inference. Each inference argued by Sunward was not only speculative in itself, but it rested upon predicate speculation, thus compounding the error. The following excerpt from Sun-ward’s closing argument to the jury illustrates the point:
*520In December of ’75, we’ve got sales of $7 million and employees of 45. Now, there is a big jump between 1975 and 1979 and, with the courtesy of the Dun & Bradstreet computer program, which obliterates everything in between, we don’t know what was happening in that period, but let’s suppose that you have been a subscriber to Dun & Bradstreet since 1975 and keep at file on companies, and you take a look at Exhibit 26, which shows sales of $7 million for this outfit called Wedg-Cor, a subsidiary of Marvel Steel, and you take a look at Exhibit 30 four years later, not quite four years later, and you see Marvel Brute Steel doing business as Marvel Steel WedgCor, and as the witness Hannaford, I believe it was, acknowledged on the stand, if you were to make the connection between the names — and offhand, I can’t think of how many companies named Wedg-Cor or Marvel Steel would fit into Jamestown, North Dakota — if you were to make that connection, the logical inference would be that something very bad had happened to this company in the meantime and that its sales had shrunk to a tenth of their former size, and that employees had gone down from 45 to 5.
R. Vol. XI at 652-53. In that argument, in order even to reach the inference counsel sought, the jury had to speculate on: (a) receipt of a report in 1975; (b) retention of that report — or accurate memory of its contents — for four years; (c) its comparison by the recipient with the 1979 report; and (d) a mental adjustment to a change in company names. Sunward then speculates that the 1979 report recipient would infer that “something very bad had happened to this company.” That final speculation must still co-exist with others which are just as plausible. For example, the recipient may have disregarded the report because it stated that Sunward refused to give information, and that sales were the reporter’s estimates; or the recipient may have known that the report was inaccurate because of other information known to the recipient. Another series of speculations is required to conclude that recipients interpreted the small size reported to mean that Sunward was unfit to function in the steel building industry.
Likewise, if the crucial group of recipients was limited to competitors in the small universe of steel building manufacturers, as Sunward now seems to argue, Brief of Appellees at 19-20, a jury would be left to speculate whether the competitors started derogatory rumors because of their understanding of the credit reports or for other reasons. Sunward established at trial that competitors kept close tabs on one another. R. Vol. VI at 299. There was testimony that Sunward’s own employees were circulating rumors about the company’s financial difficulties. R. Vol. IV at 59-60, 63. Did those rumors get picked up by the competitors? Did competitors observe Sunward’s reduction in sales force and use that to start a rumor? Did it obtain information from former salespeople? Did the competitors start rumors about Sunward simply out of competitive spite, based on no information at all?
While all competing inferences do not have to be negated in order to make an asserted inference reasonable, reasonableness itself can be tested by the relevant facts and possibilities in each situation. As just illustrated, by electing not to call any of the scores of identified recipients of the reports to ask them what extrinsic facts they knew, and what they understood the reports to mean, Sunward offers little more than debater’s suppositions instead of reasonable inferences. This is especially true considering the brief and general testimony about the existence of rumors concerning Sunward’s business. This vacuum of proof is further emphasized by pure guesswork as to which of the claimed defamatory meanings was supposedly understood by recipients: financial distress, incompetence, or inability to produce and stand behind its product because of small size alone? It would have been impossible for a jury to make that determination on the evidence presented.
*521Although a jury is entitled to draw reasonable inferences from circumstantial evidence, reasonable inferences themselves must be more than speculation and conjecture. Galloway v. United States, 319 U.S. 372, 63 S.Ct. 1077, 87 L.Ed. 1458 (1943). For example, in Daniels v. Twin Oaks Nursing Home, 692 F.2d 1321 (11th Cir. 1982), the court rejected as too speculative inferences that a nursing home was negligent and therefore liable for a resident’s wrongful death when he wandered off while under the nursing home’s care. Upholding the district court’s grant of judgment n.o.v., the court stated that, “a jury will not be allowed to engage in a degree of speculation and conjecture that renders its finding a guess or mere possibility. Such an inference is infirm because it is not based on the evidence.” Id. at 1326 (citation omitted). See also Hanson & Orth, Inc. v. M/V Jalatarang, 450 F.Supp. 528, 540-41 (S.D.Ga.1978) (evidence that a shipowner was negligent in relation to a fire aboard his ship was too speculative); Thomas v. Young, 282 F.Supp. 52, 55 (E.D. Wis.1968). The line between “reasonable inferences” and mere speculation is impossible to define with any precision. However, the Third Circuit has effectively described the process of distinguishing between reasonable inferences and impermissible speculation:
The line between a reasonable inference that may permissibly be drawn by a jury from basic facts in evidence and an impermissible speculation is not drawn by judicial idiosyncracies. The line is drawn by the laws of logic. If there is an experience of logical probability that an ultimate fact will follow a stated narrative or historical fact, then the jury is given the opportunity to draw a conclusion because there is a reasonable probability that the conclusion flows from the proven facts. As the Supreme Court has stated, “the essential requirement is that mere speculation be not allowed to do duty for probative facts after making due allowance for all reasonably possible inferences favoring the party whose case is attacked.” Galloway v. United States, 319 U.S. 372, 395, 63 S.Ct. 1077, 1089, 87 L.Ed. 1458 (1943).
Tose v. First Pennsylvania Bank, N.A., 648 F.2d 879, 895 (3rd Cir.), cert. denied, 454 U.S. 893, 102 S.Ct. 390, 70 L.Ed.2d 208 (1981).
In this case, we are not confronted with difficult line-drawing determinations. Inferences that the reports were understood as defamatory and that they caused or contributed to Sunward’s financial difficulties are here supported only by speculation and conjecture. The record is devoid of evidence that anyone ever understood the credit reports in the defamatory manner inferred by the plaintiff. The only evidence offered was a chronological rendition of events which, among other things, indicated that the first inaccurate Dun & Bradstreet report preceded Sunward’s financial downturn, rumors of difficulty, and associated problems. Sunward’s argument based on this evidence consists of “reasoning from sequence to consequence, that is, assuming a causal connection between two events merely because one follows the other.” Edward J. Sweeney & Sons, Inc. v. Texaco, Inc., 637 F.2d 105, 116 (3rd Cir. 1980), cert. denied, 451 U.S. 911, 101 S.Ct. 1981, 68 L.Ed.2d 300 (1981).8 The inferenc*522es required to establish proof of defamation in this case do not follow from the evidence and must be rejected.
In this case it was especially important for the evidence fairly to disclose what the recipients understood by these credit reports. The language of the reports was both neutral and qualified. Liability turned on extrinsic facts and multiple alleged innuendos which are clearly more complicated than those found in the usual defamation case. Many complex economic factors are apparent in the case, from changes in the economy generally, to internal forces at work in this business. Furthermore, a determination that the reports were considered to be defamatory directly and seriously impacts the major issue of damages, see infra Section III. B., as well as going to the issue of liability. Under such circumstances courts should be particularly careful not to permit plaintiffs to establish so many conclusions with so much argument and so little evidence.
In partial explanation of its failure to produce any direct evidence that the Dun & Bradstreet reports were understood by recipients as defamatory, Sunward states:
Plaintiffs were faced with the prospect of tracing a pernicious rumor back to its source. Such a task would be difficult under the best of circumstances; given the D & B confidentiality requirements and the status of those receiving the reports, the task is impossible.
Brief of Appellees at 20-21 (emphasis added). It is one thing to argue impossibility as a theoretical proposition, however, and quite another thing to demonstrate it. Sunward offers no evidence to support its claim of impossibility. More than one hundred recipients of these reports were identified. They included competitors, banks, suppliers, bonding companies, and others. R.Vol. IV at 49-52, 61-62; R. Vol. VI at 299. It is not credible that direct evidence of defamatory understanding by recipients was totally unavailable from a group that large and diverse. Reluctance of some to testify must be balanced against the fact that they can be examined under oath, and that letters, memoranda and other documents can be produced under subpoena duces tecum. Furthermore, by requiring such a process, and direct evidence, Dun & Bradstreet is accorded the right to cross-examine, test the evidence, and develop its own case — rights which are denied by allowing Sunward to rely entirely on multiple suppositions. As for alleged difficulties with rumors, they need not be traced to a source as Sunward complains. Sunward has already asserted that the source was report recipients and a sufficient number of those recipients has been identified to test the source theory.
Sunward also contends, in essence, that the potential availability of presumed damages controlled proof of defamation. The proposition is stated as follows:
Under the law of Colorado, as properly interpreted by the trial court, Sunward did not have to prove special damages. Damages were presumed because of the nature of the defamatory report. D & B’s approach would destroy the benefit of the presumption. The presumption exists to allow a defamed party to recover damages in circumstances when, because of the nature of the defamation, substantial damages are likely to have occurred, but proof of those damages may be difficult or impossible____ If, as D & B suggests, a plaintiff must nonetheless produce a witness to testify as to his defamatory interpretation of the statement, the benefits of the presumption would be lost.
Brief of Appellees at 19 (emphasis added). That argument puts the case exactly backwards. It assumes proof of defamation, and proof of the nature of the alleged defamation, to justify the conclusion of entitlement to presumed damages. Then, going full circle, it contends that the benefit of the presumption is lost by requiring *523proof of the fact and nature of the defamation. It is elementary that proof of defamation (along with falsity) precedes all else in a defamation case. See Restatement § 558; see also Walters v. Linhof 559 F.Supp. 1231, 1234 (D.Colo.1983). That is also the point of this entire section. It is equally elementary that proof of the nature of the defamation controls whether or not damages will be presumed. That point is discussed in Section III. B., infra.
Sunward implies, but makes no real attempt to develop an argument, that the court’s pre-trial ruling granted presumed damages as a matter of law. There was no such ruling. The court denied Dun & Bradstreet’s motion for summary judgment and ruled in the process that the reports were capable of the defamatory innuendo alleged — which it has the right to do. Restatement § 614. It then submitted the matter for jury determination which, in a case like this, it must do if the case has survived motions to dismiss. Id.
Furthermore, Sunward’s argument confuses proof of special damage with proof of defamation. While proof of special damage and proof of defamation often overlap, they are not equivalent. Take, for example, a false report which by virtue of known extrinsic facts is interpreted by a recipient bank to impute insolvency. Special damage does not necessarily result. The bank may have already decided on other grounds not to extend credit, or may extend credit despite the report because further investigation or a correction shows the report to be false, or the bank may not even have been asked to extend credit and does not advise anyone else of the report. The incident causes no out-of-pocket or other economic loss to the subject of the report. Thus, there is no special damage. Testimony by bank officials as to how they interpreted the report in light of extrinsic facts known to them goes directly to proof of defamation. An imputation of insolvency proved by that testimony may then be used in evaluating entitlement to presumed damages. Obviously, the plaintiff is not robbed of the benefits of the presumption by requiring “a witness to testify as to his defamatory interpretation of the statement,” as Sunward contends.
In sum, Sunward’s position confuses the order of proof. Presumption of damages neither precedes nor excludes a determination of defamation; rather, it follows and is dependent on that determination. The intertwined nature of liability and damage issues in defamation cases, together with interchangable or similar terms, has created difficulties in this regard, but the point is clear. We reemphasize that the first element of a defamation case is as stated in § 558 of the Restatement: “To create liability for defamation there must be: (a) a false and defamatory statement concerning another.” (emphasis added).
Jury instructions in this case perpetuated the problem. The district court instructed the jury that it must find the reports were false, defamatory, caused damage, and were published with reckless disregard. R.Vol. XI at 718. However, the instructions did not define defamation except as part of a later instruction on presumed damages. R.Vol. XI at 719. And they completely omitted the requirement that the jury must determine whether recipients of the reports understood them in the defamatory sense (and which defamatory sense) alleged by Sunward. The jury’s function in that regard is settled law in Colorado and elsewhere. See M.F. Patterson Dental Supply Co. v. Wadley, 401 F.2d 167, 169 (10th Cir.1968); Walters, 559 F.Supp. at 1236; Rocky Mountain News Printing Co. v. Fridborn, 46 Colo. 440, 104 P. 956, 959 (1909); see also Restatement § 614(2).9
We hold that the evidence of defamation in this case was insufficient to go *524to the jury. Although the issue is not directly raised on appeal, it is a fair extension of our holding also to hold that the instructions to the jury on the issue of defamation were insufficient and erroneous for the reasons stated herein.
As indicated at the conclusion of this opinion we have decided not to reverse this case outright, but to remand for a new trial. For purposes of retrial, it is impossible to state in advance what evidence will be sufficient to go to the jury with respect to Sunward’s burden of proving that recipients of these reports understood them to carry any particular defamatory meaning alleged. However, it is doubtful that the issue could go to a jury without testimony from recipients. It is also doubtful that just one or two recipients could establish the case. General guidance is offered by the Restatement § 559 comment e, which refers to a “substantial and respectable minority.” See Burns v. McGraw-Hill Broadcasting Co., 659 P.2d 1351, 1357 (Colo.1983).10 Since proof of defamation not only goes to establishing liability but also directly affects whether or not presumed damages may be awarded, see infra Section III. B., the seriousness of the issue of defamation cannot be overemphasized.
II.
ABUSE OF PRIVILEGE — DEFINITION OF RECKLESS DISREGARD
The district court ruled, and Sunward does not dispute, that credit reports are conditionally privileged in Colorado.11 The issue here concerns abuse of privilege.
The district court generally articulated the standard for abuse of privilege to be knowledge that the reports were false, or reckless disregard for whether the reports were true or not. The parties are in apparent agreement that such is the standard which should prevail in this case, and no issue on the point is raised on appeal. Therefore, we do not review the basic standard itself. It is the jury instruction defining reckless disregard which is in dispute.
The problem arises from the various ways reckless disregard is defined, depending on context. As a constitutional standard (federal or Colorado) a strict, subjective “serious doubts as to the truth of the publication” definition is employed. In the *525general tort law of Colorado, however, reckless disregard is largely placed in the standard-of-care category, somewhere at or even beyond gross negligence. Common law abuse of privilege cases in Colorado concentrate on a publisher’s motive in defamation cases and add that nuance to reckless disregard when used in that context.12
As the following analysis indicates, we reject Dun & Bradstreet's position that this *526case implicates constitutional First Amendment protections — state or federal — which entitle it to the “serious doubts” definition of reckless disregard. We do, however, find that the jury instruction defining reckless disregard was erroneous. As defined, it permitted the jury to include within the definition of reckless disregard conduct close to ordinary negligence. The tort law of Colorado places reckless disregard much further along the continuum of care — so much so as to constitute a difference in kind. Thus, by way of illustration only, the instruction impermissibly defines something akin to gross negligence in terms close to ordinary negligence.
As the foregoing overview indicates, our analysis focuses first on Dun & Bradstreet’s constitutional definition argument, then discusses Colorado common law definitions of reckless disregard.13
The court rejected differing definitions of reckless disregard offered by the parties and opted for a definition offering the least protection to Dun & Bradstreet. The instruction permitted the jury to conclude that any lack of care beyond ordinary negligence constitutes abuse of the privilege. The entire instruction, with the disputed portion emphasized, is as follows:
The defendant abused the privilege if you find that when it published the reports in question it knew the reports to be false, or acted with reckless disregard for whether the reports were true or not.
R.Vol. XI at 720.
* * * * * *
Reckless disregard implies a higher degree of improper conduct than negligence. A failure to exercise ordinary or reasonable care in ascertaining the truth of published material does not, standing alone, constitute reckless disregard. Such failure, however, may be considered as an element of reckless disregard.
R.Vol. XI at 719 (emphasis added).
Dun & Bradstreet challenges that definition of reckless disregard, and claims entitlement to the following test adopted by the Colorado Supreme Court in Diversified Management, Inc. v. Denver Post, Inc., 653 P.2d 1103 (Colo.1982):
Recklessness implies a higher degree of culpability than negligence. A failure to exercise ordinary or reasonable care in ascertaining the truth of published material does not, standing alone, constitute recklessness. The gist of the meaning of recklessness is that these Defendants had a high degree of awareness that the statements published were probably false.”
Id. at 1109 (emphasis in original).
Diversified was decided on constitutional grounds, both federal and Colorado. It involved an action brought by a private individual and his corporation against the Denver Post and one of its reporters over two articles published in the paper. The articles dealt with certain “widespread and ongoing land-development schemes of questionable propriety” which the Colorado court declared to be a “matter of public or general concern.” Id. at 1108. Because the subject matter was one of public concern the court considered the case to involve core First Amendment values protected by state and federal constitutions from the reach of common law defamation standards. Consequently it applied the constitutional “actual malice” standard developed by the United States Supreme Court in the well-known line of cases beginning with New York Times v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964). See Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc., 472 U.S. 749, 105 S.Ct. 2939, *52786 L.Ed.2d 593 (1985); Gertz v. Robert Welch, Inc., 418 U.S. 323, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1973); Rosenbloom v. Metromedia, Inc., 403 U.S. 29, 91 S.Ct. 1811, 29 L.Ed.2d 296 (1971); Curtis Publishing Co. v. Butts, 388 U.S. 130, 87 S.Ct. 1975, 18 L.Ed.2d 1094 (1967). See also Bose Corp. v. Consumers Union of the United States, Inc., 466 U.S. 485, 104 S.Ct. 1949, 80 L.Ed.2d 502 (1984).14
Diversified is more protective of core First Amendment values than Gertz requires because it places matters of public concern in the same protected category as public officials and public figures, and it does so regardless of the fact that a private plaintiff is involved. The Colorado Supreme Court explains its position in this regard as follows:
In order to honor the commitment to robust debate embodied in the first amendment and to ensure sufficient scope for first amendment values, we chose to extend constitutional protection to any discussion involving matters of public concern, irrespective of the notoriety or anonymity of those involved.
The considerations which led to our adoption of Rosenbloom now cause us to conclude that first amendment values would be better honored by adopting the same definition of “reckless disregard” in cases involving public offidais, public figures, and matters of public or general concern. To the extent that Walker held otherwise, we now overrule it.
653 P.2d at 1106 (emphasis added).
As the foregoing analysis demonstrates, Dun & Bradstreet is entitled to the protective state constitutional standard approved in Diversified only if the credit reports on Sunward were matters of “public or general concern.” In Section III of this opinion we address that question in connection with the Supreme Court’s opinion in Greenmoss and conclude that the credit reports are not matters of public or general concern.
Despite the fact that the Colorado court in Diversified states it is not defining “the outer boundaries of the term ‘public or general concern,’ ” and will “look to the circumstances of the case,” id. at 1108, we do not believe that the Colorado Supreme Court would part company with the plurality opinion of the United States Supreme Court in Greenmoss on the classification of credit reports. Nothing in the decided Colorado cases suggests that Colorado would extend to credit reports its fullest constitutional protection.15 In Rowe v. Metz, 195 Colo. 424, 579 P.2d 83 (1978), the Colorado Supreme Court held that constitutional pro*528tections do not apply to cases involving private plaintiffs and statements made in a purely private context. The court permitted an award of presumed damages in such cases, at least where the challenged statements were defamatory per se, stating:
In reviewing the United States Supreme Court cases in this area, apart from Gertz, attention is focused on the need for constitutional protection of freedom of the press and free speech whenever public officials, public figures, or issues of public concern are involved. Compare, New York Times Co. v. Sullivan, 376 U.S. 254, 84 S.Ct. 710, 11 L.Ed.2d 686 (1964), with Rosenbloom v. Metromedia, Inc., 403 U.S. 29, 91 S.Ct. 1181, 29 L.Ed.2d 296 (1971). A cogent restatement of this proposition appears in New York Times Co. v. Sullivan, supra, where it was said:
“Purely private defamation has little to do with the political ends of a self-governing society. The imposition of liability for private defamation does not abridge the freedom oí public speech or any other freedom protected by the First Amendment, [footnote omitted.]” New York Times Co. v. Sullivan, 376 U.S. at 301-302, 84 S.Ct. at 737, 11 L.Ed.2d at 721 (Goldberg, J. concurring); quoted in Calero v. Del Chemical Corp., 68 Wis.2d 487, 228 N.W.2d 737 (1975).
In this light we think the balance should be struck in favor of the private plaintiff where his reputation has been injured by a nonmedia defendant in a purely private context.
Id. 579 P.2d at 84-85 (emphasis added). See Kuhn v. Tribune-Republican Publishing Co., 637 P.2d 315, 319 (Colo.1981) (“This actual malice standard was meant to remove the inhibitory effect of defamation laws, in essence creating a constitutional privilege for good faith critics of public officials.”) (emphasis added); Walters v. Linhoff, 559 F.Supp. 1231, 1236 (D.Colo. 1983) (“[P]ublic hearings and free and open communication by citizens with public agencies and public officials give the First Amendment its raison d’etre.”), Russell v. McMillen, 685 P.2d 255, 258 (Colo.App. 1984) (“[T]he reporting of a matter involving a public official is protected.”). See also Burns v. McGraw-Hill Broadcasting Co., 659 P.2d 1351 (Colo.1983).
We conclude, therefore, that Dun & Bradstreet is not entitled to the subjective “awareness of probable falsity” or ‘^serious doubts” test, approved in Diversified for determining reckless disregard. That conclusion is not altered by the only two Colorado cases, in addition to Diversified, cited to us by Dun & Bradstreet, Dominguez v. Babcock, 696 P.2d 338 (Colo.App.1984), and Willis v. Perry, 677 P.2d 961 (Colo.App. 1983). Both cases involved public officials or public figures and matters of public interest.16
Likewise, we are not persuaded by Dun & Bradstreet’s reliance on § 600 of the Restatement, which refashions the abuse of privilege standard as a result of the Supreme Court opinion in Gertz. In comment b to § 600 the American Law Institute states that: “reckless disregard as to truth or falsity exists when there is a high degree of awareness of probable falsity or serious doubt as to the truth of the statement.” The same comment then proceeds to refer to public officials and public figures. Section 600 of the Restatement, as defined in comment b, is not “black letter law” as asserted by Dun & Bradstreet. Brief of Appellant at 25. The American Law Institute makes clear that its position with respect to conditional privilege in the wake of Gertz is still tentative and awaits further development of the law. See Restatement, “Special Note on Conditional Privileges and the Constitutional Requirement of Fault,” pp. 259-60; id. § 580B comments c, d, f, and 1; id. § 595 comment *529h (credit agencies); id. § 599 comment d.17 In any event, we do not regard comment b to § 600 as the definition of reckless disregard in credit reporting cases in Colorado.
Our conclusion that Dun & Bradstreet’s requested instruction is wrong does not mean that the jury instruction given by the district court is correct. A definition of reckless disregard meaning simply that it "implies a higher degree of improper conduct than negligence,” without further elucidation, is not supported by any Colorado case, or any other authority. Sunward concedes that “the instruction given by the trial court was not entirely satisfactory to either Sunward or D & B.” Brief of Appellees at 23 n. 10.
[Sunward] sought a definition of reckless disregard such as that set forth in Williams v. Burns, 463 F.Supp. 1278, 1283 (D.Colo.1979) (“Colorado has defined reckless disregard as ‘an act destitute of heed or concern for consequences, especially foolishly heedless of danger; headlong, rash; without thought or care of consequences.’ ”) (quoting Dixson v. Newsweek, Inc., 562 F.2d 626 (10th Cir.1977)).
Id. at 23 n. 10. Nothing in the authorities we have reviewed justifies a definition of reckless disregard in Colorado as simply any conduct more culpable than ordinary negligence. To the contrary, reckless disregard is generally placed at the far end of the continuum of care, short of intentional acts. The point is made in § 500 of the Restatement, defining reckless disregard of safety. Negligence is contrasted with reckless disregard, in part, as follows:
the difference between reckless misconduct and conduct involving only such a quantum of risk as is necessary to make it negligent is a difference in the degree of the risk, but this difference of degree is so marked as to amount substantially to a difference in kind.
Restatement § 500 comment g (emphasis added). The section also points out that reckless disregard “is often referred to as ‘wanton or willful misconduct.’ ” Id. § 500. See also Restatement § 282 comment e. In fashioning its own tort definition of reckless disregard, the Colorado Supreme Court has expressly cited and relied on § 500 of the Restatement. See Coffman v. Godsoe, 142 Colo. 575, 351 P.2d 808, 814-15 (1960); Fanstiel v. Wright, 122 Colo. 451, 222 P.2d 1001, 1002-03 (1950). In Fanstiel the Court referred to a number of different descriptions of reckless disregard:
To be so classified conduct must negative both attention and concern; it must demonstrate indifference as well as inattention to consequences which may result. Millington v. Hiedloff 96 Colo. 581, 45 P.2d 937 [1935]. “Reckless,” as the word is used in the statute, is equivalent to “wanton.”
The requirements of the statute, we think, are well stated in Restatement of the Law — Torts, page 1293, chapter 19, section 500: “The actor’s conduct is in reckless disregard of the safety of another if he intentionally does an act or fails to do an act which it is his duty to the other to do, knowing or having reason to know of facts which would lead a reasonable man to realize that the actor’s conduct not only creates an unreasonable risk of bodily harm to the other but also *530involves a high degree of probability that substantial harm will result to him.” This definition has been approved in numerous jurisdictions____ [citations omitted] “Reckless disregard” has elsewhere been defined as meaning an act destitute of heed or concern for consequences, especially foolishly heedless of danger; headlong, rash; without thought or care for consequences. R.J. Reynolds Tobacco Co. v. Newby, 9 Cir., 145 F.2d 768 [1944].
Id. See also Foster v. Redding, 97 Colo. 4, 45 P.2d 940 (1935); Millington v. Hiedloff, 96 Colo. 581, 45 P.2d 937 (1935); Clark v. Small, 80 Colo. 227, 250 P. 385 (1926). In Dixson v. Newsweek, Inc., 562 F.2d 626 (10th Cir.1977), we referred to part of the foregoing language in Fanstiel and approved a jury instruction stating: “Reckless disregard implies a higher degree of culpability than negligence. Recklessly means wantonly, with indifference to the consequences." Id. at 629 (emphasis added). In Williams v. Burns, 463 F.Supp. 1278 (D.Colo.1979), the district court in Colorado, referring to Dixson, adopted the definition of reckless disregard in Colorado as “an act destitute of heed or concern for consequences, especially foolishly heedless of danger; headlong, rash; without thought or care of the consequences.” Id. at 1283. And, as previously stated, it was that definition which was offered by Sun-ward as the proper instruction to the jury in this case.
Justice Erickson of the Colorado Supreme Court, in his opinion dissenting in part in Walker v. Colorado Springs Sun, Inc., 188 Colo. 86, 538 P.2d 450, 460 (1975), stated that: “Reckless disregard as a standard, when stripped of the interpretation afforded by St. Amant v. Thompson, supra, probably means gross negligence, but certainly does not provide a settled test. See Coffman v. Godsoe, 142 Colo. 575, 351 P.2d 808 (1960); Fanstiel v. Wright, 122 Colo. 451, 222 P.2d 1001 (1950).” Id. 538 P.2d at 465.18
It is clear from Colorado authorities that the definition of reckless disregard is established in the common law of Colorado to be significantly greater than ordinary negligence.19 It is so much greater that it *531was error to fail to instruct the jury as to the seriousness of the conduct circumscribed by that term.
We are fully aware that our decision essentially adopts the reckless disregard tort definition applied in the core First Amendment values case of Walker. See supra note 18. By applying this definition to a credit report case we may appear to be creating an inconsistency with our ruling that credit reports are not matters of public concern and not at the core of First Amendment values. Admittedly, Walker dealt with matters of public concern, and there is no reason to believe that the Colorado Supreme court would have equated the constitutional privilege developed in Walker with the common law qualified privilege accorded to credit reports. However, when Diversified overruled Walker and adopted the subjective St. Amant “serious doubts” standard, the protection given core First Amendment values was greatly extended. That extension, along with the distinction between Walker’s and Diversified ’s “clear and convincing” standard and the “preponderance” standard employed here, permits the use of the Walker definition of reckless disregard in a common law context and does not preempt the application of a standard akin to gross negligence20 when non-core First Amendment values are involved. Furthermore, as we stated previously, supra pp. 29-30 and note 11, the issue of the reckless disregard standard itself is not before us, only the definition of reckless disregard.21
Sunward argues that any error was harmless since it probably did not affect the jury’s verdict and in any event the evidence was sufficient to satisfy any standard of recklessness. Brief of Appellees at 24. See McIlroy v. Dittmer, 732 F.2d 98, 105 (8th Cir.1984); Haddad v. Lockheed California Corp., 720 F.2d 1454, 1459 (9th Cir.1983). We disagree. If properly instructed, the jury may very well have found that Dun & Bradstreet did not abuse its qualified privilege in this case.
On retrial, the parties and the court are not restricted to that portion of the language used in Fanstiel v. Wright, 122 Colo. 451, 222 P.2d 1001 (1950), to which we referred in Dixson and to which the district court referred in Williams v. Burns, 463 *532F.Supp. 1278 (D.Colo.1979). At the least, however, the jury must be instructed in greater detail, and the standard of care fixed at a different point than minimally above ordinary negligence.
Both parties argue at length that the facts of this case conclusively establish their respective positions regardless of what standard is used. Because we are remanding this case for retrial under significantly different ground rules it would not be appropriate for us to address these contentions on the record before us. The evidence upon retrial may well be different in important respects.
III.
PRESUMED DAMAGES
The district court ruled prior to trial, and instructed the jury, that if Sunward was defamed by these credit reports then damage to it would be presumed.22 With the benefit of that legal presumption Sun-ward was relieved of the necessity of proving special damage, that is, of directly linking specific loss to the credit reports by competent evidence.23 Dun & Bradstreet argues that allowing presumed damages was unconstitutional and was erroneous under Colorado common law as well.
A. Constitutional Issue.
Dun & Bradstreet contends that the decision of the United States Supreme Court in Gertz, discussed in the preceding section, see supra note 14, precluded any award of presumed damages to Sunward absent a showing of New York Times “actual malice.”
During the proceedings below the Supreme Court granted certiorari in Greenmoss to determine whether the “rule of Gertz [with respect to presumed damages] applies when the false and defamatory statements do not involve matters of public concern.” 105 S.Ct. 2939 at 2941 (1985). A related issue was whether a Dun & Bradstreet credit report which was the subject of that case involved a matter of public concern. In a plurality opinion 24 the Court held that “permitting recovery of presumed and punitive damages in defamation cases absent a showing of ‘actual malice’ does not violate the First Amendment when the defamatory statements do not involve matters of public concern.” Id. at 2948. It also held “that petitioner’s credit report concerns no public issue.” Id. at 2947.
In so holding, the opinion cited with approval the Colorado Supreme Court decision in Rowe v. Metz, 195 Colo. 424, 579 P.2d 83 (1978), which in relevant particulars anticipated the holding in Greenmoss. Greenmoss, 105 S.Ct. at 2946. In Rowe the Colorado Supreme Court, after balancing the interests at stake, concluded:
In this light we think the balance should be struck in favor of the private plaintiff where his reputation has been injured by a non-media defendant in a purely private context. We hold, therefore, that in such cases, and where the remarks are defamatory per se, the common law rule permitting presumption of damages remains applicable. Accord Harley-Davidson Motor Sports, Inc. v. Markley, 279 Or. 361, 568 P.2d 1359 (1977).
579 P.2d at 85 (emphasis added).
Prior to oral argument both parties submitted supplemental briefs regarding the *533impact of Greenmoss on Dun & Bradstreet’s claim that it is entitled to the constitutional “actual malice” standard in this case. Dun & Bradstreet’s response to Greenmoss is that the “matters of public concern” test for such constitutional protection is factual in nature and requires a remand for fact findings by the district court. Supp. Brief of Appellant at 2. It cites the following language from the opinion: “[W]hether ... speech addresses a matter of public concern must be determined by [the expressions] content, form, and context ... as revealed by the whole record.” Greenmoss, 105 S.Ct. at 2947 (quoting from Connick v. Myers, 461 U.S. 138, 147-48, 103 S.Ct. 1684, 1690, 75 L.Ed.2d 708 (1983)). Dun & Bradstreet calls attention to footnote eight in the plurality opinion, which reiterates that “the protection to be accorded a particular credit report depends on whether the report’s ‘content, form, and context’ indicate that it concerns a public matter.” Id. at 2947 n. 8.
Dun & Bradstreet then cites examples of what it conceives to be public-concern characteristics of its credit reports on Sunward: the information in part replicates that contained in public filings by Sunward; purchasers and competitors were interested in the Dun & Bradstreet information on Sun-ward; the reports were broadly circulated (hundreds of responses to subscribers, plus possible secondary circulation, as opposed to only five reports in Greenmoss), and so on. Supp.Brief of Appellant at 2. Such fact finding, according to Dun & Bradstreet, must show that no matter of public concern is involved, only private matters, and that mixed public- and private-concern statements are entitled to full constitutional protection. Id. at 4. It argues that in establishing its entitlement to constitutional standards, a special verdict on each alleged defamatory statement must address form, content and context to determine whether public concern is involved. Additional arguments raise differences between Vermont and Colorado common law, state constitutional involvement in Colorado, and so on.
As a separate argument, Dun & Bradstreet contends that emphasis by the parties and district court on a distinction between media and nonmedia defendants was a crucial error of law, at odds with the analysis in Greenmoss and requiring a remand in order to allow the district court to reconsider critical rulings under a standard which would afford both kinds of defendants comparable protection where matters of public concern are involved.
We reject all of these and other arguments asserted by Dun & Bradstreet on the issue. On the record before us there is no difference between the credit reports in this case and those in Greenmoss sufficient to merit the constitutional protection in question, either state or federal. Variations between Vermont and Colorado law do not change that conclusion. Furthermore, there is no other difference sufficient to require a remand for fact finding purposes. Certainly the size of the business and the numerical difference between five reports in Greenmoss and 340 reports here have no constitutional significance; neither does the difference in credit information reported, nor the fact that some of that information reflected public filings. The public interest in requiring public filings does not necessarily and automatically attach to and follow extracts from those filings and convert private reports compiled for private purposes into a matter of public concern. Likewise, “the actual interest of the subscribers and recipients in obtaining the type of information” contained in the Sunward reports, id. at 7, fails to create any possibility of a constitutionally protected public interest in this case. We also reject the argument that qualified privilege necessarily equates on either a state or federal basis with constitutional “public concern.”
The plurality evaluation of the Dun & Bradstreet credit reports in Greenmoss applies with equal force to the reports in this case:
There is simply no credible argument that this type of credit reporting requires special protection to ensure that “debate on public issues [will] be uninhibited, ro*534bust, and wide-open.” New York Times Co. v. Sullivan, 376 U.S., at 270, 84 S.Ct., at 720.
In addition, the speech here, like advertising, is hardy and unlikely to be deterred by incidental state regulation. See Virginia Pharmacy Board v. Virginia Consumer Council, supra, 425 U.S. [748], at 771-772, 96 S.Ct. [1817], at 1830-1831 [48 L.Ed.2d 346 (1976)]. It is solely motivated by the desire for profit, which, we have noted, is a force less likely to be deterred than others. Ibid. Arguably, the reporting here was also more objectively verifiable than speech deserving of greater protection. See ibid. In any case, the market provides a powerful incentive to a credit reporting agency to be accurate, since false credit reporting is of no use to creditors. Thus, any incremental “chilling” effect of libel suits would be of decreased significance.
Greenmoss, 105 S.Ct. at 2947. See also Kansas Elec. Supply Co. v. Dun & Bradstreet, Inc., 448 F.2d 647, 649 (10th Cir. 1971), cert. denied, 405 U.S. 1026, 92 S.Ct. 1289, 31 L.Ed.2d 486 (1972).25
Finally, as Sunward points out in its supplemental brief, and as discussed previously in this opinion, at the trial of this case Dun & Bradstreet placed heavy reliance upon the Colorado Supreme Court opinion in Diversified, 653 P.2d at 1109, for its jury instructions in connection with the definition of “reckless disregard.”26 The “public or general concern” issue was a central feature in that case, the Walker case, 538 P.2d 450, and other Colorado decisions and has been a crucial underpinning of the re-suiting standard in Colorado.27 As a result, and despite references in Diversified to media/non-media distinctions, Dun & Bradstreet had both reason and opportunity in the proceedings below to develop its argument that the credit reports on Sun-ward involved matters of public concern and to make and preserve a record on the point.
We have considered all of Dun & Bradstreet’s many arguments and distinctions (including arguments that Diversified developed Colorado common law, not constitutional law) and regard them as unpersuasive. Constitutional issues do not prevail in this case and no remand is necessary on this point. The district court’s denials of Dun & Bradstreet’s proposed jury instruction based on Gertz and Diversified, and of post-trial motions based on the same point are affirmed and become the law of the case for further proceedings below. On retrial Dun & Bradstreet is precluded from reopening the issue.
B. The Business Exception Issue.
Dun & Bradstreet also invokes common law principles to challenge the allowance of presumed damages. Colorado is one of the states which treats actions alleging a libel per quod as an action for slander insofar as available damage remedies are concerned. That is, libel per quod is generally actionable only where special damages (actual economic loss alleged and proved to have resulted directly from the defamatory publication) are pleaded and proved. See Bernstein v. Dun & Bradstreet, Inc., 149 Colo. 150, 368 P.2d 780, 782-83 (1962); *535Brown v. Barnes, 133 Colo. 411, 296 P.2d 739, 741 (1956); Knapp v. Post Printing & Publishing Co., 111 Colo. 492, 144 P.2d 981, 985 (1943); Fort v. Holt, 508 P.2d 792, 793 (Colo.App.1973); Inter-State Detective Bureau, Inc. v. Denver Post, Inc., 29 Colo. App. 313, 484 P.2d 181, 133 (1971).
Under generally recognized exceptions to that rule, however, special damages need not be pleaded and proved, and presumed damages may be recovered, if the defamatory communication falls within any one of four subject matter categories which have been accorded special treatment over the years: imputations of unchastity to a woman, or serious sexual misconduct; imputation of a criminal offense; imputation of a loathsome disease; or defamation affecting business, trade, profession or office. Bernstein, 368 P.2d at 783 (citing Prosser, Libel Per Quod, 46 Va.L.Rev. 838, 844 (1960)). See also Wegner v. Rodeo Cowboys Ass’n, 290 F.Supp. 369, 373-75 (D.Colo.1968); Big O Tire Dealers, Inc. v. Goodyear Tire & Rubber Co., 408 F.Supp. 1219, 1234-35 (D.Colo.1976), modified, 561 F.2d 1365 (10th Cir.1977), cert. dismissed, 434 U.S. 1052, 98 S.Ct. 905, 54 L.Ed.2d 805 (1978); Fort, 508 P.2d at 793; Restatement §§ 570-574. For convenience, the latter category is referred to as the “business exception.” It is set out and defined in § 573 of the Restatement as part of the rules governing cases of slander.
The Colorado rule on libel per quod is at odds with the position of the American Law Institute. In § 569 of the Restatement28 the Institute continues to adhere to the historic rule that, unlike slander, damage is presumed in all libels.29 Section 569 also defines a business libel more broadly than does § 573, relating to slanders. See Restatement § 569 comment e. In any event, the point is of only marginal significance because, as indicated, Colorado does not follow the Institute’s views as set out in § 569.30 That conclusion is not disputed by either party. Furthermore, in the present appeal neither party challenges the general applicability of the business exception in Colorado, even though it has not been specifically adopted by the Colorado Supreme Court but has only been cited in dicta. See Bernstein, 368 P.2d at 783. See also Wegner, 290 F.Supp. 369; Big O Tire, 408 F.Supp. 1219; Fort, 508 P.2d at 793.
Prior to trial the district court ruled as a matter of law that the credit reports on Sunward fell within the business exception, and that, therefore, Sunward could recover presumed damages. The court stated:
In the instant action, plaintiffs, multimillion dollar affiliate corporations with hundreds of employees, were characterized as a business employing only five persons with sales in the range of $500,-000 to $750,000. This is not a good reference for a company engaged in the manufacture and sale of steel buildings and could reflect on plaintiffs’ fitness to conduct its business. I hold that the facts of this case fall within this recognized exception, that damages are presumed and need not be specifically alleged as argued by defendant.
568 F.Supp. at 606 (emphasis added). It is this ruling, reflected in an instruction permitting presumed damages to be found by the jury, which is in dispute.
*536Dun & Bradstreet asserts that the business exception is not broad enough to include neutral language that merely is “not a good reference,” and “could reflect” on fitness to conduct business. Rather, it argues that the business exception is limited to direct imputations of “incompetence, dishonesty or misconduct which is incompatible with the proper conduct of plaintiffs’ profession or business.” Bernstein, 368 P.2d at 784.31 See also Prosser on Torts § 112 at 758 (4th ed. 1971) (“the [business] exception was limited to defamation of a kind incompatible with the proper conduct of the business, trade, profession or office itself.”) (emphasis added). Indeed, § 573 of the Restatement, on which the district court itself relied, specifies publications as slanderous which ascribe “conduct, characteristics or a condition that would adversely affect his fitness for the proper conduct of his lawful business.” Restatement § 573 at 191 (emphasis added). The district court also relies on Wegner, which emphasizes § 569 of the Restatement describing defamatory utterances as those “imputing any misconduct whatever in the conduct of the other’s calling.” Sunward, 568 F.Supp. at 606 (quoting Wegner, 290 F.Supp. at 374) (emphasis added).32
The problems with this issue arise largely from and are a direct reflection of the errors committed at the level of proof of defamation in this case. Multiple defamatory innuendoes were argued to the jury by Sunward but the general verdict, along with insufficient evidence of defamation, makes it impossible to know what the defamation, if any, was. Thus, analysis is compromised at the threshold. Which innuendo? 33
Sunward’s main contention is that the credit reports were interpreted by recipients as imputing a financially straitened condition to Sunward due to an inferred precipitous decline in its business. The district court’s opinion was predicated on the point that proof of special damages is not needed “where mercantile dishonesty or insolvency is imputed.” Sunward, 568 F.Supp. at 606. Sunward suggests that the companion inference is gross incompetence on the part of management. Either inference, if proved, would subject Dun & Bradstreet to liability without proof of special damages under definitions contained in § 573 of the Restatement, and, therefore, would automatically fall within the definitions contained in § 569 comment e. Under Colorado law, false imputations of business insolvency or financial embarrassment are defamatory. In McKenzie v. Denver Times Publishing Co., 3 Colo.App. 554, 34 P. 577 (1893), the Colorado Court of Appeals made it clear that imputations of financial difficulty, where a business is concerned, are libelous per se and actionable without the necessity of proving special damage, that is the plaintiff is entitled to presumed damages in such a case. The court summarized the rule as follows: “The law has special regard for the credit of merchants and traders, and words which impute to them insolvency, financial difficulty or embarrassment, or dishonesty or fraud, are actionable in themselves, without the necessity of alleging or proving special damages. Upon this proposition *537the authorities are unanimous.” Id. 34 P. at 577. See also Cinquanta v. Burdett, 154 Colo. 37, 388 P.2d 779, 780-81 (1964); Kobey v. Eddy, 21 Colo.App. 140, 121 P. 948 (1912); Annotation, Imputation of Insolvency as Defamatory, 49 A.L.R.3d 163, (1973).
There remains, however, the final category of defamatory meaning alleged by Sunward, based on size alone, which bears independent consideration. To the extent Sunward argues, or the district court’s ruling is based upon the general proposition, that an inaccurate report of small size alone constitutes a libel falling within the business exception we reject it. Such rejection includes the district court’s conclusion that smallness in the steel building industry “is not a good reference,” and “could reflect” on fitness. Sunward, 568 F.Supp. at 606. Not one case is cited to us in direct support of the proposition that it is defamatory within the meaning of the business exception to underestimate the size of a business. In Seaboard Warehouse Terminals, Inc. v. Dun & Bradstreet, Inc., 328 F.Supp. 291 (S.D.Fla.1971), aff'd, 458 F.2d 160 (5th Cir.1972), a report allegedly understating the level of plaintiff’s creditworthiness was held not to be libelous per se, and plaintiff was required but unable to plead special damages under libel per quod. The court stated:
Even though such a credit rating might in truth be an understatement of the plaintiff’s financial worth, in actual practice a “fair” credit appraisal is not per se incompatible with the proper exercise of the plaintiff’s trade or business and would not in the natural course of events cause damage to the plaintiff.
Id. at 292. We adhere to that general sentiment.
In this country it is not disgraceful or suggestive of unfitness or misconduct to be a small business. General charges that a small business is unable to stand behind its warranties or provide an adequate engineer’s certification are debatable. A small business may not only be highly scrupulous and ethical about its warranties and engineering work, it may give better service than its larger competitors in an effort to capture more of the market. The speculative inference that a small business is inefficient, anemic, or in decline is balanced against the equally plausible inference that the business is robust, technically advanced, and in a strong growth pattern because it is very good and staffed with exceptional people.
Furthermore, evidence that a few people failed to patronize Sunward solely because it was small would not support a charge of misconduct or unfitness within the business exception. It probably would not even support a charge of defamation despite the accepted definition of a defamatory communication as one lowering estimation or deterring persons from dealing. The law of defamation does not cover every publication error which may lower business traffic, e.g., a mistake in an address, name, or telephone number. Defamation includes elements of hatred, ridicule, or contempt. Restatement § 559 comment b. There is an actively derogatory aspect to defamatory language which is missing in these types of errors.
There may be instances where size alone carries a defamatory meaning, but as the proposition, is presented in this case it would take proof of special damage to establish it. Nothing in this record, as it now stands, satisfactorily establishes any actionable defamation based on size alone. Certainly there was no evidence at trial sufficient to justify application of the business exception to such an innuendo. Instructions permitting presumed damages on this basis were erroneous.
On retrial, if Sunward asserts multiple innuendos it must do so with specificity, and special interrogatories or verdicts must be employed, probably in conjunction with proof of defamation, i.e., proof that recipients understood the reports in a particular way.34 A finding that recipients under*538stood the credit reports to mean that Sun-ward was insolvent, in financial difficulty, or unfit to conduct its business, because of a steep decline in sales and employees would permit a presumption of damages under the business exception. Of course, such a finding must include a foundation, particularly as to the 1979 fiscal year, that the recipient had knowledge of Sunward’s size and interpreted the credit report to mean that a sharp decline had occurred. Years following fiscal 1979 must, of course, be treated separately, since actual declines had occurred. A recipient’s knowledge of the actual decline and interpretation of the credit reports would have to be specially treated, in context.
If the jury finds that recipients only interpreted the reports to mean that Sun-ward was small, then special damages must be proved. In that connection, the pleadings (which do allege , special damage) may need clarification. Any such clarification, however, would only be as permitted by the district court, within its discretion, to the extent it would have permitted amendments or modifications at the time of its published rulings and thereafter.
IV.
DAMAGES
Dun & Bradstreet attacks Sunward’s damage evidence on multiple grounds: improper introduction of special damages without proof of causation; impermissible speculation; and no foundation for expert testimony. The damage award is also challenged as excessive. Because this case is being remanded for a new trial it is only partially relevant to analyze these contentions, and Dun & Bradstreet’s remittitur argument is entirely mooted by the remand.
Based on the district court’s pretrial ruling and subsequent instruction to the jury that damages could be presumed in this case, Sunward relied entirely on a presumption of general damages. It elected not to present any proof of special damage, i.e., actual economic loss, reasonably established as to amount and proved to have been proximately caused by publication of the credit reports. See, Continental Nut Company v. Robert L. Berner Co., 393 F.2d 283, 286 (7th Cir.), cert. denied, 393 U.S. 923, 89 S.Ct. 254, 21 L.Ed.2d 259 (1968); Bernstein v. Dun & Bradstreet, Inc., 149 Colo. 150, 368 P.2d 780, 782-84 (1962) ; Fort v. Holt, 508 P.2d 792, 793 (Colo.App.1973). See generally M.F. Patterson Dental Supply Co. v. Wadley, 401 F.2d 167 (10th Cir.1968); Cinquanta v. Burdett, 154 Colo. 37, 388 P.2d 779, 780 (1963) ; Brown v. Barnes, 133 Colo. 411, 296 P.2d 739 (1956); Knapp v. Post Printing & Publishing Co., 111 Colo. 492, 144 P.2d 981, 985 (1943); Dorr v. C.B. Johnson, Inc., 660 P.2d 517, 520 (Colo.App. 1983); Inter-State Detective Bureau, Inc. v. Denver Post, Inc., 29 Colo.App. 313, 484 P.2d 131 (1971); R. Sack, Libel, Slander and Related Problems at 346-47 (1980); Restatement §§ 621, 622, 622A and 904.
Ascertainment of presumed general damages is difficult at best and unavoidably includes an element of speculation. See Kendall v. Lively, 94 Colo. 483, 31 P.2d 343, 344 (1934) (presumed damages “are recoverable by inference of law and require no evidence.” (citing Newell, Slander and Libel § 721 (4th ed. 1924)). See, also Prosser and Keeton on Torts § 112 at 788 (5th ed. 1984) (jury is permitted to estimate amount of damages); Restatement § 621 comment a (“This presumption of general damage to reputation from a defamatory publication that is actionable per se affords little control by the court over the jury in assessing the amount of damages.”). As the Supreme Court of Colorado observed in Rowe v. Metz, 195 Colo. 424, 579 P.2d 83 (1978):
This state very early adopted the rule permitting presumption of damages in certain so-called per se situations. See, McKenzie v. Denver Times Publishing Co., 3 Colo.App. 554, 34 P. 577 (1893). *539The rationale for this rule derived from the difficulty of proving damages in these instances. This is particularly true where, as here, the defamatory remarks relate to the conduct of an individual’s business affairs. It is the rare case in which a slander will destroy business profits in such a way that the loss can be directly traced to the slanderous remarks. Thus, due to the intangible nature of the harm resulting from certain slanders and libels, the common law retained the notion of presumed damages for various types of defamatory utterances. See generally, Restatement (Second) of Torts, §§ 569-570.
Id. 579 P.2d at 84. See generally Big O Tire Dealers, Inc. v. Goodyear Tire & Rubber Co., 408 F.Supp. 1219 (D.Colo. 1976) , modified, 561 F.2d 1365 (10th Cir. 1977) , cert. dismissed, 434 U.S. 1052, 98 S.Ct. 905, 54 L.Ed.2d 805 (1978). The same view is expressed in the plurality opinion of the Supreme Court in Greenmoss.
The rationale of the common law rules has been the experience and judgment of history that “proof of actual damage will be impossible in a great many eases where, from the character of the defamatory words and the circumstances of publication, it is all but certain that serious harm has resulted in fact.” W. Prosser, Law of Torts § 112, p. 765 (4th ed. 1971); accord, Rowe v. Metz, supra, 195 Colo., at 425-426, 579 P.2d, at 84; Note, Developments in the Law — Defamation, 69 Harv.L.Rev. 875, 891-892 (1956). As a result, courts for centuries have allowed juries to presume that some damage occurred from many defamatory utterances and publications. Restatement of Torts § 568, comment b, at 162 (1938) (noting that Hale announced that damages were to be presumed for libel as early as 1670). This rule furthers the state interest in providing remedies for defamation by ensuring that those remedies are effective.
105 S.Ct. at 2946.
In considering presumed general damages it is usually permissible, as a matter of background only, to show a general decline in sales and a general diminution of profits without the necessity of naming particular customers or proving why customers are not purchasing the plaintiff’s product. See Cook v. Safeway Stores, Inc., 266 Or. 77, 511 P.2d 375, 378-79 (1973) (quoting Newell, Slander and Libel § 751 at 839 (4th ed. 1924)); see also C. McCormick, Damages § 116 (1935); cf. Vojak v. Jensen, 161 N.W.2d 100, 108 (Iowa 1968) (diminution of business may be shown to prove general damages, but preliminary foundation required).
However, absent proof of causation it is not permissible to suggest the loss of specific customers, orders, profits or other specific pecuniary loss. See, e.g., Vojak, 161 N.W.2d 100; DeLashmitt v. Journal Publishing Co., 166 Or. 650, 114 P.2d 1018 (1941). See also Maheu v. Hughes Tool Co., 569 F.2d 459 (9th Cir.1977); Continental Nut Co., 393 F.2d at 286. It is equally impermissible to use general background information regarding a decline in sales or profits to imply causation and special damage where the plaintiff relied solely on the presumption of damage, having elected not to prove special damage. Id. Of course, presumed general damages may be augmented by proof of special damage.
As previously stated, Sunward elected not to attempt to prove any special damage but to rely solely on the presumption of general damage. It then proceeded, however, to present a specific damage theory and specific damage figures, based on a calculation of “lost profits” presented by its expert, Richard Clark, an accountant and former part-time employee of Sun-ward. Mr. Clark’s calculation of lost profits was based upon selective assumptions and projections. First, he attempted to identify “lost” sales by taking 80% of the average sales of members of the Metal Building Manufacturers Association (MBMA) for the years 1980-1982.35 The 80% assumption was employed on the *540ground that Sunward’s sales for the fiscal years ended February 28, 1978-1980 amounted to approximately 80% of the MBMA sales average. Pl.Ex. 66.36 No variables, such as geographic market, member variables within the MBMA, and so on, were isolated and analyzed by Mr. Clark. He relied solely upon his 80% figure, nothing more, except for one adjustment at trial to account for farm/commercial mix. R.Vol. VII at 370.
Having derived an assumed sales figure for fiscal years ended February 28, 1981-1983, Mr. Clark then applied assumed profit margins of 6.96% and 1.85% to the assumed sales figures. Those profit margins were based on an average of Sunward’s profits for the fiscal years ended February 28,1978-1980 (6.96%), and profit realized in fiscal year ended February 29, 1980 alone (1.85%). Pl.Ex. 66. Using the assumed sales and profit figures Mr. Clark then estimated that Sunward should have earned between $7,637,119 and $11,294,524 for the fiscal years ended February 28, 1981-1983. Id. The jury clearly tied its deliberation to those figures, awarding about one-half of the lower amount.37
Like the sales figures, the assumed profit percentages of 1.6% and 6.96% took no variables into account. For instance, profit margins were declining each year during the fiscal years ended February 28, 1978-1980 from 18%, to 10%, to 1.85%. See Ex. 66, R. Vol. VII at 390-93. Those declines were occurring despite annual increases in sales. Mr. Clark acknowledged that had this trend continued, profits would have been zero in 1980. Id. Furthermore, Mr. Clark took into account interest rate assumptions which did not accord with Sun-ward’s actual interest expenses; profits from nonrecurring capital expenditures; and made other assumptions, both implicit and explicit, which were purely speculative. It is not surprising that Mr. Clark’s lowest projection of profits for the years 1981-1983 ($7,637,119) was a higher figure than Sunward had ever earned in a three-year period. Other inconsistencies abound, but the most notable departures from reality relate to the general downturn in the economy in the early 1980’s, Mr. Wirth's altered manner of doing business (from commissioned salesmen to dealers) and other forces at work within Sunward itself.
Sunward contends that the various inconsistencies, speculations, and omissions in the Clark study go only to weight, not to admissibility, citing DeVries v. Starr, 393 F.2d 9 (10th Cir.1968), in which this court considered a challenge concerning evidence of plaintiffs’ lost customers and a resulting lost profits computation. We concluded in DeVries that the “arguable validity” of factors entering into plaintiffs’ lost profits calculation “goes to the evaluation and probative significance of the computation, not to the entitlement to consideration, of such loss of profits, as was reasonably established.” Id. at 20. DeVries is distinguishable from the presumed general damage problems encountered in this case, but the following excerpts are instructive. Quoting from our earlier opinion in Hoffer Oil Corp. v. Carpenter, 34 F.2d 589, 592 (10th Cir.1929), cert. denied, 280 U.S. 608, 50 S.Ct. 158, 74 L.Ed. 651 (1930), we said:
The general rule is that, where the cause and existence of damages has been established with requisite certainty, recovery will not be denied because such damages are difficult of ascertainment. * * *
In Straus v. Victor Talking M(ach). Co., supra, page 802 of 297 F., the court said: “The constant tendency of the courts is
*541to find some way in which damages can be awarded where a wrong has been done. Difficulty of ascertainment is no longer confused with right of recovery.” * * *
A reasonable basis for computation, and the best evidence which is obtainable under the circumstances of the case and which will enable the jury to arrive at an approximate estimate of the loss, is sufficient.
While the damages may not be determined by mere speculation or guess, it is enough if the evidence shows the extent of the damages as a matter of just and reasonable inference, although the result be only approximate. However, the plaintiff must establish his damage by the most accurate basis possible under the circumstances. He must produce the best evidence reasonably obtainable.
393 F.2d at 17 (emphasis added).
The underscored language focuses on the problems we perceive in Mr. Clark’s testimony. First, that testimony is so speculative that it violates the requirements that conclusions as to damage must have a rational basis and be the result of just and reasonable inference. Mr. Clark’s calculations were nothing more than an exercise in creative mathematics. His “charts are an ‘array of figures conveying a delusive impression of exactness in an area where a jury’s common sense is less available than usual to protect it.’ ” Maheu, 569 F.2d at 476 (quoting Friendly, J. in Herman Schwabe, Inc. v. United Shoe Machinery Corp., 297 F.2d 906, 912 (2d Cir.), cert. denied, 369 U.S. 865, 82 S.Ct. 1031, 8 L.Ed.2d 85 (1962)).
Second, and more fundamentally, the testimony of Mr. Clark impermissibly crossed the line between presumed general damage and proof of special damage. The jury was led to infer that the “lost profits” were proximately caused by the credit reports, and represented special damages resulting from those reports. Thus, Sunward attempted to accomplish indirectly what it openly and consciously elected not to do directly, that is prove that the credit reports caused it identifiable pecuniary loss. Presumed general damages must be approached in an entirely different and, by definition, more general manner.
On retrial neither the district court nor the parties will be assisted much by these findings. The chief difficulty, as we have indicated, is in the nature of presumed damages themselves. They are “an oddity of tort law” and permit the “largely uncontrolled discretion of juries to award damages where there is no loss,” potentially resulting in “gratuitous awards of money damages far in excess of any actual injury.” Gertz v. Robert Welch, Inc., 418 U.S. at 349, 94 S.Ct. at 3012. See Restatement § 621 comment b.
Presumed damages are not well suited to cases like this. The more reasonable view of damages in these situations is expressed by Justice Brennan in his dissent in Green-moss, where he stated:
[T]he reputational interest at stake here is that of a corporation. Similarly, that this speech is solely commercial in nature undercuts the argument that presumed damages should be unrestrained in actions like this one because actual harm will be difficult to prove. If the credit report is viewed as commercial expression, proving that actual damages occurred is relatively easy. For instance, an alleged libel concerning a bank’s customer may cause the bank to lower the credit limit or raise the interest rate charged that customer. The commercial context does not increase the need for presumed damages, but if anything reduces the need to presume harm. At worst the commercial damages caused by such action should be no more difficult to ascertain than many other traditional elements of tort damages. See, e.g., Russell v. City of Wildwood, 428 F.2d 1176, 1181 (CA3 1970) (future earnings); Seffert v. Los Angeles Transit Lines, 56 Cal.2d 498, 509, 15 Cal.Rptr. 161, 168, 364 P.2d 337, 344 (1961) (Traynor, J., dissenting) (pain and suffering).
105 S.Ct. at 2964 n. 16.
But that is not the law of this case. Assuming all prerequisite elements are es*542tablished, Sunward can claim the presumption of general damage. The district court has so ruled and we find no sufficient basis for error as to the conclusion that such damages might be available. Therefore on retrial both parties are necessarily potentially exposed to some risk in a jury determination (estimation) of presumed damages. Of course, if Sunward elects to attempt to prove special damages the situation will be altered.
The burden falls upon the district court to minimize the risk of an unreasonable damage award by educating the jury as fully as possible with respect to its task, without, of course, suggesting that damages must be awarded. If figures showing Sunward’s 56.5% decline in sales, Pl.Ex. 66, par. 4, are presented, the jury must be cautioned that they are to be used as background information only, and that they are not to assume either causation or amount of damage from those figures.38 Any figures showing Sunward’s decline in sales must be tempered by the demonstrated pri- or decline in profits, and an instruction that loss of profits, not sales, is the ultimate characteristic of damage. When the plaintiff elects to pursue only the presumption of general damages, the presentation of a damage argument must also be general, however frustrating that may be to all concerned.
y.
CONCLUSION
Our findings in this opinion may be summarized generally as follows. For Sun-ward to establish liability it was first required to prove that the credit reports were understood by recipients to have the specific defamatory meanings alleged by Sun-ward. It failed to do so; and the jury was underinstructed on this crucial issue. Even upon proper proof of defamation, Sunward was still required to prove that Dun & Bradstreet abused its qualified privilege to make certain mistakes in the issuance of credit reports, by conduct amounting to reckless disregard. That requirement failed because the jury was improperly instructed as to the definition of reckless disregard. Thus, Sunward’s case failed on both of the major requirements for establishing liability: proof of defamation, and proof that Dun & Bradstreet abused its qualified privilege.
Proof of damage was also defective. Sunward failed to prove that it was entitled to a presumption of damage, and, therefore, relieved of the necessity of pleading and proving special damage directly attributable to the credit reports. Presumption of damage was tied directly to proof of defamation in that the presumption only applied if recipients of the reports understood them to attribute financial distress or incompetence to Sunward. There was no acceptable proof in that regard. Finally, proof of the amount of damage was defective.
Sunward’s failure to prove defamation would normally result in outright reversal of the judgment. However, after careful deliberation we have determined to remand this case for a new trial. We do so in the interests of justice because it appears that the pretrial ruling on presumption of damages contained language which may have led Sunward to try its case on certain erroneous assumptions.
It is because of our decision to remand for a new trial that this opinion has been expanded to cover points which otherwise would have been unnecessary, only the principal ones of which are summarized here. In the process we have affirmed a number of important rulings by the district court. Certain other rulings are declared to be the law of the case on retrial.
We have considered all the arguments made by the parties, addressing those we considered necessary on the record as presented, taking the remand into account. For the reasons discussed, certain rulings by the district court are affirmed and others reversed; the judgment is reversed, *543and the case remanded for a new trial consistent with this opinion.
. Combined annual sales for Sunward for fiscal years 1973-1982 were as follows:
[[Image here]]
Ex. 64, R. Vol. IV at 81.
. There is a cause and effect disagreement over the timing of the reduction in commissioned sales personnel. In 1978 Sunward decided to gradually phase out its sales force and convert to sales through dealers. Mr. Wirth testified that he was forced to accelerate the changeover during the time period when the erroneous credit reports were being issued. R. Vol. IV at 79-80, 85-87.
. Under common law pleading, when the defamatory meaning of a communication depended upon extrinsic circumstances, those circumstances, as pleaded in the complaint, were referred to as the "inducement.” The plaintiffs explanation of the alleged defamatory words in light of the extrinsic circumstances was called the "innuendo.” Restatement (Second) of Torts § 563(f) (1977). While pleading practices have been liberalized, a plaintiff must still employ the same structure of extrinsic facts and innuendo drawn from the alleged defamatory words in light of those facts to prove defamation in a case of libel per quod. See Prosser and Keeton on Torts, § 111 at 780, 782 (5th Ed. 1984). See generally Inter-State Detective Bureau, Inc. v. Denver Post, Inc., 29 Colo.App. 313, 484 P.2d 131, 133 (1971).
. The 1975 report on Sunward came to light only after the case was filed. However, the other essential external facts relied upon at trial were obviously known when suit was filed.
. We disregard the testimony of Linda and Sydney Frakes regarding a telephone report that Sunward was "under investigation.” The district court ruled as a matter of law that such a notation is not defamatory. Sunward does not appeal that ruling.
. Unless otherwise noted, all further references to the Restatement in this opinion are to the Restatement (Second) of Torts (1977).
. Technically, the requirement that a declaration must be understood in a defamatory sense applies to all defamation cases and includes both libel per se and libel per quod. For practical purposes, however, the issue does not often arise in libel per se cases. For example, where there is a false report of bankruptcy it is obviously unnecessary to prove that the recipient of the report knows the meaning of the word, Situations where a declaration libelous per se is not understood by the recipient because, for instance, a foreign language is involved, fall within the definition of what constitutes publication. See Restatement § 577.
. This form of reasoning represents a logical fallacy known as the post hoc ergo propter hoc (after this and therefore because of this) fallacy. Other courts have held that a conclusion based upon such reasoning is not a reasonable inference but is mere speculation and conjecture. See, e.g., Loesch v. United States, 645 F.2d 905, 914-15, 227 Ct.Cl. 34 (rejecting an inference of a taking based upon evidence that erosion of plaintiffs’ riverbanks was not a problem until after the government constructed certain dams and locks on the river), cert. denied, 454 U.S. 1099, 102 S.Ct. 672, 70 L.Ed.2d 640 (1981); Edward J. Sweeney & Sons, 637 F.2d at 115-17 (evidence that competitors’ complaints to supplier preceded supplier’s termination of agreement with plaintiff was insufficient evidence to permit a reasonable inference that defendant terminated plaintiffs distributor agreement because of competitor’s complaints); Dodge Motor Trucks, Inc. v. First National Bank, 519 F.2d 578, 584 (8th Cir. 1975) (evidence that a seller, in selling a car to a buyer who later went bankrupt, had relied on a letter (allegedly a guaranty of credit) written by the defendant bank was *522insufficient to hold the bank liable; “merely because the letter preceded the injury and the injury followed hard on the heels of the letter does not establish the relation of cause and effect" (quoting Ligget v. Levy, 233 Mo. 590, 136 S.W. 299, 303 (1911)).
. Restatement § 614(2) provides: “The jury determines whether a communication, capable of a defamatory meaning, was so understood by its recipient.” In Martin v. Outboard Marine Corp., 15 Wis.2d 452, 113 N.W.2d 135 (1962), for exam-pie, the Wisconsin Supreme Court held it was error not to submit to the jury the question of whether recipients of a communication understood it in the defamatory sense alleged by the plaintiff.
. Even producing witnesses who testify that they understood the communication in its defamatory sense may be insufficient to establish the defamation. See Brown v. Barnes, 133 Colo. 411, 296 P.2d 739, 740-41 (1956). Proof that the communication was understood as defamatory will depend upon the quality of the testimony measured against other evidence in the case.
. The Colorado Supreme Court has not yet passed on the question of the privilege, if any, accorded to reports issued by a credit reporting agency. Upon reviewing the authorities, the district court concluded that Colorado would follow the majority of states in extending a qualified privilege. Sunward Corp. v. Dun & Bradstreet, Inc., 568 F.Supp. 602, 606-08 (1983). See Greenmoss, 472 U.S. 749, 105 S.Ct. 2939, 2963, 86 L.Ed.2d 593 (1985) (Brennan, J., dissenting) ("It is worth noting in this regard that the common law of most states ... recognizes a qualified privilege for reports like that at issue here. See Maurer, Common Law Defamation and the Fair Credit Reporting Act, 72 Geo.L.Rev. [sic] 95, 99-105 (1983).”). However, the possible intrusion of constitutional fault standards into the area of common law privilege (depending on findings as to public concern and the impact of Greenmoss on the constitutional requirement of fault under Gertz v. Robert Welch, Inc., 418 U.S. 323, 94 S.Ct. 2997, 41 L.Ed.2d 789 (1973)) will undoubtedly cause some states to review their position on this subject. See infra note 14 for a discussion of Gertz and other Supreme Court cases. The passage of state and federal Fair Credit Reporting Acts, as well as continuing judicial reevaluation and adaptation of common law principles to credit reporting agencies may prompt further review of this issue. Accordingly, all of the cited authorities, as they relate to this issue, should be viewed with some reservation. See, e.g., Maurer, Common Law Defamation and the Fair Credit Reporting Act, 72 Geo.L.J. 95 (1983); McNulty, The Gertz Fault Standard and the Common Law of Defamation: An Argument for Predictability of Result and Certainty of Expectation, 35 Drake L.Rev. 51 (1985-86); Watkins & Schwartz, Gertz and the Common Law of Defamation: Of Fault, Nonmedia Defendants, and Conditional Privileges, 15 Texas Tech L.Rev. 823 (1984); Annotation, Sufficiency of Showing of Malice of Lack of Reasonable Care to Support Credit Agency’s Liability for Circulating Inaccurate Credit Report, 40 A.L.R.3d 1049, 1052 (1971); Annotation, Libel and Slander: Report of Mercantile Agency as Privileged, 30 A.L.R.2d 776 (1953).
. There is an established body of common law authority in Colorado which generally defines abuse of privilege in the usual common law terms of ill will, spite, intent to injure, bad faith, hostility, express malice and similar terms based on improper motive or state of mind — all categorized by courts and legal scholars as common law malice, or simple malice. See Abrahamsen v. Mountain States Telephone & Tel. Co., 117 Colo. 422, 494 P.2d 1287, 1289 (1972); Coopersmith v. Williams, 171 Colo. 511, 468 P.2d 739, 741 (1970); Ling v. Whittemore, 140 Colo. 247, 343 P.2d 1048, 1052 (1959); Bereman v. Power Pub. Co., 93 Colo. 581, 27 P.2d 749, 751 (1933); Hoover v. Jordan, 27 Colo.App. 515, 150 P. 333, 336 (1915); Denver Pub. Warehouse Co. v. Holloway, 34 Colo. 432, 83 P. 131, 133 (1905); Melcher v. Beeler, 48 Colo. 233, 110 P. 181, 184-85 (1910). See also Prosser on Torts, § 2 at 9-14 (4th ed. 1971); Veeder, The History and Theory of the Law of Defamation (pt. 2), 4 Colum.L.Rev. 33, 35-38 (1904); Hallen, Character of Belief Necessary for the Conditional Privilege in Defamation, 25 Ill.L.Rev. 865 (1931); Developments in the Law — Defamation, 69 Harv. L.Rev. 875, 930 (1956); and authorities cited supra note 11. The district court in this case has previously alluded to the malice requirement. See Williams v. Burns, 463 F.Supp. 1278, 1283 (D.Colo.1979).
It is generally agreed that "malice" (whether "actual malice” used in constitutional issue cases or "simple” or "common law” malice), has become too confusing and has lost its legal and practical usefulness. See Prosser and Keeton on Torts, § 115 at 834 (5th ed. 1984). Colorado now advises against its use in jury instructions. Walker v. Colorado Springs Sun, Inc., 188 Colo. 86, 538 P.2d 450, 459 (1975). Some writers urge abandonment of the common law malice standard due to its emphasis on improper motive. See Watkins & Schwartz, Gertz and the Common Law of Defamation: Of Fault, Nonmedia Defendants, and Conditional Privileges, 15 Tex. Tech L.Rev. 823, 870-75 (1984); Comments, Liability and Damages in Libel and Slander Law, 47 Tenn.L.Rev. 814, 840-42 (1980). See also Kalur, Exploration of the “Outer Limits": The Misdirected Evolution of Reckless Disregard, 61 Den. L.J., 43, 60-65 (1983). However, the great variety of defamation fact situations demands a reasonable flexibility and choice of defamation standards. That includes credit reporting situations. See, e.g., National Apparel Adjustment Council, Inc. v. Dun & Bradstreet, Inc., 42 A.D.2d 58; 345 N.Y.S.2d 40 (1973) (feud with plaintiff resulted in a deliberate attempt to injure business with false credit reports).
Elements of the subjective common law malice standard have been present and approved by this court in recent times in diversity cases involving credit reporting agencies, and other matters. See Lawrence v. Moss, 639 F.2d 634, 638 (10th Cir.), cert. denied, 451 U.S. 1031, 101 S.Ct. 3021, 69 L.Ed.2d 400 (1981); Anderson v. Dun & Bradstreet, Inc., 543 F.2d 732, 739 (10th Cir.1976); Hall v. Hercules, Inc., 494 F.2d 420, 423 (10th Cir.1974); Kansas Elec. Supply Co. v. Dun & Bradstreet, Inc., 448 F.2d 647, 649 (10th Cir.1971), cert. denied, 405 U.S. 1026, 92 S.Ct. 1289, 31 L.Ed.2d 486 (1972).
Reckless disregard has been used as part of the common law malice standard in various cases, and has been separated from the personal and moral aspects of the more usual simple malice definitions. See e.g., ABC Needlecraft Co. v. Dun & Bradstreet, Inc., 245 F.2d 775, 777 (2d Cir. 1957); Dun & Bradstreet, Inc. v. Robinson, 233 Ark. 168, 345 S.W.2d 34 (1961). See also authorities cited supra note 11. It is also adopted as the current standard by the American Law Institute in § 600 of the Restatement, but we do not regard that section, as defined in comment b, as the prevailing view of the law in Colorado regarding privilege extended to credit reports. Prior to Gertz, the American Law Institute equated abuse of privilege with ordinary negligence. See Restatement "Special Note on Conditional Privileges and the Constitutional Requirement of Fault,” pp. 259-61; id. § 580B comments c, d, f, and 1; id. § 593 comment c; id. § 595 comment h (credit agencies); id. § 599 comment d.
Although this abbreviated review suggests that Colorado’s abuse of privilege cases do not require (as the district court implicitly and correctly found) application of a subjective motive or state of mind standard in credit reporting situations, these cases cannot be ignored. The standards which they describe are susceptible of adaptation, and the reckless disregard standard is consistent with their broad interpretation which appears to extend immunity considerably beyond ordinary negligence, or conduct close to that standard. Finally, the considerable comment in the Restatement and elsewhere (see the articles cited above and supra note 11) about the Gertz constitutional requirement of fault eliminating the rationale for conditional privileges probably does not apply here. As mentioned, the Colorado abuse of privilege cases appear to go beyond mere negligence in overcoming strict liability; and, in any event, Greenmoss may have removed the application of the Gertz fault requirement to private plaintiff/private matter cases. See infra note 14 for a discussion of Gertz and other Supreme Court cases.
. We are not called upon to further analyze reckless disregard as it is used in the Colorado statute providing for punitive damages, but additional difficulty is obviously injected by its use in that other (and, in our view, even more stringent) context. It is also not necessary to our analysis to separate definitions depending on the object modified by reckless disregard, i.e., reckless disregard as to truth or falsity, as to consequences, as to rights and feelings, and so on. Cases are not consistent with respect to the role of these modified objects. In this case disregard of truth and of consequences would be indistinguishable from a practical standpoint.
. Those cases prohibit a public official or public figure from recovering damages for a defamatory falsehood unless such person proves with “convincing clarity" that the statement was made with "actual malice” — that is, with knowledge that the statement was false or with reckless disregard of whether it was false or not. Reckless disregard is defined to mean "that the defendant in fact entertained serious doubts as to the truth of his publication." St. Amant v. Thompson, 390 U.S. 727, 731, 88 S.Ct. 1323, 1325, 20 L.Ed.2d 262 (1968). See also Garrison v. Louisiana, 379 U.S. 64, 74, 85 S.Ct. 209, 215, 13 L.Ed.2d 125 (1964) ("high degree of awareness of their probable falsity” required.)
In Gertz, the Supreme Court partially extended the New York Times actual malice standard to defamation cases brought by private plaintiffs. In such cases, so long as fault amounting to at least negligence is established, recovery of actual damages is permitted under applicable state law defamation standards. However, absent a showing of New York Times actual malice, awards of presumed and punitive damages are not permitted. In the recent case of Green-moss, the Supreme Court confirmed that Gertz applies only where matters of public concern or interest are involved. Full discussions of the evolution of constitutional protections in defamation cases are contained in Greenmoss, Walker v. Colorado Springs Sun, Inc., 188 Colo. 86, 538 P.2d 450 (1975), Diversified, and our opinion in Dixson v. Newsweek, Inc., 562 F.2d 626, 628-30 (10th Cir.1977).
. Full constitutional protection includes the higher level "clear and convincing" standard of proof, rather than the "preponderance” standard employed by the district court in this case, without objection from the parties. Dun & Bradstreet does not refer to the point on appeal. We know of no case where the court has attempted to compromise on the extent of protection provided by employing the full "actual malice” constitutional standard but reducing the quantum of proof standard.
. In Dominguez the court relied principally on DiLeo v. Koltnow, 200 Colo. 119, 613 P.2d 318 (1980). The DiLeo case was analyzed by the Colorado Supreme Court in Diversified, 653 P.2d at 1107. By initiating contact with newspapers and reports with respect to his grievance, DiLeo made himself a public figure, and his grievance a matter of public interest.
. The Institute also notes that different rules may apply to libel per quod situations since Gertz did not expressly rule on the constitutional requirement of at least negligence in such cases. See Restatement § 580B comment d. The Supreme Court's opinion in Greenmoss will add still further dimension to the inquiry. In his concurring opinion in Greenmoss Justice White suggests that "the Gertz requirement of some kind of fault on the part of the defendant is also inapplicable in cases such as this.” 105 S.Ct. at 2953. The dissenting opinion takes just the opposite view. Id. at 2957 (Brennan, J., dissenting). The plurality opinion does not discuss the Gertz fault requirement at all. Further confusion is added in libel per quod situations by the American Law Institute’s suggestion in comment d to § 580B of the Restatement that there may be a constitutional requirement of recklessness in such cases, followed by the Institute’s suggestion that regular principles of negligence would be appropriate to the same degree that the negligence standard required by Gertz applies to the issue of truth or falsity.
. Walker involved a libel suit brought by owners of an antique shop against a newspaper, its editor and publisher, and one of its reporters, for three articles and an editorial concerning an alleged purchase of stolen goods by the plaintiffs. The Colorado Supreme Court acknowledged that the plaintiffs were neither public officials nor public figures, but declared the matter to be one of public or general concern entitled, therefore, to the constitutional protection of the “actual malice" standard developed by the United States Supreme Court. There were, however, two significant differences in the application of that standard. Although the Supreme Court in Gertz declared that the actual malice standard was required only for the imposition of presumed or punitive damages in libel cases involving private plaintiffs, the Colorado Supreme Court chose to apply the actual malice standard at the threshold of liability in the same fashion as cases involving public officials or public figures. The second difference was a decision by the Colorado Supreme Court to define "reckless disregard" as used in the "actual malice” standard in terms of existing Colorado tort law definitions rather than the "serious doubts" definition employed by the United States Supreme Court. The protections extended to the defendant in Walker were based upon the Colorado Supreme Court’s interpretation of Article II, Section 10 of the Colorado Constitution. Proceeding on the same basis, the Colorado Supreme Court overruled Walker in Diversified and extended, at the threshold stage, the full constitutional “actual malice” standard, including the "serious doubts" definition of reckless disregard to the publication of all matters of public or general concern, whether or not the plaintiffs are private or public officials or figures.
. There is no doubt that Colorado is free to fashion any definition of reckless disregard which it may deem appropriate for these types of cases. Gertz and Greenmoss make clear there is no constitutional impediment. Chief Justice Burger suggests that an ordinary negligence definition could be employed. Greenmoss, 105 S.Ct. at 2948 (Burger, C.J. concurring). Furthermore, in Burns v. McGraw-Hill, 659 P.2d at 1360, the Colorado Supreme Court suggested an intention not to be confined to any single standard but to fit an appropriate standard to each type of case. However, in adopting the reckless disregard standard in Walker the Colorado Supreme Court specifically referred to existing tort definitions of the term and showed no inclination to make a change. Doubtless we have not described the end of the matter with respect to Colorado law. Constitutional law developments, Fair Credit Reporting Acts, and cases *531from other jurisdictions applying standards of care from negligence, see, e.g., Hood v. Dun & Bradstreet, Inc., 486 F.2d 25, 31-32 (5th Cir. 1973), cert. denied, 415 U.S. 985, 94 S.Ct. 1580, 39 L.E.2d 882 (1974), to New York Times "actual malice,” see, e.g., Dun & Bradstreet, Inc. v. O’Neil, 456 S.W.2d 896 (Tex.1970), will likely result in the development of a specific standard for credit agency cases by the Colorado Supreme Court. See the discussion of various standards with respect to credit agencies, Restatement § 595 comment h. See also the extended discussions with respect to conditional privilege and abuse of privilege, Restatement § 580B comments c, d, f, and 1; id. § 593 comment e; id. § 599 comment d; id., "Special Note on Conditional Privileges and the Constitutional Requirement of Fault,” pp. 259-61. Any such standard at or close to ordinary negligence will likely not be framed in terms of reckless disregard since an attempt to force yet another definition onto that term will render it as confusing and meaningless as “malice” or "actual malice” have become. For an analysis of multiple approaches to the subject, see the articles and authorities cited supra notes 11 and 12.
. Reckless disregard and gross negligence are not necessarily the same. See Restatement § 282 comment e, and § 500-503. Some cases view gross negligence as the same or something less than recklessness. See e.g., Fidelity Leasing Corp. v. Dun & Bradstreet, Inc., 494 F.Supp. 786, 790 (E.D.Pa.1980); Baltensperger v. United States, 174 F.Supp. 601, 605 (D.Neb.1959); Smith v. Stepp, 257 N.C. 422, 125 S.E.2d 903, 905 (1962). Other cases equate gross negligence and recklessness. See e.g., Hong Kong Export Credit Ins. Corp. v. Dun & Bradstreet, Inc., 414 F.Supp. 153, 160 (S.D.N.Y.1975); Burk Royalty Co. v. Walls, 616 S.W.2d 911, 920 (Tex.1981); Fassett v. Santiam Loggers, Inc., 267 Or. 505, 517 P.2d 1059, 1060 (1973). See ako Anderson and Pagliuca, The Colorado Supreme Court’s Developing Defamation Guidelines: Colorado Enters the Quagmire, 59 Den.LJ. 627 (1982).
. We are also aware that conditional privilege in this case gives Dun & Bradstreet more protection than would be accorded a newspaper which published the same material where no public issue is concerned, or a recipient of a defamatory credit report who republished the defamatory material. As we have pointed out the issue of privilege is not before us. However, the context in which a subscriber requests and receives credit reports is qualitatively different than the unprivileged examples mentioned.
. In general, common law damage remedies in a defamation case potentially include: (1) compensatory damages, "which may be either (a) general or (b) special; (2) punitive or exemplary; and (3) nominal." Prosser and Keeton on Torts § 116A at 842-48 (5th ed. 1984). See Restatement §§ 621, 622, 622A, 623. See generally O’Cana v. Espinosa, 141 Colo. 371, 347 P.2d 1118, 1119 (1960).
. The district court’s instruction to the jury was as follows:
If you are satisfied that the reports would tend to harm their reputation or tend to deter others from dealing with them, then damages are presumed as a matter of law. A defamatory statement rarely will affect business profits in such a way that the loss can be traced directly to the statements.
R. Vol. XI at 721.
. Further references to the opinion in Greenmoss are to the plurality opinion of Justice Powell.
. Other courts, in varying contexts, have consistently held that credit reports do not involve matters of public concern or are not protected speech under the First Amendment. See Millstone v. O’Hanlon Reports, Inc., 528 F.2d 829, 833 (8th Cir.1976); Hood v. Dun & Bradstreet, Inc., 486 F.2d 25, 29 (5th Cir.1973), cert. denied, 415 U.S. 985, 94 S.Ct. 1580, 39 L.Ed.2d 882 (1974); Grove v. Dun & Bradstreet, Inc., 438 F.2d 433, 436-37 (3rd Cir.), cert. denied, 404 U.S. 898, 92 S.Ct. 204, 30 L.Ed.2d 175 (1971); Wortham v. Dun & Bradstreet, Inc., 399 F.Supp. 633, 638-41 (S.D.Tex.1975), aff’d, 537 F.2d 1142 (5th Cir.1976); Roemer v. Retail Credit Co., 44 Cal. App.3d 926, 934, 119 Cal.Rptr. 82, 87 (1975).
. See supra discussion Section II.
. The emphasis in the Colorado cases, and in the proceedings below, on the public speech values to be protected also overrides any argument for a remand to consider the issue without regard to the media. It is significant that no such remand was deemed necessary in Green-moss, although the Vermont courts placed heavy emphasis on the media/non-media criteria.
. Section 569 of the Restatement provides as follows:
§ 569. Liability Without Proof of Special Harm — Libel
One who falsely publishes matter defamatory of another in such a manner as to make the publication a libel is subject to liability to the other although no special harm results from the publication.
. The obligatory references setting forth differing views of the state of the law on this subject are: Eldredge, The Spurious Rule of Libel Per Quod, 79 Harv.L.Rev. 733 (1966); Prosser, More Libel Per Quod, 79 Harv.L.Rev. 1629 (1966); Prosser, Libel Per Quod, 46 Va.L.Rev. 839 (1960). See Veeder, The History and Theory of the Law of Defamation (pt. 1), 3 Colum.L.Rev. 546 (1903), (pt. 2), 4 Colum.L.Rev. 33 (1904); Comments, Liability and Damages in Libel and Slander Law, 47 TennX.Rev. 814, 819-22 (1980); Restatement § 568 comment b.
. In Walker v. Colorado Springs Sun, Inc., 188 Colo. 86, 538 P.2d 450, 460 (1975), the Colorado Supreme Court declined an invitation to abolish the distinction between libel per se and libel per quod.
. Other Colorado cases include Cinquanta v. Burdett, 154 Colo. 37, 388 P.2d 779, 780-81 (1963) (imputation of "insolvency or mercantile dishonesty”); Kendall v. Lively, 94 Colo. 483, 31 P.2d 343, 344 (1934) (disparaging statements concerning products); Kobey v. Eddy, 21 Colo. App. 140, 121 P. 948 (1912) (“calculated [attempt] to injure the reputation and credit”); Dorr v. C.B. Johnson, Inc., 660 P.2d 517 (Colo. App.1983) (statements to the effect that employees engaged in reprehensible conduct inconsistent with job performance).
. Simultaneous reliance by the district court on § 573 of the Restatement (relating to slander) and Wegner, which selected a definition from § 569 of the Restatement (relating to libel), draws attention to the disputed position of the American Law Institute on this subject described earlier.
. In Avins v. White, 627 F.2d 637, 646 (3d Cir.), cert. denied, 449 U.S. 982, 101 S.Ct. 398, 66 L.Ed.2d 244 (1980), a general defamation verdict was reversed because the court found it impossible to know which of several alleged incidents the jury relied upon, and only one was considered actionable.
. In Asbill v. Housing Authority of Choctaw Nation, 726 F.2d 1499 (10th Cir.1984), we point*538ed out the difficulties of reviewing general verdicts when multiple claims are involved, and urged the use of special verdicts or interrogatoríes in those and similar situations. This case is one of those situations. See authorities cited in Asbill, id. at 1504 n. 11.
. As previously explained, Sunward was on a fiscal year basis ending February 28. MBMA *540figures were on a calendar year basis.
. There was no real average or history at all. For the fiscal year ending February 28, 1978, Sunward’s sales were only 65.3% of the MBMA. The following two years the percentages were 82.7 and 80.2, respectively, which hardly constitutes an adequate base from which to make projections in any event. See Pl.Ex. 66.
. The lower lost profit figure first presented to the jury by Mr. Clark was $7,694,976, later adjusted at trial to $7,637,119. R. Vol. VII at 370. The jury award of $3,847,000 is one-half of the first figure, rounded to the nearest thousand.
. It is, of course, permissible for the parties to introduce a broad range of facts about Sun-ward’s business, its operations, personnel, problems and successes, as background information.