*305Modern class actions are an important innovation in our legal system. They are a rare departure from "a principle of general application in Anglo-American jurisprudence that one is not bound by a judgment in personam in a litigation in which he is not designated as a party or to which he has not been made a party by service of process." Hansberry v. Lee , 311 U.S. 32, 40, 61 S.Ct. 115, 85 L.Ed. 22 (1940). When properly employed, class actions efficiently provide consistent outcomes in resolving large numbers of disputes involving common questions of law or fact.
As with many innovations, however, these benefits come with risks.1 The very features of class actions that provide efficiency and consistency at the same time present serious risks, including the risk of due process violations. For a plaintiff class, absent class members' grievances are litigated not directly by the class members themselves but by class representatives. Defendants in those actions can be sued by a class of hundreds or thousands of people who are relieved of their normal responsibility to file and prosecute their own case. For a defendant class, concerns are arguably even greater with the very real possibility that judgments can be entered against absent defendants without those defendants receiving notice of the underlying action itself. See 1 Joseph M. McLaughlin, McLaughlin on Class Actions § 4:46 (15th ed. 2018).
Rule 23 of the Federal Rules of Civil Procedure is a structural safeguard in our legal system developed to provide the benefits of class actions while simultaneously avoiding the risks associated with them. As a leading commentator on class actions explains, the primary function of Rule 23"is, through its various parts, to ensure the protection of absent class members' rights and, hence, the justification of the binding effect of the resulting judgment." 1 William B. Rubenstein, Newberg on Class Actions § 1:10 (5th ed. 2011). Thus, to appropriately strike this balance of maximizing benefits and minimizing risks, compliance with Rule 23 is necessary.
In this appeal, Objectors challenge a number of the district court's decisions, including several findings regarding the Rule 23 requirements. Addressing those decisions, the majority affirms the district court's denial of attempted class wide opt-outs and dismisses an appeal of an order denying a motion to intervene as untimely. Maj. Op. at 304. The majority reverses the district court's certification and settlement approval decisions, holding that the district court abused its discretion in certifying the Speaks settlement class under Rule 23(a) and in approving the final settlement as fair, reasonable and adequate under Rule 23(e). Maj. Op. at 304. I agree that the district court erred on these issues and, therefore, join with the majority in reversing the district court's order. However, I write separately because, although I agree with the result the majority reaches, I reach those results for different reasons.
I.
I first address the district court's finding regarding the adequacy of the class representatives. Because the district court found the class representatives to be adequate when intra-class conflicts existed regarding the patronage theory of recovery, I join the majority in finding that the district *306court abused its discretion in certifying the Speaks class.
In considering the adequacy of the class representatives, we look to Rule 23(a). That subsection requires adequacy of class representatives-along with numerosity, typicality and commonality-for a class to be certified. Although a district court is rightly afforded wide discretion in its adequacy determination under Rule 23(a), this discretion is not absolute or unlimited. Amchem Products, Inc. v. Windsor , 521 U.S. 591, 625-28, 117 S.Ct. 2231, 138 L.Ed.2d 689 (1997) and its progeny make clear that a district court abuses its discretion when it finds class representatives to be adequate despite the presence of intra-class conflicts. The concerns about intra-class conflicts relate to the same due process concerns identified above. "The premise of a class action is that litigation by representative parties adjudicates the rights of all class members, so basic due process requires that named plaintiffs possess undivided loyalties to absent class members." Broussard v. Meineke Disc. Muffler Shops, Inc., 155 F.3d 331, 338 (4th Cir. 1998).
Decisions from the Supreme Court and this Court provide guidance on the issue of intra-class conflicts. In Amchem , the Supreme Court considered whether a settlement class consisting of victims of asbestos exposure satisfied the requirements of Rule 23. 521 U.S. at 591, 117 S.Ct. 2231. In considering the requirements of Rule 23(a)(4), the Court emphasized the purpose of the adequacy requirement, noting "[t]he adequacy inquiry under Rule 23(a)(4) serves to uncover conflicts of interest between named parties and the class they seek to represent. [A] class representative must be part of the class and possess the same interest and suffer the same injury as the class members." Id . at 625-26, 117 S.Ct. 2231 (citing General Tel. Co. of Sw. v. Falcon , 457 U.S. 147, 157-58 n.13, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982) ) (internal citation omitted) (internal quotation marks omitted). The Court identified one such intra-class conflict, finding that "[i]n significant respects, the interests of those within the single class are not aligned. Most saliently, for the currently injured, the critical goal is generous immediate payments. That goal tugs against the interest of exposure-only plaintiffs in ensuring an ample, inflation-protected fund for the future." Id . at 626, 117 S.Ct. 2231. In affirming the decision to decertify the class in part because of this conflict, the Court concluded "[t]he settling parties, in sum, achieved a global compromise with no structural assurance of fair and adequate representation for the diverse groups and individuals affected." Id . at 627, 117 S.Ct. 2231.
Ortiz v. Fibreboard , 527 U.S. 815, 856-58, 119 S.Ct. 2295, 144 L.Ed.2d 715 (1999) affirmed the principle established by Amchem and further explained its meaning. In yet another asbestos case, the Supreme Court in Ortiz considered whether a mandatory settlement class certified under a limited fund theory satisfied the requirements of Rule 23. Id . at 821, 119 S.Ct. 2295. In considering the fairness of the distribution of the settlement fund among class members, the Supreme Court applied the adequacy principles set out in Amchem and identified two types of conflicts within the class. First, reiterating its holding in Amchem , the Court identified a conflict between class members with present injuries and future injuries. Id . at 856, 119 S.Ct. 2295. The Court noted "it is obvious after Amchem that a class divided between holders of present and future claims ... requires division into homogenous subclasses under Rule 23(c)(4)(B)." Id . Second, the Court identified a conflict between class members related to the value of individual class members' claims. Id. at 857, 119 S.Ct. 2295. Noting that some class *307members necessarily had "more valuable claims" than others, the Court found that this misalignment of interests required subclasses. Id . According to the Court, both types of conflict were "well within the requirement of structural protection recognized in Amchem ." Id .
After Ortiz , it is clear that Amchem extends beyond conflicts between current and future damages. Our Court, a year earlier, recognized this in Broussard v. Meineke Disc. Muffler Shops, Inc., 155 F.3d 331 (4th Cir. 1998). There, this Court followed Amchem in determining a nationwide class of franchise holders was improperly certified. Broussard , 155 F.3d at 337-39. According to this Court, the putative class of franchise holders contained "manifest conflicts of interests," largely driven by class members' diverging interests in the amount and type of recovery. Id . at 338. This Court noted, "The interest of former franchisees in damages and of many EDP franchisees in restitution reveals an obvious initial schism within the putative class regarding the appropriate remedy for [the defendant's] alleged wrongdoing." Id . In describing the prejudice caused by these conflicts of interests, this Court observed that the "three distinct groups have potentially divergent aims" and that the remedy sought by the plaintiffs "is consistent with-and only with-the aims [of one of the groups]." Id . at 339 (internal quotation marks omitted).
Broussard also suggests that intra-class conflicts infringe the due process rights not only of absent plaintiff class members but also of defendants. In that case, the presence of intra-class conflict forced the defendant to respond to inconsistent legal theories and damages claims some of which involved individualized considerations. See id . at 338 ("[O]ne group of [absent plaintiffs] sought to intervene below and appear as amici on appeal, claiming their ongoing business relationship with [the defendant] and their interests in the long-term financial health of the company were imperiled by plaintiff's efforts to wring a large damage award out of defendants."). While the defendant in this case did not raise these issues below or on appeal, in the parallel state proceeding it argued that the differences in arguably a less diverse class prevented class certification.2
Here, the district court concluded that there were no intra-class conflicts in this case that precluded certification or necessitated subclasses under Rule 23(a)(4). In reaching this conclusion, the district court attempted to distinguish Amchem , characterizing the differences in class members as mere differences in damages between class members. Because of that, the district court concluded that the differences in class members did not create diverging interests or a fundamental conflict. The district court noted "[a]ny differences between class members only arise with respect to the amount of damages each member seeks to recover. Such differences do not create adverse or diverging interests." J.A. 4738 (citing In re Pet Food Products Liab. Litig ., 629 F.3d 333, 345-46 (3d Cir. 2010) and In re Ins. Brokerage Antitrust Litig ., 579 F.3d 241, 272-73 (3d Cir. 2009) ).
It is clear from the record that the district court gave thoughtful consideration to this issue. However, the district *308court did not address either Ortiz or Broussard . Moreover, the pertinent issue-the proper scope of Amchem -ultimately is one of law.3 The majority in effect concludes that the district court applied Amchem too narrowly. I agree. In the same way that class certification was improper in Amchem because of differences in the timing of payments, in Ortiz because of the differences in strength of the class members' claims and in Broussard because of inconsistent claims and damages theories and the individualized issues among class members, the patronage theory4 asserted by some class members makes the class certification improper in this case.5
In order to satisfy the requirements of Rule 23(a), Amchem and its progeny require a named class representative to be not only common and typical of those of the entire class, but also to adequately represent the interests of all. When those requirements are not met, the due process rights of the absent class members, as well as defendants, are prejudiced and classes should not be certified.
II.
Turning next to the district court's findings regarding the adequacy of class counsel under Rule 23(g), I agree with the majority that the district court abused its discretion in finding class counsel to be adequate, but for a different reason. In contrast to the majority's approach, if the issue needs to be reached at all, I would simply find that the district court necessarily abused its discretion because it appointed class counsel for a class that was improperly certified due to the presence of intra-class conflicts.
Instead of relying upon the inadequacy of class representatives to conclude that counsel was therefore inadequate, the majority questions the manner in which the district court evaluated the adequacy of class counsel. The majority finds that "the district court improperly rejected the collusion argument based on the mere presence of the mediator" and "did not meaningfully grapple" with the North Carolina court's findings about collusion. Maj. Op. at 292, 293. Respectfully, the record does not support this conclusion, and affording the findings of the district court appropriate deference, I cannot agree that the district court erred in its findings regarding collusion. I write separately on this issue to address my disagreement with the majority's analysis and to illustrate the ways in which the district court meaningfully grappled with the North Carolina court's findings of collusion.6
*309As an initial matter, the district court carefully considered Objectors' arguments that the North Carolina court orders were binding on its own analysis of the adequacy of class counsel. Regarding the North Carolina court's certification order, the district court correctly concluded that the certification of the Fisher-Lewis class did not impact the Speaks action. Likewise, the district court correctly concluded that the North Carolina case management order could not limit the parties or counsel in the Speaks action. In reaching this decision, the district court noted that the Objectors offered no case law in support of their positions. Finally, the district court concluded that the North Carolina court's interlocutory order dated January 17, 2018 was not binding on its own analysis.
These decisions are not surprising since, in advancing these arguments, the Objectors fundamentally mischaracterized the relationship between the two parallel actions. It is well settled that a parallel action has no effect on a competing action until a final judgment is reached. See 3 William B. Rubenstein, Newberg on Class Actions § 10:33 (5th ed. 2011). That is even true for the certification decision. "A common misunderstanding is that certification of a class in one suit somehow forestalls other related cases; in fact, only the preclusive effect of a final judgment can formally achieve that end and a decision simply certifying a class is not itself a final judgment." Id . Simply put, the North Carolina court's interlocutory decisions did not apply to the Speaks case.
Turning now to issues raised in the majority's analysis, the record reveals that the district court did not improperly reject the collusion argument based on the mere presence of the mediator and did meaningfully grapple with the North Carolina court's findings about collusion. Regarding the role of the mediator, the district court admittedly viewed the mediator as a factor in rejecting the collusion arguments. As the court noted, the Speaks settlement was reached after a mediation led by a well-respected North Carolina mediator. The court correctly cited to authority that stands for the proposition that the presence of an unbiased mediator is some evidence against the presence of collusion. See Jones v. Singing River Health Servs. Found ., 865 F.3d 285, 295 (5th Cir. 2017), cert. denied sub nom. Almond v. Singing River Health Sys ., --- U.S. ----, 138 S.Ct. 1000, 200 L.Ed.2d 252 (2018) ("The involvement of an experienced and well-known mediator is also a strong indicator of procedural fairness.") (internal quotation marks omitted).
However, the district court's decision on alleged collusion was based on more than the involvement of the mediator. Despite the district court's conclusion about the legal effect of the North Carolina court *310orders (or lack thereof), the district court nevertheless considered issues they raised concerning the adequacy of class counsel. Specifically, despite the majority's contention to the contrary, the district court considered at length issues raised by the North Carolina court's interlocutory order dated January 17, 2018. Admittedly, the court's treatment of the issues raised by the North Carolina order are not all contained in the same part of the order. However, when the order is considered as a whole, it meaningfully grappled with the North Carolina court's findings cited by the majority on pages twenty-nine and thirty above.
Regarding notice, the district court considered the issue of whether the Fisher-Lewis class was adequately notified of the Speaks mediation. For the same reasons set forth in the preceding paragraphs, the court found that it was. Specifically, the court rejected "the contention that the Fisher/Lewis class members have some interest, separate and apart, from the interests of the other class members in this case." J.A. 4742.
The district court also considered the issue of whether the notices issued in the Speaks case adequately described the effect of the Speaks settlement on the Fisher-Lewis claims. It reviewed the contents of the actual notices and found the short form notice, the settlement website and the long form notice contained the necessary information about the effect of the settlement. See J.A. 4779 ("The long form notice also clearly details how the Speaks action affects the Fisher/Lewis action.... The settlement website, which is listed on the short form notice and the media publications, provides additional information about the settlement and the Fisher/Lewis action....").
The district court also considered the relative inactivity in Speaks prior to the settlement. It found that while admittedly there was less formal discovery in the Speaks case than the Fisher-Lewis case, there was extensive information discovery and information exchange. See J.A. 4764 ("Here, the parties engaged in informal discovery and class counsel received numerous documents from the Cooperative .... The mediation process provided class counsel with sufficient information about the merits of the case.")
Last, the district court considered whether the record contained evidence of collusion. After reviewing the oral arguments made at two hearings and the extensive briefing on this issue, the district court assessed whether actual evidence of collusion had been offered. The court concluded that there was no evidence presented of collusion besides the timing of the Speaks settlement.
In sum, the record reveals that the district court erred in appointing class counsel but not in the way the majority suggests. The district court erred in appointing class counsel for a settlement class that should not have been certified to begin with because the class representatives were not adequate. However, its conclusions regarding allegations of collusion were not error.7 On this issue, the *311district court applied the law reasonably and based its decision on the record before it.
III.
I turn last to the district court's finding that the settlement was fair, reasonable and adequate. Since the class representatives in this case were not adequate, approval of the settlement was not appropriate in these circumstances. The Supreme Court has found "[a] fairness hearing under subdivision (e) can no more swallow the preceding protective requirements of Rule 23 in a subdivision (b)(1)(B) action than in one under subdivision (b)(3)." Ortiz , 527 U.S. at 858-59, 119 S.Ct. 2295. In this limited sense, I join the majority in reversing the district court's approval of the settlement. However, I write separately to express my concerns about the majority's reasoning for reversing the district court.
Like a district court's certification decision, a district court's approval of a class action settlement is afforded substantial deference and should be reversed only upon a clear showing that the district court abused its discretion. Berry , 807 F.3d at 614.
Significantly, the majority does not find fault with the legal factors the district court considered in reaching its decision. And it does not find fault with the district court in determining the most important of those factors are the relative strength of the plaintiffs' claims and the existence of difficulties of proof or strong defenses. It also does not object to the way the district court evaluated those issues as they relate to the Speaks class. Indeed, the majority describes the district court's analysis as thorough.
Instead, the majority finds the district court abused its discretion because, to the majority, many of the weaknesses the district court found with the Speaks class claims do not apply to the Fisher-Lewis class claims or at least apply with less force. Specifically, the majority contends that the statute of limitations and business judgment rule defenses are less compelling as to the Fisher-Lewis class claims than the district court found. The majority then criticizes the district court's reasoning as to why the $ 76 million settlement offer from 2005 in the Fisher-Lewis action did not make the settlement here inadequate and as to why the failure of a case in Georgia with claims very similar to the Fisher-Lewis class claims suggested weaknesses of the claims in this case.
In my view, the majority here deviates from reviewing the district court's decisions under the abuse of discretion standard. Instead, it substitutes its judgment for the district court's. The district court evaluated each claim for relief. It considered the defenses faced by the plaintiffs, including the statute of limitations and business judgment rule defenses. It considered and found persuasive a nearly identical action brought in Georgia resulting in no recovery. It considered the $ 76 million settlement offer in the North Carolina state case in 2005, determining that offer did not mean the settlement here was inadequate because, in its view, the litigation risks surrounding the claims have changed significantly since 2005 in favor of the Cooperative.
The district court also considered the differences between the Speaks class claims and the Fisher-Lewis class claims. Specifically, the court reasoned:
[C]ounsel ... failed to explain at the fairness hearing or anywhere else how *312including these additional causes of action would provide a route for the Fisher/Lewis plaintiffs to litigate that case through final judgment and final appeal and recover any greater relief than $ 24 million, given that each of these claims appears to rest on the same fundamentally flawed allegation that the Cooperative unlawfully retained reserves.
J.A. 4768. The court then added "[a]t the fairness hearing, this court asked counsel ... to discuss the strength of plaintiffs' claims in Fisher/Lewis and Speaks and explain how plaintiffs in either case could recover anything approaching $ 24 million if they litigated to final judgment and final appeal." J.A. 4770. The district court found the absence of a meaningful response significant.
Reasonable minds could of course differ on the strengths and weaknesses of the Speaks class claims. Indeed, if one hundred judges or lawyers were asked their views about the strengths and weaknesses of the claims here and about an appropriate value of the case, there would likely be one hundred different responses. That is why it is critical that we steadfastly adhere to the applicable standard of review. Under the abuse of discretion standard, our inquiry is simply whether the district court identified and applied the appropriate legal principles and whether it based its decision based on evidence in the record. Here, the district court did that. When that occurs, an appellate court should not substitute its judgment for the district court's.
As noted above, I agree with the majority that the order approving the class settlement should be reversed. However, my agreement is based on the adequacy of the class representative issue as described above. My agreement is not based on the district court's evaluation of the fairness and adequacy of the settlement. In my view, the district court carefully considered the appropriate legal issues and based its decision on a reasonable interpretation of the record.
IV.
In conclusion, I join the majority in affirming the district court's denial of attempted class wide opt-outs and in dismissing the appeal of the order denying the motion to intervene. I also join the majority in reversing the district court's order certifying the class and approving the final settlement. While in the end, the majority and I arrive at the same destination, I arrive by a different path.
One need not look far or long to find current events that illustrate this point. Recent news coverage of social media platforms and the opioid crisis underscore the old truth that innovation can be a double-edged sword.
In pointing out the Cooperative's lack of consistency, I note that this characteristic appears to be contagious in this case. In the parallel state action, the Fisher -Lewis class was apparently not limited to patronage members, see definition of Fisher -Lewis class at Fisher v. Flue-Cured Tobacco Coop. Stabilization Corp., 369 N.C. 202, 794 S.E.2d 699, 704 (2016), but the class nevertheless argued for certification. It is hard to reconcile that position with the one it takes in this case.
While we should defer to the district court on matters within its discretion, an appellate court need not defer to a district court's error of law. Berry v. Schulman , 807 F.3d 600, 608 (4th Cir. 2015)
As described by the majority, the most glaring conflict in the class is the conflict between class members with traceable patronage interests in the Cooperative's reserves and those class members without such interests. Class members with patronage interests can trace their interest in the Cooperative's reserves to their individual contributions to the reserves over time, and class members without those patronage interests cannot. Patronage interests accrued over time and largely derived from contributions made during profitable years in the Cooperative's history and from assessments under the No Net Cost Act.
While typicality and commonality are not specifically raised on appeal, the same issues about which the Objectors complain in the class representative argument also apply to those issues. As the Supreme Court has instructed, typicality, commonality and adequacy of class representative often merge into the same or similar analysis. Gen. Tel. Co. of Sw. v. Falcon , 457 U.S. 147, 157 n.13, 102 S.Ct. 2364, 72 L.Ed.2d 740 (1982). That was in fact the case in Broussard .
While I disagree with the majority on this issue, I in no way minimize the importance of allegations and findings about ethical violations. Compliance with the Rules of Professional Responsibility are fundamental to our profession. My concurrence should not be construed to suggest allegations of ethical misconduct should not be taken seriously or carefully evaluated in the appropriate forum. They should. However, those findings were neither binding on the district court nor necessarily relevant to the district court's adequacy determination here. See, e.g. , W. Morgan-E. Lawrence Water & Sewer Auth. v. 3M Co., 737 F.App'x 457, 465 n.10 (11th Cir. 2018) ("Although state ethical rules may be relevant to our determination of whether approval of the settlement was appropriate, a violation of state ethical rules is neither necessary nor sufficient to establish a violation of federal Rule 23(g)."). What was and remains relevant is whether the district court considered the objections and, in doing so, made decisions for which there is legal and evidentiary support. As noted above, the district did that here.
Under the abuse of discretion standard, an appellate court gives "substantial deference" to a district court's certification and settlement decisions, "recognizing that a 'district court possesses greater familiarity and expertise than a court of appeals in managing the practical problems of a class action.' " Berry, 807 F.3d at 608 (quoting Ward v. Dixie Nat'l. Life Ins. Co. , 595 F.3d 164, 179 (4th Cir. 2010) ). In applying this standard, an appellate court may not substitute its own ideas of fairness for the those of district court. Flinn v. FMC Corp. , 528 F.2d 1169, 1172 (4th Cir. 1975). Rather, an appellate court may reverse a district court's certification or settlement decision only when the district court has committed an error of law or a clear error in finding of fact. Berry , 807 F.3d at 608.