MEMORANDUM RE: DISCOVERY “FENCE” AND COST ALLOCATION
BAYLSON, District Judge.I. Introduction
Substantial disputes about the scope of discovery, and sharing the costs of discovery, before a determination is made whether this case should be certified as a class action, have divided the parties. The disputes are substantial and the cost of the discovery requested by Plaintiffs is considerable. This issue, where Defendant’s financial exposure will drastically increase if a class is certified, appears to be one of first impression.
Plaintiffs in these cases signed contracts to become members of a health/fitness organization and allege that they thereafter encountered deception and breaches concerning their desire to terminate their membership.1 Plaintiffs filed a consolidated class action Complaint on October 18, 2011 (ECF No. 35). The claims contained in the most recent Complaint are as follows:
1. All plaintiffs — Kenneth J. Silver, Joshua Vaughn, Lori Bohn, Sharon N. Lockett, and Justin P. Bronzell — claiming for breach of contract.
2. Plaintiff Joshua Vaughn, claiming for violations of the Florida Deceptive and Unfair Trade Practices Act.
3. Plaintiff Lori Bohn, claiming for violations of the Washington consumer Protection Act.
II. Summary of the Case Management Conferences
As a result of various filings by the parties, the Court ascertained that this case warranted active discovery management. The Court has consistently ruled that the parties must focus on appropriate discovery so that they could proceed to a hearing, following Plaintiffs’ anticipated Motion for Class Determination, as envisioned by In re Hydrogen Peroxide Antitrust Litig., 552 F.3d 305 (3d Cir. 2008).2 Four separate discovery conferences have taken place since January 2012, as follows:
1. January 31, 2012 (Order of February 1, 2012 (ECF No. 47))' — The parties had been unable to agree on a discovery plan, and the Court authorized the initiation of discovery on both class action and merits, and directed that discovery should focus on the claims of the named plaintiffs, who are proposed class representatives, and merits issues that may be relevant as part of the class action determination. In addition, the Court directed *333that discovery should be initiated and completed promptly with respect to Defendant’s corporate documents that set forth policies and practices that apply to the issues in the case and Plaintiffs’ claims. The Court allowed depositions of corporate officers and employees after documents were produced, and set some deadlines, which proved to be unduly optimistic and were later vacated.
2. April 25, 2012 (Order of April 26, 2012, (ECF No. 51)) — The Court made a number of suggestions about having cost effective discovery focusing on class action issues, but also considering merits to the extent they may be relevant on the class action certification. Following this conference, the parties agreed on a Stipulated Protective Order protecting confidential information. (ECF No. 48)
3. May 8, 2012 (Order of May 10, 2012 (ECF No. 57)) — The filings of the parties, including a Motion to Compel of May 4, 2012 by Plaintiffs (ECF No. 52), revealed significant differences between the parties and a further discovery conference was held. The Court considered the arguments of the parties concerning discovery, but did not make any specific rulings, noting that Defendant was continuing production of papers, documents and electronically-stored information (“ESP’). Plaintiffs had also noticed Rule 30(b)(6) depositions, concerning the contents and location of Defendant’s ESI and paper documents.
4. May 22, 2012 — A further evidentiary hearing regarding discovery disputes was held on May 22, 2012, which is described in more detail below. Another conference was scheduled for June 28, 2012. However, the parties notified the Court by letter dated May 29, 2012, that they had reached agreement concerning their discovery disputes; accordingly, the Court entered an Order on May 30, 2012 (ECF No. 61), denying Plaintiffs’ Motion to Compel as moot.
A pretrial conference had been set for June 28, 2012, but counsel requested it be postponed until the end of July.
Despite the discovery conferences and the parties’ having stated that they had reached agreement, Plaintiffs submitted a letter to the Court on July 31, 2012 reporting that many of the same issues raised in the Motion to Compel of May 4, 2012 remained unresolved. Also on July 31, 2012, Defendant denied the allegations in Plaintiffs’ letter and requested time to submit a more complete response, which Defendant submitted on August 7, 2012 (ECF # 66). Plaintiffs filed a reply on August 8, 2012 (ECF #67). The dialogue back and forth between counsel is similar to a Verdian opera scene where a tenor and a bass boast of their qualities to compete to win over the fair princess.
III. Creating a Discovery “Fence”
Discovery need not be perfect, but discovery must be fair. In determining the boundaries of appropriate discovery in any case where the scope of discovery is subject to disputes, I have found it useful to adopt, as both a metaphor and a guide to determine what discovery is appropriate, a “discovery fence.” The facts that are within the discovery fence are discoverable, and relevant materials should be produced; the facts that are outside the fence are not discoverable and documents or information need not be produced in discovery.
There are two other consequences of adopting a discovery fence. First, the “fence” itself must be a “flexible fence.” A judge should always be willing to reexamine the contours of the fence depending on new facts that are uncovered, unforeseen discovery expenses, or the judge’s changing perception about what is fair. The “fence” can bulge or contract as case-specific circumstances require. Counsel should not hesitate in bringing to the Court new facts warranting a change in the fence.
Secondly, not all fact gathering must come from the opposing party. Each party can and should always conduct its own investigation of matters inside and outside the “fence,” the results of which may warrant a change to the fence boundaries.
IV. Discovery in a Putative Class Action — Economic Aspects
As the Third Circuit noted in Hydrogen Peroxide, the concept of treating a civil action as a class action dramatically changes *334the strategies and economic considerations of the parties and their counsel. In an ordinary civil action, involving named parties against named defendants, there is usually a well-defined range of economic consequences. However, a class action, if allowed by the Court, dramatically increases the economic pressure on the defendant.
In a class action, particularly when damages are sought under Rule 23(b)(3), the defendant must defend against the class as the Court has defined it. If there is a class in this case, it will most likely number in the thousands or tens of thousands people — anyone who joined an LA Fitness Club and later cancelled or sought to cancel their membership. Thus, instead of facing the five named plaintiffs, bringing the case at the present time, Defendant will face exposure to a multitude. The damages sought, if the case goes to trial, would be more difficult to estimate and could be many millions of dollars.
Much has been written in the legal literature about the effect of the 1966 revisions to Fed.R.Civ.P. 23 in this context. The thrust of recent appellate holdings on class actions has been to put significant limits on their scope. Any observer of class action jurisprudence over the last fifty years knows that courts have become much more exacting and demanding that class certification will be fair to a defendant. Compare Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 94 S.Ct. 2140, 40 L.Ed.2d 732 (1974), with, Wal-Mart Stores, Inc. v. Dukes, — U.S.-, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011). Courts must also recognize that there have been many instances, as a result of securities fraud, product defects, or conspiracies that involve price-fixing under the antitrust laws, e.g., where many thousands of people have been significantly damaged. In these eases, class actions have proven an efficient method of transferring unlawfully obtained wealth from the offenders to the victims with generous fee awards to class counsel.
V. Asymmetrical Discovery
Another important aspect of this case is that the discovery is asymmetrical. In other words, Plaintiffs, consumers who were members of LA Fitness for brief periods and allege that they had difficulty cancelling their memberships, number only five. If the class action is denied, perhaps another small group of individuals will intervene to join them as named parties. Even in aggregate, these individuals have very few documents. Perhaps they have kept a copy of their membership contract, or copies of correspondence with LA Fitness.
On the other hand, Defendant LA Fitness has millions of documents and millions of items of electronically stored information (“ESI”). Thus, the cost of production of these documents is a significant factor in the defense of the litigation.
Since Plaintiffs have authorized their counsel to seek a class action, the scope of discovery from Defendant is greatly enlarged. To some extent, discovery is necessary for the parties to advocate, and for the Court to determine, whether a class is appropriate under Rule 23. The scope of discovery, on the merits, will be even further enlarged. Virtually all of this discovery will be directed to Defendant.
Both sides legitimately want to limit their own costs of pretrial discovery. Defendant wants to limit its production of information as much as possible under the applicable rules, in part to save costs, but also, candidly, to provide Plaintiffs with as little material as possible from which to find evidence for use at trial (or in settlement negotiations).
Plaintiffs naturally want to limit their own costs of discovery, and will not voluntarily assume responsibility for the costs of Defendant producing Defendant’s documents.
Professor Coffee wrote about problems posed by asymmetrical discovery in 1987, as follows:
Asymmetric litigation stakes necessitate that the plaintiffs’ attorneys, as independent entrepreneurs, minimize costs, while defense attorneys can persuade their own clients to litigate according to a more luxurious style. Closely related to this first scenario is a second: defendants may exploit asymmetric stakes by increasing the ante while the plaintiffs respond by engaging in low cost, ‘feigned’ litigation that can give the action the appearance of being *335frivolous. Even if defendants cannot force plaintiffs to match them dollar for dollar, they still may be able to demonstrate that protracted litigation is unprofitable for the plaintiffs because the eventual fee recovery will neither cover the plaintiffs’ attorneys’ opportunity costs nor compensate them for the risk associated with their contingent fee. A third possibility is that plaintiffs’ attorneys can diversify their risks by managing a portfolio of individual actions; their tactics may then make sense if their hope is either to identify from among this portfolio a risk averse defendant who will settle generously or to discover a ‘smoking gun’ that changes the litigation odds in plaintiffs’ favor after only a limited search. Finally, there remains the original possibility that a litigation cost differential may sometimes enable plaintiffs’ attorneys to engage in practices that resemble extortion.
At bottom, the tradeoff between the asymmetric stakes and cost differential effects is indeterminate.
John C. Coffee, The Regulation of Entrepreneurial Litigation: Balancing Fairness and Efficiency in the Large Class Action, 54 U. Chi. L. Rev. 877, 893 (1987).
This Court does not in any way suggest that counsel is acting otherwise than in the interests of their clients, but economic motivation and fairness are relevant factors in determining cost shifting of disputed discovery burdens.
The fact that Defendant has more documents than Plaintiffs does not necessarily mean that Defendant’s production should be limited. However, as in this case, where the cost of producing documents is very significant, the Court has the power to allocate the cost of discovery, and doing so is fair. If Plaintiffs’ counsel has confidence in the merits of its case, they should not object to making an investment in the cost of securing documents from Defendant and sharing costs with Defendant.3
In this case Plaintiffs are represented by the very successful and well regarded Philadelphia firm of Berger & Montague, which has had outstanding successes for many years in prosecuting class actions, winning hundreds of millions of dollars for their clients, and undoubtedly and deservedly, substantial fees for themselves. If the Berger & Montague firm believes that this case is meritorious, it has the financial ability to make the investment in discovery, to the extent the Court finds that cost sharing is otherwise appropriate.4
VI. Allocating the Cost of Discovery Before a Decision on Class Certification.
The rules for discovery, and some case law, have long allowed a trial judge to shift the cost of pretrial discovery. In Oppenheimer Fund, Inc. v. Sanders, the Supreme Court stated:
[u]nder th[e] [discovery] rules, the presumption is that the responding party must bear the expense of complying with discovery requests, but he may invoke the district court’s discretion under Rule 26(c) to grant orders protecting him from ‘undue burden or expense’ in doing so, including orders conditioning discovery on the requesting party’s payment of the costs of discovery.
437 U.S. 340, 358, 98 S.Ct. 2380, 57 L.Ed.2d 253 (1978) (emphasis added).5
Despite this broad authorization of cost shifting by the Supreme Court over thirty *336years ago, very few Courts took advantage of this authority before the advent of ESI. Indeed, the impact of ESI on the issue of cost shifting is clear from the civil discovery rules themselves — their only specific mention of cost shifting is found in Rule 26(b)(2)(B), titled “Specific Limitations on [ESI]”:
*335[T]he court must limit the frequency or extent of discovery otherwise allowed by these rules or by local rules if it determines that ... the burden or expense of the proposed discovery outweighs its likely benefit, considering the needs of the case, the amount in controversy, the parties' resources, the importance of the issues at stake in the action, and the importance of the discovery in resolving the issues.
*336A party need not provide discovery of electronically stored information from sources that the party identifies as not reasonably accessible because of undue burden or cost. On motion to compel discovery or for a protective order, the party from whom discovery is sought must show that the information is not reasonably accessible because of undue burden or cost. If that showing is made, the court may nonetheless order discovery from such sources if the requesting party shows good cause, considering the limitations of Rule 26(b)(2)(C). The court may specify conditions for the discovery.
Although by its terms this rule is limited to ESI, courts are permitted to apply similar cost allocations in cases involving extensive production of paper discovery.
Many reported decisions apply standards for requiring that the requesting party share discovery costs. Recently, most of these cases involve ESI. The leading opinion is undoubtedly Judge Scheindlin’s Zubulake v. UBS Warburg LLC (“Zubulake I”), 217 F.R.D. 309 (S.D.N.Y.2003), where she enunciated a seven-factor test that should govern decisions about cost allocation. Those factors are:
1. The degree to which the request for information is designed to discover germane information;
2. The availability of the same information from different sources;
3. The cost of production as compared to the amount in controversy;
4. The cost of production as compared to the resources of each party;
5. The parties’ relative abilities to control discovery costs and their incentives to control costs;
6. The degree of importance of the issues being decided in the litigation; and
7. The relative benefits to each of the parties in obtaining the information at issue.
Id. at 322.
Shifting the cost burdens of discovery, both for ESI and paper discovery, is no longer rare. Still, no decision has been found specifically addressed to allocation of costs as part of a substantial discovery dispute prior to the class certification decision. In this case, at least some of the discovery is relevant to whether a class should be certified, particularly given the searching inquiry required by Hydrogen Peroxide.
One instructive case involving ESI is Clean Harbors Envtl. Servs., Inc. v. ESIS, Inc., No. 09 C 3789, 2011 WL 1897213 (N.D.Ill. May 17, 2011) (Cox, Mag. J.). In this ease, the plaintiff filed a motion, after completion of discovery, seeking to shift to the defendants approximately $91,000 in discovery costs that resulted from the plaintiff having produced ESI stored on backup tapes. In order to produce the ESI, the plaintiff needed to physically pull backup tapes from an offsite facility, load those tapes onto its system to extract certain data, and then retain a third party vendor to process, search, filter, or otherwise access 166 gigabytes of data. Id. at *2. Among other things, the court explained that in cases involving cost shifting, an important consideration is whether the records in question are stored in an accessible or inaccessible format — the process required to access the plaintiffs ESI stored on backups tapes was sufficiently cumbersome to render it inaccessible. Id. at *2 (citing Zubulake v. UBS Warburg LLC (“Zubulake II”), 216 F.R.D. 280, 291 (S.D.N.Y.2003)). For this and other reasons — including (1) the parties’ similar financial footings, (2) the unquestionable relevance of the discovery sought, (3) that the discovery was needed by all parties, and (4) that neither the plaintiff nor the defendants had failed to control costs — the court deemed cost shifting appropriate, ordering that the plaintiff and the defendants share the discovery costs equally. Id. at *3-5.
Several recent cases have applied cost-sharing to non-ESI discovery. In Simms v. Ctr. for Corr. Health & Policy Studies, 272 *337F.R.D. 36 (D.D.C.2011), Chief Judge Royce C. Lamberth endorsed shifting a plaintiffs discovery costs to the defendant, albeit in a minor amount. Citing Oppenheimer Fund, Inc., 437 U.S. at 358, 98 S.Ct. 2380; Peskoff v. Faber, 251 F.R.D. 59, 61 (D.D.C.2008), and Fed.R.Civ.P. 34(a) advisory committee’s note, Judge Lamberth held that because the plaintiff established that she lacked the financial resources to pay $300 in copying costs necessary to comply with the defendant’s discovery requests, she had overcome the presumption that she was responsible for those costs and satisfied her discovery burden by making the requested documents available for inspection and copying at the defendant’s expense. Id. at 40-41.
Am. Int’l Specialty Lines Ins. Co. v. NWI-I, Inc., 240 F.R.D. 401 (N.D.Ill.2007) (Mason, Mag. J.), involved a motion by the plaintiff to compel the defendants to review for relevance more than 19,000 boxes of paper documents. The discovery scenario facing the court was that (1) the defendants had limited financial resources, (2) the dispute covered only a narrow set of issues for which broad discovery would not ordinarily be necessary, and (3) any effective review would necessarily be overbroad because targeting likely relevant documents was impossible, as the only insight into the contents of the documents was a demonstrably unhelpful index.6 Id. at 409-410.
The court ultimately denied the plaintiffs motion to compel the defendants to review all the documents, holding instead that it was appropriate to require the plaintiff and the defendants to share the review costs equally:
After considering the foregoing factors and the particular circumstances present here, this Court finds that it would be unduly burdensome and expensive to require the [defendants] to review all of the ... documents in order to produce only those documents that are responsive to plaintiffs requests. Accordingly, plaintiffs request for such an order is denied. Nevertheless, this Court finds that plaintiff is entitled to discover whether the ... documents contain relevant information. Furthermore, we find that it would be equally burdensome, expensive and unfair if we permitted defendants to rest on their “keys to the warehouse” method of responding to plaintiffs discovery requests. We cannot allow defendants to merely offer plaintiff access to the ... documents because doing so shifts the undue burden entirely to the plaintiff. Therefore, in light of the resources of the parties, the limited issues in this case and the fact that there are more than 19,000 boxes of documents in storage, we find it appropriate to place some limits on discovery related to the ... documents.
Based on the foregoing, this Court orders the parties to work together to create a discovery plan with respect to the ... documents. The parties must take into consideration the following: (1) from this point forward, all fees and costs associated with discovery related to the ... documents will be shared equally between plaintiff and the defendants; (2) the 1,778 boxes with no labels or labels that do not sufficiently describe their contents should be reviewed to determine whether the boxes contain any relevant information; and (3) the parties should agree upon what additional discovery related to the ... documents is necessary, if any, and how long it will take to complete that discovery. By requiring the parties to share the cost, this Court encourages the parties to come up with a discovery plan that is both time and cost efficient.
Id. at 412-13 (citation and footnotes omitted).
In Foreclosure Mgmt. Co. v. Asset Mgmt. Holdings, LLC, Civil Action No. 07-2388-DJW, 2008 WL 3822773 (D.Kan. Aug. 13, 2008), Asset Management Holdings, LLC (“Asset Management”) sought an order compelling Foreclosure Management Co. (“FMC”) to produce certain documents. FMC argued, among other things, that production presented an undue burden and, therefore, that production either should not be ordered at all, or, if ordered, be accompanied by a requirement that Asset Manage*338ment share costs. Responding to FMC’s arguments, the court stated:
The Court finds that FMC has failed to adequately support its undue burden objection. As a preliminary matter, FMC fails to provide the Court with any evidence showing the expenditure of time, effort or money that would be necessary to produce the requested documents. While FMC has provided an “estimate” that it would likely take 2,000 hours to search its storage facilities, locate the appropriate files, and search the invoices, this explanation is not detailed enough to convince the Court that such a search would result in the expenditure of great labor, great expense, or considerable hardship to FMC. Even assuming arguendo that FMC had submitted evidence to demonstrate such a burden or hardship, FMC has failed to establish that the burden or hardship would be undue and disproportionate to the benefits Asset Management would gain from the document production. As discussed above, the Court finds the requested [documents] to be directly relevant and important to establishing Asset Management’s defenses and counterclaims. Thus, the Court declines to deny the Motion to Compel based on undue burden.
The Court will now consider FMC’s request that the cost of producing these documents be shifted to Asset Management. Under the discovery rules, the presumption is that the responding party must bear the expense of complying with discovery requests. Notwithstanding this presumption, a court has discretion under Rule 26(c) to condition discovery on the requesting party’s payment of the costs of the discovery. In making such a determination, courts often consider the factors set forth in Rule 26(b)(2)(C)(iii), which are whether “the burden or expense of the proposed discovery outweighs its likely benefit, considering the needs of the ease, the amount in controversy, the parties’ resources, the importance of the issues at stake in the action, and the importance of the discovery in resolving the issues.” Taking into consideration these factors, the Court finds that it would be appropriate to apportion the costs of this discovery equally between FMC and Asset Management. The parties are directed to confer to determine the most cost efficient method for production of these documents.
Id. at *7 (quotations and citations omitted).
In Schweinfurth v. Motorola, Inc., No. 1:05CV0024, 2008 WL 4449081 (N.D.Ohio Sept. 30, 2008), a class action, the court ordered the plaintiffs to pay half of the discovery costs because the plaintiffs’ requested discovery, though relevant, (1) was huge in size and scope, (2) did not pertain to the defendant’s products that were the subject of the named plaintiffs’ claims, and (3) would not necessarily provide material admissible at trial.
Cost shifting was also recently applied to discovery of ESI in this jurisdiction in Universal Del., Inc. v. Comdata Corp., Civil Action No. 07-1078, 2010 WL 1381225 (E.D.Pa. Mar. 31, 2010) (Perkin, Mag. J.).
This Court has applied cost allocation in at least one other case. In IP Venture, Inc. v. Sony Elees., No. 09-497 (D.Del.), the plaintiff was an investor in lawsuits asserting that various patents had been improperly obtained and should be declared invalid. The plaintiff itself manufactured nothing and had no business other than its investments in these lawsuits. The defendant Sony, as most would recognize, is a multi-billion dollar, mul-ti-national corporation with thousands of inventions and millions of documents. Because of the asymmetrical discovery in the case, the Court informed the parties that it was likely to impose on the plaintiff at least some portion of the expenses associated with its extensive discovery requests. Order of June 21, 2010 (ECF No. 48). Perhaps lacking some confidence in either its economic ability to sustain this burden or in the merits of the claim, the plaintiff promptly settled the case.7
*339VII. Plaintiffs’ Motion to Compel filed May 4, 2012.
In this motion, Plaintiffs sought production of documents in two ESI categories:
1. “Member notes” (the LA Fitness term for electronic data reflecting internal discussions and actions pertaining to LA members) relating to customer complaints which were generated in five states over the course of a sampling of 60 months designated by Plaintiffs, and
2. E-mails to and from five individuals who created or implemented Defendant’s cancellation procedures and from two individuals who Defendant identified as having direct responsibility over handling cancellations (who were identified by name).
Plaintiffs had also asserted that Defendant should bear all costs associated with Plaintiffs review of Defendant’s paper documents located at the Iron Mountain facility in California. Defendant had previously advised that paper documents relating to membership issues were stored at Iron Mountain. A “litigation hold” had been put on these documents so that they were not being destroyed, as they would have been in the absence of this litigation. Plaintiffs suggested that Defendant be required to continue the litigation hold and incur expenses pertaining to facilitating the inspection of the documents at the Iron Mountain facility.
Defendant’s response (EOF #4) details the parties’ various negotiations prior to Plaintiff filing the Motion to Compel. The Court’s notes of the conference held on May 22, 2012, indicate that the parties had reached some agreements. First, certain depositions had already been taken of corporate representatives, which gave Plaintiffs further details about the type of documents that Defendant generated and how they were maintained. Second, Plaintiffs had already inspected some of the documents at the Iron Mountain facility in California. Third, the parties had a tentative agreement that the Members Notes ESI would be reviewed by Defendant for members in five states over a 30 month period, although there were some disputes as to details. Fourth, Defendant would preserve the documents at Iron Mountain.
As the Court recalls, the major issues left open after the May 22, 2012 conference were (1) cost allocation, (2) production of internal memoranda by LA Fitness officers and employees, and (3) certain details concerning the production of ESI.
At the hearing on May 22, 2012, the Court made only one ruling: Plaintiffs would not be required to pay for Defendant’s counsel to review Defendant’s Member Notes prior to production. This is not a case in which there is likely to be any privileged documents in the material requested, because the Member Notes are basically complaints from consumers. Indeed, not only is it highly unlikely that privileged materials will be contained in the Member Notes, but Plaintiffs have agreed to return to Defendant any documents that they come across that are Privileged, and not to use them in the litigation.
VIII. Discovery Disputes Still Pendiny
Plaintiffs’ letter brief of July 31, 2012 lists in detail Plaintiffs’ document requests to *340which Plaintiffs contend that Defendant has not fully complied. Generally, Plaintiffs’ requests fall within the following basic categories:
1. A variety of internal memoranda and correspondence relating to Defendant’s policies and practices regarding cancellations— including whether Defendant required members to use pre-printed forms for cancellations, whether customers were informed of such, and whether customer cancellations were processed at the local club level prior to January 1, 2012.
2. Documents related to Defendant’s responses to consumer complaints referred by the Better Business Bureau and other state consumer protection agencies.
3. Interrogatory responses regarding the identities of Defendant’s managers who were directly involved in handling customers’ cancellation notices and complaints.
4. Interrogatory responses regarding Defendant’s explanations for the reasons that many cancellations notices were not processed.
5. Privilege logs for all documents withheld on grounds of privilege.
Defendant’s initial response details the amount of discovery which it has already provided, apparently at its own expense. Defendant notes that it reviewed thousands of E-mails itself, rather than negotiate a narrowing of the search terms. In total, Defendant asserts that it has reviewed (1) over 500,000 Member Notes from five states for 30 months looking for certain terms, (2) over 1,000 boxes of cancellation requests, of which Plaintiffs reviewed only 70 boxes, (3) over 19,000 pages of documents, and (4) an electronic search of over 32,000 E-mails, maintained by five custodians. Defendant contends that relatively few tidbits of relevant information have been found as a result of this very extensive and expensive search.
Defendant also responds to specific requests and asserts that Plaintiffs’ demands are burdensome and not relevant to the issues in the case, particularly given that Plaintiffs have already secured a great deal of discovery, but the Court has not yet certified a class. Contrary to Plaintiffs, Defendant asserts it has conducted a reasonable search for documents relating to its policies and practices regarding the handling of cancellation notices.
Concerning Plaintiffs’ second request for documents, served after the May 22, 2012 conference, Defendant contends that this new demand for additional documents is totally improper and unduly burdensome. Plaintiffs repeat their request for documents relating to whether Defendant processed member cancellations at the local club level prior to January 1, 2012, and Defendant objects to this request as overbroad and unreasonable, particularly on the class action issue. Defendant’s response raises other disputes as well.
In Defendant’s estimate of costs, filed May 21, 2012 (ECF # 58), Defendant estimated the cost to review 60 months of Members Notes ESI at approximately $360,000. Assuming that the number of months has been reduced to 30, this is still a very elaborate and expensive undertaking, particularly, if as Defendant has represented, a sampling of these Member Notes has exhibited only an extremely small proportion with any evidence probative of Plaintiffs’ claims.
In addition, Defendant asserted that it will cost $219,000 to perform the search and production of E-mails from the active accounts from the seven custodians requested by Plaintiffs.
Lastly, Defendant notes that the Iron Mountain storage costs are $300 per month for over 1,000 boxes of documents, consisting primarily of cancellation notices. Defendant would like to be relieved of this expense by either destroying the documents or having Plaintiffs undertake the cost of maintaining them.
The parties’ latest submissions to the Court, dated July 31, August 7 and August 8, 2012, are to some extent advocacy for their arguments on the merits, and lack complete symmetry in their discussion of Plaintiffs’ requests and Defendant’s objections.
Plaintiffs’ reply dated August 8, 2012, asserts that there was an agreement between the parties on May 25, 2012 which “is in full force and effect and is being carried out.” *341However, the only agreement of which this Court is aware is reflected in a June 26, 2012 letter from counsel that “LA Fitness will produce Member Notes for 30 months across five states as well as E-mails of five custodians.” Notably, Plaintiffs’ opening letter-brief, dated July 31, 2012, and Defendant’s response dated August 7, 2012, make no reference to a May 25,2012 agreement.
Plaintiffs’ August 8, 2012 reply brief concentrates heavily on Defendants’ failure to produce “internal memoranda” and correspondence on various topics. Defendant asserts that large amounts of internal memoranda have already been produced on relevant topics. Plaintiffs, however, contend that they are entitled to “all responsive internal documents.” The Court disagrees with this sweeping characterization of a defendant’s obligations in a case of this nature, prior to class action certification. Concerning the sampling of executives’ e-mails, Defendant also claims bur-densomeness, but the Court believes that these files, as limited by the parties’ agreement, are potentially germane, and are of the type generally produced in commercial litigation. It also appears that this search, as limited by agreement, may have already been completed.
Regarding the Member Notes ESI, Defendant contends that there is a substantial difference between the parties about the appropriate scope of review. This dispute has been the subject of much back-and-forth negotiating between the parties. Defendant makes a convincing showing that the huge volume of boxes of materials stored in Iron Mountain contain disproportionately few relevant documents, and in any event, whatever germane nuggets of information they may contain, Plaintiffs should bear the entire costs of production inspection, and preservation.
The exchange of correspondence between the parties also reflects disagreement regarding what Defendant has or has not been produced. The Court cannot resolve these issues through an exchange of correspondence. The parties should meet with each other in the same conference room for however long is necessary to determine to what extent Defendant has searched for and/or produced documents. If necessary, depositions may be appropriate of additional LA Fitness custodians with knowledge about the existence and location of documents related to the class action issue. The Court notes that three such depositions have already taken place.
IX. The Court’s Rulings on Discovery and Cost Shifting Issues
Based on the legal discussion above and extensive review of the parties positions, the Court mandates cost allocation as fair and appropriate. The Court concludes that where (1) class certification is pending, and (2) the plaintiffs have asked for very extensive discovery, compliance with which will be very expensive, that absent compelling equitable circumstances to the contrary, the plaintiffs should pay for the discovery they seek. If the plaintiffs have confidence in their contention that the Court should certify the class, then the plaintiffs should have no objection to making an investment. Where the burden of discovery expense is almost entirely on the defendant, principally because the plaintiffs seek class certification, then the plaintiffs should share the costs.
Plaintiffs seek to represent a very extensive class, and if, as Plaintiffs anticipate, their class action motion is granted, this case will suddenly turn from a routine ease to a major financial exposure for Defendant. The Hydrogen Peroxide decision and its progeny require that the Court make a very detailed analysis as to whether Plaintiffs can meet their Rule 23 burdens. Plaintiffs have already amassed, mostly at Defendant’s expense, a very large set of documents that may be probative as to the class action issue.
The Court is persuaded, it appearing that Defendant has borne all of the costs of complying with Plaintiffs’ discovery to date, that the cost burdens must now shift to Plaintiffs, if Plaintiffs believe that they need additional discovery. In other words, given the large amount of information Defendant has already provided, Plaintiffs need to assess the value of additional discovery for their class action motion. If Plaintiffs conclude that additional discovery is not only relevant, but important *342to proving that a class should be certified, then Plaintiffs should pay for that additional discovery from this date forward, at least until the class action determination is made.
The Court is firmly of the view that discovery burdens should not force either party to succumb to a settlement that is based on the cost of litigation rather than the merits of the case.
Therefore, the Court will require Plaintiffs forthwith to detail to Defendant the additional documents they have requested, but contend have not yet been produced and are desired before their class action motion is filed. Plaintiffs’ list should be specific as to what searching of ESI, or hard documents, is required. Defendant shall respond to Plaintiffs’ request within fourteen (14) days, summarizing Defendant’s internal costs for providing this information. These costs may include the appropriately allocated salaries of individuals employed by Defendant who participate in supplying the information which Plaintiffs request, including managers, in-house counsel, paralegals, computer technicians and others involved in the retrieval and production of Defendant’s ESI. Within seven (7) days of receiving this information, Plaintiffs shall advise Defendant if they are willing to make the payment requested by Defendant. If so, the requisite amount shall be paid to Defendant promptly and before Defendant initiates production. The Court reserves the right to make an allocation of these costs depending upon the outcome of the class action motion and/or the merits of the case.
The Court will hold a further pretrial conference on September 20, 2012, when the Court will set a date for completion of class action discovery, the filing of a motion for class certification, a briefing schedule, and a hearing date.
If either party requests any depositions before the next pretrial conference, that are reasonably related to the class action discovery issues, those requests shall be honored and each party shall bear its own costs for counsel and transcripts.
Returning to the “discovery fence” metaphor, and in making a determination of what facts are inside and outside the fence, the Court recognizes that Plaintiffs will need to show “predominance” of common issues as an important part of their burden in meeting the Hydrogen Peroxide standard for class certification. In particular, Plaintiffs will need to establish that “ ‘proof of the essential elements of the[ir] causefs] of action’ ” will not require “individual treatment” of each plaintiffs claims. Hydrogen Peroxide, 552 F.3d at 311 (quoting Newton v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 259 F.3d 154, 172 (3d Cir.2001)). As the Third Circuit explained in Sullivan v. DB Invs., Inc., 667 F.3d 273, 299 (3d Cir.2011) (en banc), the Supreme Court’s recent decision in Walr-Mart Stores, Inc. v. Dukes, — U.S. -, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011), requires Plaintiffs to establish that “common questions generate common answers ‘apt to drive the resolution of the litigation’” — in other words, that Plaintiffs show “a common course of conduct in which the defendant engaged with respect to each individual.” (quoting Dukes, 131 S.Ct. at 2551). Thus, under Dukes, the most relevant discovery at this stage of the case is that which will illuminate the extent to which Defendant’s membership cancellation policies and practices are set and followed nationally; Plaintiffs must show either that individual managers have no discretion or that there is a “common mode of exercising discretion that pervades the entire company.” Dukes, 131 S.Ct. at 2553-55.
A. Accordingly, the Court concludes that the following are inside the fence, but with Plaintiffs paying the cost:
1. Corporate documents stating Defendant’s practices and policies applicable to joining LA Fitness and cancellation of memberships.
2. A reasonable sample of Members Notes (but it appears that these have already been produced as part of Defendant’s ESI).
3. Documents reflecting Defendant’s use of preprinted forms, and whether this practice was uniform throughout the United States.
4. Corporate documents stating corporate policy and practice as to how managers or *343employees should respond to requests for cancellations.
5. Corporate documents concerning discretion of local managers or employees to vary from corporate policies.
6. Documents stored at Iron Mountain.
7. Additional ESI related to categories 1-4.
B. The Court further concludes that the following are outside the fence, not discoverable at this stage of the case, even if Plaintiffs offered to pay costs:
1. Any requests characterized by Plaintiffs as a demand for production of “all” documents of a general category (but Defendant should produce “all” documents where the categorization is specific).
2. Requests targeted at determining whether decisions are made at any particular level of managers.
3. Internal memoranda other than as described above.
4. Defendant’s response to consumer complaints, including Chambers of Commerce, Better Business Bureau.
5. Documents relating to how individual cancellation notices are processed.
X. Conclusion
The Court expects communication between counsel to proceed in accordance with the foregoing memorandum, and for counsel to propose a joint, or separate, scheduling proposal before the next pre-trial conference.
An appropriate Order follows.
. After the first case was started in 2010, and the second case was filed in 2011, Defendant filed extensive Rule 12 Motions in both cases, raising a number of issues. In a lengthy Opinion dated September 12, 2011 (ECF No. 33, Boeynaems v. LA Fitness Int'l, LLC, 2011 WL 4048512 (E.D.P.A.2011)) the Court granted, in part, Defendant's Motion to Dismiss and also consolidated the two cases because they make basically similar claims.
. The Court referred counsel to two other cases where the Court held two-day evidentiary hearings on Rule 23 Motions as an indication of how the undersigned interprets the duties of a District Court under Hydrogen Peroxide.
. Although these costs may not necessarily be taxable against Defendant in the event of a successful trial, see Race Tires Am. v. Hoosier Racing Tire Corp., 674 F.3d 158 (3d Cir.2012), this Court would likely allow Plaintiffs’ counsel’s out-of-pocket costs to be deductible from any award to a class, assuming a class was certified.
. In the absence of a putative class action, where a party or counsel is unable to afford assuming the other party's costs of discovery, cost shifting may not be appropriate.
. Rule 26(b)(2)(C)(iii) sets out the standard for what constitutes an "undue burden":
. The plaintiff had established that the index either lacked or had inaccurate descriptions of the contents of at least 2,000 boxes of documents. Id. at 409-10.
. This Court also notes the fairly extensive commentary in the legal literature on cost allocation for discovery, particularly where it is asymmetrical. Articles on the topic include (but are not limited to):
1. Martin H. Redish & Colleen McNamara, Back to the Future: Discovery Cost Allocation and Modem Procedural Theory, 79 Geo. Wash. L.Rev. 773 (2011) (questioning the usual practice of requiring a producing party to *339bear the expenses associated with its opponent’s discovery requests because, inter alia, it creates a litigation subsidy that enriches the requesting party);
2. Laura E. Ellsworth & Robert Pass, Cost Shifting in Electronic Discovery, 5 Sedona Conf. J. 125 (2004) (discussing Fed.R.Civ.P. 26(b)(2) and 34(b), along with significant federal judicial treatment of cost shifting with respect to ESI discovery);
3. Edward Pekarek, The Shifting Tide of ESI Discovery Cost Allocation, 1899 PLI/Corp 405 (2011) (discussing various state and federal provisions concerning ESI discovery cost shifting); and
4. Vlad Vainberg, When Should Discovery Come with a Bill? Assessing Cost Shifting for Electronic Discovery, 158 U. Pa. L.Rev. 1523 (2010) (discussing the weaknesses and strengths of the various multifactor tests for cost shifting and describing recent trends in judicial orders involving cost shifting).
A number of treatises also address cost allocation for discovery, including (but not limited to);
1. Barbara J. Rothstein, Ronald J. Hedges & Elizabeth C. Wiggins, Federal Judicial Center, Managing Discovery of Electronic Information: A Pocket Guide for Judges 17-20 (2d ed. 2012);
2. Richard L. Marcus, Edward F. Sherman & Howard M. Erichson, Complex Litigation: Cases and Materials on Advanced Civil Procedure 510-27 (5th ed. 2010); and
3. Steven S. Gensler, Federal Rules of Civil Procedure: Rules and Commentary 520, 526-28 (2008).