*1 This matter is again before the Court because the parties have been unable to complete the limited discovery permitted by [93] this Court's June 27, 2011 Order. The discovery relates to plaintiffs' only remaining claim in this case: that defendant Test Masters Educational Services, Inc. (“TES”) violated the D.C. Consumer Protection Procedures Act (“CPPA”) by misleading plaintiffs into taking TES's LSAT preparation course when plaintiffs actually meant to enroll in the Robin Singh Educational Services's “TestMasters” LSAT preparation course. See Compl. [ECF 81–3] ¶¶ 1–2; 28 D.C. Code §§ 3901 et seq.

Plaintiffs sought limited discovery related to the CPPA claim in May 2011, see Pls.' Mot to Compel [ECF 81] (“First Mot. to Compel”), and the Court granted that request in its June 27, 2011 Order. Plaintiffs were unsatisfied with TES's responses to their discovery requests, and therefore filed the instant Motion to Compel, for Issue Related Sanctions, and for Monetary Sanctions [ECF 101] (“Second Mot. to Compel”). Plaintiffs argue that (1) TES has spoliated emails between TES representatives and LSAT students or potential LSAT students, (2) TES has spoliated recordings of telephone calls between TES representatives and LSAT students or potential LSAT students, (3) TES improperly redacted some of the documents it did produce, and (4) plaintiffs are entitled to fees and costs for bringing both its initial motion to compel and the instant motion. TES resists each claim, although it does concede that some emails and perhaps phone calls were not produced because they were accidentally lost or destroyed. A hearing was held on plaintiffs' motion on August 8, 2012. For the reasons explained below, the Court will grant in part and deny in part plaintiffs' motion.

I. Legal Standards

1The law regarding sanctions for loss or destruction of evidence is complicated. The first question is whether Federal Rule of Civil Procedure 37 or the Court's inherent power provides the source of authority for issuing sanctions here. Rule 37 authorizes an award of sanctions when a party does not obey a discovery order, and sanctions available under that Rule include “directing that the matters embraced in the order or other designated facts be taken as established for purposes of the action,” prohibiting the disobedient party from introducing evidence on the issue, striking pleadings, dismissing the action, and rendering a default judgment. Fed. R. Civ. P. 37(b)(2)(A). In situations where a party has committed discovery abuses but Rule 37 does not apply, a court may instead issue an appropriate sanction under its inherent power. Shepherd v. Am. Broadcasting Cos., Inc., 62 F.3d 1469, 1474 (D.C.Cir. 1995). Inherent power sanctions include, but are not limited to, “fines, awards of attorneys' fees and expenses, contempt citations, disqualifications or suspensions of counsel, and drawing adverse evidentiary inferences or precluding the admission of evidence.” Id. at 1475.

TES contends that Rule 37 does not apply here, arguing that it cannot be said to have disobeyed the Court's discovery order by failing to produce documents that no longer existed. TES's Opp. to Pls.' Second Mot. to Compel [ECF 105] (“TES Opp.”) at 9–10. Plaintiffs counter that allowing a party to escape Rule 37 by spoliating documents before a discovery order had issued would create a “ ‘race to the shredder.’ ” Pls.' Reply in Supp. of Second Mot. to Compel [ECF 110] (“Pls.' Reply”) at 4 (quoting United Medical Supply Co., Inc. v. United States, 77 Fed.Cl. 257, 269 (Fed.Cl.2007)). Several courts have addressed the question whether loss or destruction of documents before a discovery order is issued can be sanctioned under Rule 37, but there is neither a clear majority view nor an obvious trend in these decisions.[2] Compare Barsoum v. NYC Housing Auth., 202 F.R.D. 396, 399 (S.D.N.Y.2001)(“where spoliation prevents a party from complying with a discovery order, sanctions may be awarded pursuant to Rule 37(b) of the Federal Rules of Civil Procedure”) (citations omitted); Turner v. Hudson Transit Lines, Inc., 142 F.R.D. 68, 72 (S.D.N.Y.1991) (same); In re Air Crash Disaster Near Chicago, Illinois on May 25, 1979, 90 F.R.D. 613, 620–21 (N.D.Ill.1981) (same), with Unigard Sec. Ins. Co. v. Lakewood Eng'g & Mfg. Corp., 982 F.2d 363, 367 (9th Cir. 1992) (finding that “the principles underlying Rule 37 [do not] extend to situations where compulsion orders would be futile, because the evidence has been destroyed”) (internal quotation marks omitted); United Medical Supply Co., 77 Fed.Cl. at 268 (same). Courts that invoke Rule 37 in spoliation situations do so to avoid creating perverse incentives by allowing a party to avoid an anticipated discovery order by destroying documents. SeeTurner, 142 F.R.D. at 72. Courts that rely solely on the inherent power, on the other hand, point out that the “literal language” of Rule 37 does not encompass pre-order spoliation, and the inherent power is meant to fill precisely this sort of gap. Unigard, 982 F.2d at 367.

*2 2The reason this question matters is that the culpability required to impose sanctions under Rule 37 and under a Court's inherent authority may be different. By its terms, Rule 37 imposes no bad faith or willfulness requirement; it simply allows a court to issue any “just orders” necessary to punish past discovery abuses or deter future abuses. Hence, several courts have found that a party that negligently disobeys a discovery order may be sanctioned under Rule 37. See Design Strategy, Inc. v. Davis, 469 F.3d 284, 296 (2d Cir. 2006); Southern States Rack & Fixture, Inc. v. Sherwin–Williams Co., 318 F.3d 592, 596 (4th Cir. 2003). Rule 37's standard is modified, however, in the case of electronic discovery. Under Rule 37(e), “[a]bsent exceptional circumstances, a court may not impose sanctions under these rules on a party for failing to provide electronically stored information lost as a result of the routine, good-faith operation of an electronic information system.” According to the Advisory Committee Notes to Rule 37(e) (then Rule 37(f)), Rule 37(e) was meant to establish an “intermediate standard” between negligence and recklessness that would be applicable to the loss of electronic data. Advisory Committee Notes on Rule 37—2006 Amendment, Subdivision (t). Because this case is about the loss of electronically stored information, this “intermediate standard” applies.

3The law regarding the culpability required to issue sanctions under the Court's inherent power is somewhat muddled in this Circuit. In Shepherd, a case focused on the sanction of default, the D.C. Circuit held that “[a] sanction for failure to preserve evidence is appropriate only when a party has consciously disregarded its obligation to do so.” 62 F.3d at 1481. Later, however, several district courts have held that negligent destruction or negligent failure to preserve evidence can justify an adverse inference instruction. More v. Snow, 480 F.Supp.2d 257, 275 (D.D.C. 2007) (“a court may employ an adverse inference due to a party's failure to preserve evidence, even if deliberate or reckless conduct is not present”) (citations omitted); Mazloum v. D.C. Metro. Police Dep't, 530 F.Supp.2d 282, 292–93 (D.D.C. 2008) (same). In a recent case, the D.C. Circuit agreed. Talavera v. Shah, 638 F.3d 303, 312 (D.C.Cir. 2011) (“Streufert admits to knowing or negligent destruction of his interview notes insofar as the destruction was not accidental ...; non-accidental destruction of his notes supports an inference that the notes would have contained information favorable to [plaintiffs] claim.”). Hence, a court can issue at least lesser sanctions under its inherent power without finding reckless or malicious conduct. In selecting the appropriate sanction, however, a court should consider “the degree of negligence or bad faith involved.” More, 480 F.Supp.2d at 275 (citations omitted).

4With these standards in mind, the Court concludes that it need not decide whether this motion should be addressed under Rule 37 or its inherent powers. Normally, a court must answer this question because some sanctions are justifiable under Rule 37 but not under a court's inherent powers. In the area of electronic data loss, however, the culpability required to justify an award of sanctions is actually less than it is under Rule 37. Hence, the Court can rely on its inherent powers in deciding on an appropriate sanction, and it is of no moment whether Rule 37 also justifies the award.

56Whatever sanction the Court selects, it must “properly calibrate the scales to ensure that the gravity of an inherent power sanction corresponds to the misconduct.” Shepherd, 62 F.3d at 1479 (citations omitted). In addition, in issuing an adverse inference instruction, a court must consider whether

*3 Bolger v. Dist. of Colum., 608 F.Supp.2d 10, 31 (D.D.C. 2009) (citations omitted). Although that test is usually discussed in the context of adverse inference instructions, it naturally applies to any other issue-based sanctions that flow from alleged spoliation of evidence, and the Court will use it as a guide throughout this opinion.

II. Discussion

A. Lost emails

Per the Court's June 27, 2011 Order, plaintiffs requested all emails between TES representatives and LSAT students or potential LSAT students from 2003 to 2010. TES, however, produced almost no emails predating August 2007. TES explains that its server crashed in August 2007 and all of TES's email from before that date was irretrievably lost. See TES Opp., Ex. L (“Israni Decl.”), ¶ 8.

Standing alone, a loss of emails due to a server crash might not warrant any sanction. Indeed, as noted, Rule 37(e)provides that a court “may not” impose sanctions “for failing to provide electronically stored information lost as a result of the routine, good-faith operation of an electronic information system.” In this case, however, several problems with TES's evidence raise the question whether the pre-August 2007 emails were, in fact, lost “as a result of the routine, good-faith operation of an electronic information system.”

The August 2007 crash first came up at the Rule 30(b)(6) deposition of Roger Israni, the president of TES. Pls.' Second Mot. to Compel, Ex. 3 (“Israni Depo.”) at 16. Although Israni testified that TES had an IT professional, Andy Nguyen, Israni stated that he, rather than Nguyen, was the “primary person” who ensured that “automated data backup is operating correctly.” Id. at 165–67. This task included “configur[ing] the backup, set[ting] it at a certain interval, and ... check[ing] to make sure it's working properly.” Id. Israni testified that he did not know whether Nyugen ever checked to ensure that data backup was functioning properly, and Israni could not recall whether he had ever instructed Nyugen to do so. Id.

When asked about the August 2007 server crash, Israni testified that TES had “[a] RAID array. And something was wrong with the RAID array that created some corruption.” Id. at 196. As a result of that corruption, Israni explained, TES “lost the databases that was on that—on that computer” in August 2007. Id. at 197. Although Israni was a Rule 30(b)(6) deponent and had been told that TES's efforts “to identify, locate, and produce electronic documents, including ... emails” would be a subject of his deposition, his testimony about the crash was somewhat confused. Pls.' Second Mot. to Compel at 12. When asked to describe the computer that had crashed, Israni testified that it was “[j]ust one random computer in our office.” Israni Depo. at 197. When pressed about whether the desktop was backed up, he stated that the desktop was actually acting as a server. Describing the crash itself, Israni testified:

*4 Id. at 199–200. Israni did not know whether TES had kept any of the failed hard drives. When asked whether TES had tried to reuse them, he said: “We might have. I don't know.” Id. at 200. Israni was equally vague in explaining TES's efforts to recover data after the crash:

Id. at 203. Israni also testified that he didn't know what data had been lost in the crash, though he did affirm that “to [his] knowledge,” no “information that [he] wanted to find for this lawsuit [was] inaccessible because it was on that desktop that crashed.” Id. at 201.

Although Israni's testimony was confused or vague during much of the deposition, TES made no effort to amend or supplement any of his answers after the deposition. But three and a half years later, in response to this sanctions motion—which was based in large part on Israni's 30(b)(6) deposition testimony—Israni wrote a thorough and detailed declaration with a new explanation for the August 2007 crash. See Israni Decl. ¶¶ 2–11. In the declaration, Israni reiterated that TES had a RAID array server system, where only a crash of two or more hard drives should cause data loss. Israni Decl. ¶ 6. But, contrary to his deposition testimony, he averred:

Id. ¶¶ 7–8.

The question now before the Court is how to make sense of these two accounts in light of the test governing issue-based sanctions. As explained above, that test considers whether the party from whom discovery is sought had a duty to preserve the information; whether the party is culpable in the loss or destruction of the information; and whether the Court can conclude that the lost or destroyed information would have helped the movant. Bolger, 608 F.Supp.2d at 31.

7TES does not seriously dispute that it had a duty to preserve data relevant to this litigation beginning, at the latest, when it was served with the complaint in 2004. See Smith v. Cafe Asia, 246 F.R.D. 19, 21 n. 1 (D.D.C.2007). Emails between TES representatives and LSAT students would have been relevant to, at a minimum, plaintiffs' claims under the CPPA. Accordingly, the first element of the sanctions test is satisfied.

8The second question is whether the loss of the emails “was accompanied by a culpable state of mind.” Bolger,608 F.Supp.2d at 31. In evaluating this issue, the Court is troubled by Israni's two different accounts of the August 2007 crash. If Israni was the president and, according to him, the “primary person” who handled data backup at TES's small office—in addition to being a 30(b)(6) deponent—he certainly should have been very familiar with a crash that occurred less than a year before the deposition that allegedly erased all of TES's email. Indeed, although Israni never specified what data besides email was lost in the crash, it seems likely that the crash Israni described in his deposition would have destroyed a large amount of data. Plaintiffs' expert states that the crash Israni described would have resulted in the loss of all data saved in the RAID array, not just emails, and TES did not challenge that statement at the motions hearing. Pls.' Reply., Ex. B at 18.

*5 Despite the possibly catastrophic nature of the crash and Israni's responsibility for data backup at TES, Israni's deposition testimony suggested that he barely remembered the event. He testified that only one hard drive crashed—which normally would not cause permanent data loss in a RAID 5 array—and offered only an extremely vague description of why the data had been permanently lost. See Israni Depo. at 199–200 (“If you try to rebuild [the hard drive that crashed]—and I think in this case, if I remember correctly, this is where the problem came about, when it's rebuilt, if it's done incorrectly, then it can corrupt the other databases. I'm not saying that's what happened in this particular case, but I'm just trying to describe an instance where it's not pure redundancy.”). Israni never mentioned any sort of backup server. Yet, at the time of his declaration, he was confident that two hard drives had crashed and that the backup server was corrupted. Israni Decl. ¶¶ 7–8. Moreover, in his deposition, Israni said that TES had not called in any outside help, that he did not know what happened to the hard drives, and that he did not know whether Mr. Nguyen had been at all involved in the post-crash recovery. Israni Depo. at 199–203. Three and a half years later, however, Israni remembered that TES had “immediately” sought to retrieve information from backup servers and then hired a “computer expert,” who examined both the server and the backup image before concluding that nothing could be salvaged. Israni Decl. ¶ 8.

There are two separate problems here. First, Israni obviously arrived unprepared for his Rule 30(b)(6) deposition and made no effort to supplement his answers until confronted with a sanctions motion years later. As several courts have found, such ambush-by-declaration is unacceptable:

Rainey v. Am. Forest & Paper Ass'n, Inc., 26 F.Supp.2d 82, 94 (D.D.C.1998). This case shows exactly why that rule exists: plaintiffs have now invested considerable time and energy in preparing a motion to compel, the basic factual premises of which are undermined by Israni's new declaration. Moreover, the time plaintiffs invested in taking the Rule 30(b)(6) deposition of Israni is rendered effectively worthless if TES can now simply replace his deposition answers with ones that better serve their current purposes.

Even setting aside fairness issues, the Court simply cannot credit both the deposition and the declaration because they contain too many inconsistencies. Given the necessity of making a choice between the two accounts, the Court finds Israni's deposition testimony to be the more credible explanation, as it is somewhat more plausible, was written much closer to the actual crash and was not tailored to respond to a motion to compel. Hence, both for that reason and because a party cannot supplement a 30(b)(6) deposition in this way, the Court will look only to Israni's deposition in determining whether TES was culpable in losing the pre-August 2007 emails.

Based on the deposition, the Court concludes that TES was at least negligent in the loss of the emails. Although TES's use of the RAID 5 array to back up the data was appropriate, TES's efforts to retrieve the data after the crash were sorely lacking. According to the deposition, Israni made no effort to find out what data had been lost or to seek outside assistance (or even the assistance of TES's IT professional) in restoring the data, and he may not have even preserved the failed hard drives. As the president of TES, Israni must have known that some data in TES's email would be relevant to the pending litigation. Given that, making such a lackluster effort to retrieve the data—indeed, it is not clear from the deposition that any effort was made—constitutes a conscious disregard of TES's preservation obligations that can fairly be described as gross negligence or recklessness.[3]

*6 9Plaintiffs bolster their contention that TES should be sanctioned for loss of emails by arguing that TES's production of post-August 2007 material was also incomplete. Pls.' Second Mot. to Compel at 10–21. TES has repeatedly stated that it has produced all relevant post-August 2007 emails, but plaintiffs argue that certain gaps in the production indicate that spoliation occurred. Specifically, plaintiffs point to (1) the lack of any course cancellation emails in TES's production, even though it is undisputed that some course cancellation emails were sent; (2) the lack of internal emails among TES employees, and (3) the lack of emails from Will Howorth, even though plaintiffs independently obtained copies of emails between Howorth and TES representatives from Howorth himself. With respect to course cancellation emails, TES responds that such emails were sent through “batch processing” that leaves no copy of the email on the email server. TES Opp. at 17. Hence, TES has only produced a few samples of the text of such emails. Contrary to plaintiffs' contentions, the Court does not see any significant differences between TES's text samples and the cancellation email plaintiffs' counsel obtained independently from an LSAT student. Compare TES Opp., Ex. U with Pls.' Status Report [ECF 97], Ex. K at 6. Hence, although the lack of cancellation emails is a bit odd, TES's explanation is not incredible, and that absence does not show that spoliation occurred. The absence of a single internal email between TES employees is even stranger, but TES's counsel explained at the hearing that, because of the ongoing litigation, TES employees talked about LSAT courses in person rather than memorializing their conversations in emails. Tr. of Mot. Hearing (Aug. 8, 2012) at 37–38. That explanation is not completely unbelievable, and plaintiffs have not persuasively shown that post–2007 internal emails must have existed. TES did not provide any explanation for its failure to produce Will Howorth's email, but because neither party really focused on this issue, the Court will not conclude that this alone indicates any systemic gaps in TES's production. In sum, although the Court is again troubled by possible problems with TES's post–2007 production—particularly in light of the loss of pre–2007 emails—the Court cannot conclude by a preponderance of the evidence that any post-August 2007 emails were negligently, recklessly, or intentionally lost.

1011Having found that the loss of the pre-August 2007 emails was accompanied by a culpable state of mind, the Court turns to the third and final element of the test: whether there is any evidence that “the lost evidence would have supported the claims or defense of the party that sought it.” Bolger, 608 F.Supp.2d at 31 (citations omitted); see also Mazloum, 530 F.Supp.2d at 293 (asking whether evidence “would have been of the nature alleged by the party affected by its destruction”) (citations omitted). As several courts have pointed out. however, requiring the movant to prove with particularity the contents of a spoliated email can be unfair. See E.I. du Pont de Nemours and Co. v. Kolon Indus., Inc., 803 F.Supp.2d 469, 498–99 (E.D.Va. 2011); United States ex rel. Miller v. Holzmann, No. 95–1231, 2007 WL 781941, at *1 (D.D.C. Mar. 12, 2007). Hence, a party may satisfy this third factor by “establishing that the spoliated material addressed topics, or falls into categories of documents, that would be favorable to the movant's case.” Kolon, 803 F.Supp.2d at 498–99.

In support of their claim that the destroyed emails would be relevant and helpful, plaintiffs point to the declarations of two LSAT students who mistakenly signed up for TES's LSAT courses. Both students declare that they asked TES employees questions (one via email, one via phone) indicating that they were confused about the difference between TES's course and Robin Singh's course, and both say that TES employees avoided clarifying the situation. Pls.' Mem. in Opp. to Mot. for Sanctions [ECF 62], Ex. 6 at 10–12; First Mot. to Compel, Ex. M at 18–23. This is precisely the sort of evidence that would support plaintiffs' CPPA claims, which are based on both misrepresentations and material omissions by TES employees. That plaintiffs have located two students who claim that TES misled them suggests that there may be other such students, and conversations with those other students would be found in TES's email archives (or in phone recordings, which are discussed further below). TES has done nothing to refute this argument. Hence, and in light of the intrinsic difficulty of proving the relevance of spoliated data, the Court finds that plaintiffs have made out the third element of the test.

12Plaintiffs have therefore established that they are entitled to some sort of issue-based sanction to cure the prejudice from spoliated evidence. The next question is what sanction is “calibrate[d]” to make up for the loss of the relevant evidence. Shepherd, 62 F.3d at 1479 (citations omitted). Plaintiffs suggest four possibilities: (1) a “[f]inding and jury instruction that defendant intended to deceive consumers into believing it was the well-reputed TestMasters,” (2) a “[f]inding and jury instruction that Defendant made false statements and material omissions in order to deceive Plaintiffs and others into believing that it was the nationally known and well-reputed TestMasters,” (3) “[a] prohibition on TES introducing any evidence or argument as to what statements its representatives made to Plaintiffs or other consumers regarding the identity of their company,” or (4) “an instruction to the jury that as a result of TES's failure to produce certain evidence [within its control], the jury must presume that the evidence would have been unfavorable to TES.” Pls.' Second Mot. to Compel at 27.

*7 1314In order to prove their CPPA claims, plaintiffs must show that TES made material misstatements, omitted material facts, or passed their services off as those of another. 28 D.C. Code § 28–3904(e), (f), (s). Plaintiffs need not prove that TES misled students intentionally in order to show a violation of the CPPA, but they are eligible for punitive damages only if they prove by clear and convincing evidence that TES acted with “a state of mind evincing malice or its equivalent.” See Daka, Inc. v. McCrae, 839 A.2d 682, 695 n. 14 (D.C. 2003) (citations omitted). Proof of intent to deceive plaintiffs can be established by evidence of other wrongful acts; here, for instance, plaintiffs could point to deceptive conduct in 2010 to show that the plaintiffs were deceived in 2003. SeeJuergens v. Urban Title Services, Inc., 533 F.Supp.2d 64, 76 (D.D.C. 2008). Finally, in order to prevail on their claim for injunctive relief under the CPPA, plaintiffs must establish a “likelihood of future violations” of the CPPA by TES. See Am. Historical Ass'n v. National Archives & Records Admin., 516 F.Supp.2d 90, 105 (D.D.C. 2007) (citations omitted).

15Plaintiffs' first and second suggested sanctions would essentially dispose of the case by resolving the only disputed issues: what statements TES made (or failed to make) and its culpability in doing so. Either sanction would resolve both the merits of the case and the claim for punitive damages in plaintiffs' favor. The D.C. Circuit has cautioned, however, that district courts should be extremely cautious in entering either default judgments or issue-based sanctions that are the practical equivalent of default judgments. Shepherd, 62 F.3d at 1479. Because of the strong interest in deciding cases on the merits, a court may only impose a default judgment if it finds that no lesser sanction will serve to cure the prejudice from the destruction or loss of information. Id. at 1472–73. The Court finds that that high standard is not met here. TES's loss of pre-August 2007 emails does not completely preclude plaintiffs from establishing the contents of those emails; plaintiffs may seek to obtain copies of the emails from the LSAT students, as indeed they have already done in some cases, or have LSAT students testify about their contents. Indeed, as discussed below, the Court will order TES to provide email addresses for almost 200 more LSAT students to plaintiffs, which will facilitate plaintiffs' effort to recover the missing emails or obtain testimony about their contents. Hence, because this is not a situation where plaintiffs have no way to prove their case without the missing documents, an issue-based sanction that is the practical equivalent of a default judgment is not warranted.

That leaves the Court with a choice between plaintiffs' last two suggestions: prohibiting TES from putting on evidence on the issues of misrepresentation and intent or entering an adverse inference instruction. The first option, in the Court's view, is problematic. With minimal exceptions that are not relevant here, TES has produced no emails from before 2007; thus, it does not appear that TES has cherrypicked only documents that are favorable to its case for production. Nor, as previously explained, does the Court see any evidence that the post–2007 emails were only partially produced for tactical reasons. Therefore, in the Court's view, it is not appropriate to limit the already thin record in this case by excluding TES's evidence on the key issue.

That leaves the option of an adverse inference instruction, one of the most frequently employed issue-based sanctions. The Court finds that such an instruction is a fair sanction here. TES culpably lost evidence on the key issue in the case, but as previously explained, plaintiffs should be able to muster enough evidence on the issue to allow for adjudication on the merits. An evidentiary presumption that can be considered along with the rest of the evidence will make that adjudication fairer. The Court will therefore grant the fourth sanction suggested by plaintiffs, and at trial will give an adverse inference instruction stating that TES failed to produce certain emails within its control, and that the factfinder therefore must presume that the evidence contained in those emails would have been favorable to plaintiffs, i.e., would have supported the position that TES representatives made material misstatements about the relationship between TES's LSAT courses and Robin Singh's LSAT courses, omitted material information about that relationship, or tried to pass TES's service off as Robin Singh's service.[4]

B. Telephone call recordings

*8 The dispute over recordings of phone calls mirrors the dispute over emails to some extent. According to Israni, TES purchased several devices that could record telephone calls because TES wanted to have evidence of allegedly harassing calls by Robin Singh and his associates. Israni Decl. ¶ 19. Plaintiffs believe that some calls between TES representatives and LSAT students were also recorded, and they requested that TES turn over those recordings. Second Mot. to Compel at 21–27. TES maintains that no recordings of student calls exist now or have ever existed. TES Opp. at 19–22. As with the emails, however, Israni's deposition testimony and his declaration conflict.

A substantial portion of Israni's 30(b)(6) deposition focused on whether TES had ever recorded, automatically or otherwise, calls from DC-area LSAT students or prospective LSAT students. Israni was clear that TES employees had never deliberately recorded calls from LSAT students, and plaintiffs do not dispute this point. Israni Depo. at 64. Israni's testimony on whether any such calls would have been recorded automatically, however, was conflicting. He first testified that at the time of the deposition, he had to make an affirmative decision to record calls:

Id. at 63. Israni also testified, however, that some systems had an option to record calls automatically. Id. at 64. And, when pressed about whether he had told a particular caller that her call was being recorded, Israni suddenly began discussing automatic recordings of calls:

Id. at 71–72.

But in his declaration, Israni again rejected the notion that calls were automatically recorded. He wrote that:

Israni Decl. ¶ 18. Israni's declaration stated unequivocally that the call with Heidi Singer had not been automatically recorded. Id. ¶ 19. He also attempted to distinguish his deposition testimony, writing that he “understood Plaintiffs' attorney to be asking me [at the deposition] about the phone system's capabilities,” id. (emphasis added), rather than whether calls actually were recorded automatically.

*9 In his deposition, Israni made clear that TES never had a policy or practice of preserving recordings. Instead, TES allowed old recordings to be overwritten with newer ones, and would return disks with recorded calls to the vendor upon the purchase of a new system. Israni Depo. at 65–66, 73, 80–81. Israni's declaration did not contradict that testimony. The declaration did state, however, that TES conducted searches for responsive recordings in 2008 and 2011, and found none. Israni Decl. ¶ 19.

16With those facts in mind, the Court turns to the same three questions addressed in the discussion of emails: whether TES had a duty to preserve the telephone call recordings; whether TES culpably failed to do so; and whether the recordings would have been relevant and helpful to plaintiffs' case. The first question that must be resolved is whether TES actually ever recorded any calls with LSAT students or potential LSAT students; if not, there was nothing to preserve and not duty of preservation. Plaintiffs, correctly, do not argue that TES had a duty to record all of its phone calls; rather, they argue that since TES chose to record some phone calls, it assumed a duty to preserve any recordings relevant to this litigation—assuming those recordings existed. Hearing Tr. at 23.

Plaintiffs do not question Israni's testimony that the recording systems were purchased in order to record allegedly harassing calls from Robin Singh and his associates, and that TES employees never deliberately chose to record LSAT students' calls. The question, then, is whether calls from LSAT students could have been recorded automatically. Israni's declaration says that no such automatic recordings were ever made, but for the reasons already explained, the Court will not credit the declaration to the extent it contradicts the deposition. The deposition testimony is internally inconsistent and confusing, but the Court concludes that the testimony shows that some student calls were likely automatically recorded. When Israni said that calls were never automatically recorded, it appears that he was talking about the current recording system; when discussing calls recorded in the past, however, he said that “the system in some cases automatically records. I don't remember what the system was set up at this time, if it was an automatic recording. There's no way for me to know, during the actual conversation, whether the call is being recorded or not [when the system is set up to record automatically].” Isrnai Depo. at 71–72. This strongly suggests to the Court that at least one of TES's prior phone systems was set up to do automatic recordings.

If some LSAT student calls were, in fact, recorded, Israni's deposition testimony makes clear that TES recklessly failed to preserve the recordings. According to Israni, TES had no policy in place for preserving such recordings, but rather allowed recordings to be overwritten by new ones and returned disks full of recordings to the vendors. Rule 37 makes clear that the user of an electronic system must turn off such automatic overwrite features to comply with its preservation obligations, see Advisory Committee Notes on Rule 37—2006 Amendment, Subdivision (f), and the D.C. Circuit has held that “non-accidental” destruction in the face of a preservation duty may be regarded as culpable under an inherent power analysis. Talavera, 638 F.3d at 312. Under either standard, the loss of the phone call recordings was culpable.

*10 17The relevance factor, however, is not so straightforward. As explained above, plaintiffs have some evidence that TES representatives made misleading statements (or failed to correct misimpressions) in phone conversations with students. Hence, there likely were phone calls that would have been relevant to plaintiffs' case. The trouble, however, is in concluding that those phone calls were among the subset of calls that were recorded. Israni did not say how many calls were recorded, but his testimony suggests that (a) not all systems were set to record automatically and (b) those that did record automatically only recorded a small percentage of calls. Israni Depo. at 71–72. Hence, in order to find that relevant information was destroyed, the Court would have to find that the small percentage of calls that were automatically recorded overlapped with the presumably small percentage of calls in which TES representatives had occasion to discuss the difference between TestMasters and Test Masters with confused LSAT students.

Thus, the phone call recordings present quite a different situation from the emails. Ordinarily, all of TES's email would have been saved. Hence, even if only a very small percentage of emails sent by TES representatives were relevant to this litigation, that small number would have been preserved if not for TES's recklessness. But, for the reasons explained above, the same is not true of the phone calls. With a small group of recordings and a small group of relevant calls, the Court is simply not prepared to find by a preponderance of the evidence that the two sets overlapped, i.e., that any relevant calls were actually recorded and thus should have been preserved. In making this assessment, the Court bears in mind that even the least severe sanction suggested by plaintiffs is still quite harsh. The adverse inference instruction suggests to the jury that TES is a bad actor, and it allows the jury to draw inferences that might not actually be supported by the missing documents. See Morris v. Union Pacific R.R.,373 F.3d 896, 900–01 (8th Cir. 2004). Because the Court views it as unlikely that any nontrivial number of potentially misleading phone calls actually were recorded and discarded, the Court is loathe to instruct the jury to draw those inferences. The Court will therefore deny the requested sanction.

C. Redacted emails

18The Court's June 27, 2011 Order permitted plaintiffs to serve a request for production of “[d]ocuments reflecting communications with or identifying LSAT consumers (or prospective LSAT consumers) in the District of Columbia metropolitan area from 2003 through 2010.” June 27, 2011 Order [ECF 93] at 6. Plaintiffs served a request for production that tracked the language of the June 27, 2011 Order. Part of TES's response included documents entitled “Personal Info,” which contained contact information for TES students. Although the “Personal Info” documents had a field for email addresses, TES redacted email address from each of the 226 “Personal Info” documents it produced. Pls.' Second Mot. to Compel at 27–29; see also Pls.' Status Report, Ex. Q (sample Personal Info documents). TES claims that it was justified in making these redactions because plaintiffs had not previously defined the word “identify” to include email addresses, and email addresses were therefore not responsive to the request for “documents ... identifying LSAT consumers.” TES Opp. at 41–43.

There is a technical argument that TES was within its rights in redacting LSAT students' email addresses from the documents it produced. But this is the sort of petty behavior that has made discovery in this case so difficult and unpleasant. TES shall produce the documents without redacted email addresses forthwith.

D. First Motion to Compel

A court may grant attorney's fees and costs to a party who prevails on a motion to compel. In deciding whether such an award is appropriate here, the initial question is whether the Court would make any award under Rule 37or under its inherent authority. TES argues that Rule 37(a), which plaintiffs cite in their request for attorneys' fees, does not apply when the motion is decided under the Court's inherent authority. Regardless of the source of its authority, however, the Court finds that the standard laid out in Rule 37(a) provides a logical and appropriate test for whether to award expenses to a prevailing party, and the Court will therefore adopt it here.

*11 19Under Rule 37, a party who prevails on a motion to compel must be awarded “reasonable expenses incurred in making the motion, including attorney's fees” under Rule 37(a)(5)(A) unless the non-movant's opposition to the motion was “substantially justified.” Rule 37(a)(5)(A)(ii). A party is “substantially justified” in opposing a motion if reasonable people could differ in their views on the motion. Pierce v. Underwood, 487 U.S. 552, 565, 108 S.Ct. 2541, 101 L.Ed.2d 490 (1988). When a motion is granted in part and denied in part, the Court should apportion expenses, including attorney's fees, accordingly.[5] Fed. R. Civ. P. 37(a)(5)(C).

20There are two separate motions to compel here, so the Court will examine each in turn. Plaintiffs' first motion to compel sought four categories of evidence: recordings of telephone calls between TES representatives and D.C. LSAT students; names and contact information of TES representatives who spoke with D.C. LSAT students; documents reflecting communications with or identifying D.C. LSAT students; and documents reflecting TES's net worth. See generally Pls.' First Mot. to Compel. TES opposed these requests on the grounds that (1) the prior judges assigned to this case had already denied the discovery requests, and (2) plaintiffs had waived both arguments by not raising them during an appeal to the D.C. Circuit. TES's Opp. to First Mot. to Compel [ECF 91] at 3–12. TES's opposition contained only minimal responses to the substance of the requests; TES claimed that it would need more time to draft such a response, although TES inexplicably turned down plaintiffs' offer to consent to a motion for extension of time and even to prepare the motion. Id. at 22. The Court rejected both of TES's arguments in its June 27, 2011 Order as obvious mischaracterizations of the record and the law. Order [ECF 93] at 4–5. The Court therefore granted plaintiffs' first three requests in their entirety, and effectively granted the fourth as well (by requiring a certified statement of net worth instead of documents that would allow plaintiffs to compute TES's net worth). Id. at 5–6. On the first motion to compel, then, plaintiffs were completely successful. Although plaintiffs' requests for sanctions were not granted, those requests were only made in the alternative.

21The present motion to compel and for issue-precluding sanctions will be granted in part and denied in part. Plaintiffs have succeeded in their request for a sanction relating to the lost emails. The Court finds that TES's opposition to this motion was not substantially justified because it was based almost exclusively on a declaration that TES should have known was an illegitimate attempt to “supplement”—i.e., contradict—a 30(b)(6) deposition. On the other hand, the sanction for the loss of telephone recordings will not be granted, and expenses will therefore not be awarded for that aspect of the motion. Plaintiffs also succeeded on their claim that TES should produce unredacted “Personal Info” documents, but TES's opposition was “substantially justified” because it was based on a plausible, if hypertechnical, reading of plaintiffs' requests for production.

*12 The Court will therefore order TES to pay plaintiffs' reasonable expenses for the first motion and one-third of the reasonable expenses for the second motion. Plaintiffs' filing does not separate out which expenses were incurred in preparing each motion, so plaintiffs shall resubmit their request for payment with appropriately apportioned expenses. TES shall have two weeks after plaintiffs' submission to respond with any objections.

III. Conclusion

For the reasons given above, the Court will give an adverse inference instruction stating that TES failed to produce certain emails within its control, and that the factfinder therefore must presume that the evidence contained in those emails would have been favorable to plaintiffs, i.e., would have supported the position that TES representatives made material misstatements about the relationship between TES's LSAT courses and Robin Singh's LSAT courses, omitted material information about that relationship, or tried to pass TES's service off as Robin Singh's service. The Court will also order TES to produce the “Personal Info” documents without redacting student email addresses. Finally, the Court will order plaintiffs to resubmit their requested expenses for their two motions to compel. A separate Order accompanies this opinion.