Glassdoor, Inc. v. Superior Court of Santa Clara County

Opinion

Petitioner Glassdoor, Inc. (Glassdoor), operates a Web site on which workers can post “reviews” of past and current employers. Real

“3. Terrible work-life balance, except for the platform team, which do not know what to work on. For Data Science team and Game Engineering team, people usually go home after 10:00pm and have on-call duties every month.

“4. The senior management lost directions. The company has invested heavily in the platform team (there are 70-80 engineers). However, after one year, nothing has been done by that team. The CEO said in the team meeting: I don’t expect products and revenue from the platform team. I only want you can show demos. The platform is only for attracting investments from VCs.”

Under the heading “Advice,” the review stated, “Stop telling the investors and employees the unreal information. A company cannot survive forever by cheating!” The review went on to assert that employees were “Very Dissatisfied”; that they, or Doe, “ ‘Disapprove’ [of] Gabriel Leydon (CEO)”; that Doe would not recommend MZ to a friend; and that MZ’s business outlook was “Getting Worse.”

According to MZ, it notified Glassdoor on June 22, 2015, that, in its view, the post disclosed “confidential information regarding Machine Zone’s valuation and fundraising, as well as internal, confidential statements made by Machine Zone’s CEO and management regarding Machine Zone’s confidential and strategic business plans.” MZ states that the review was removed from the Web site on June 23.

MZ filed its complaint on July 1, 2015, asserting a single cause of action against Doe for breach of contract. It alleged that Doe breached the nondisclosure agreement by “disclosing to third parties Machine Zone’s confidential, non-public information.” MZ did not identify the statements in the review supposedly having this effect, nor did it specify the confidential information supposedly disclosed. Instead it broadly alleged that Doe had “provided details concerning undisclosed technology Machine Zone has and is developing, the stage of development of that technology and the scope of Machine Zone’s investment therein.” MZ further alleged that the post “quoted Machine Zone CEO Gabriel Leydon’s confidential internal statements concerning that technology.”

On July 2, Machine Zone promulgated a subpoena directing Glassdoor to produce a copy of Doe’s review as well as information identifying its author. Glassdoor produced a copy of the review but otherwise objected to the subpoena on the grounds, among others, that disclosure of the poster’s identity would violate his “right to speak anonymously under the First

A decade ago, such a contention presented a relatively novel question. Now, however, a substantial preponderance of national authority favors the rule that publishers, including Web site operators, are entitled to assert the First Amendment interests of their anonymous contributors in maintaining anonymity. (See Digital Music News LLC v. Superior Court (2014) 226 Cal.App.4th 216 , 228, fn. 12 [ 171 Cal.Rptr.3d 799 ], quoting Rancho Publications v. Superior Court (1999) 68 Cal.App.4th 1538 , 1541 [ 81 Cal.Rptr.2d 274 ] [“a nonparty ‘to civil litigation (such as a newspaper) [may] assert the constitutionally protected rights of an author to remain unknown’ ”]; McVicker v. King (W.D.Pa. 2010) 266 F.R.D. 92 , 95 [“The trend among courts which have been presented with this question is to hold that entities such as newspapers, internet service providers, and website hosts may, under the principle of jus tertii standing, assert the rights of their readers and subscribers.”]; In re Indiana Newspapers Inc. (Ind.Ct.App. 2012) 963 N.E.2d 534 , 549 [“when a third-party entity, such as a newspaper, is subpoenaed to reveal the identity of an anonymous commenter who has used that third party as a forum for his anonymous speech, the third-party has standing to contest the subpoena under the principle of jus tertii”]; Pilchesky v. Gatelli (Pa.Super.Ct. 2011) 2011 PA Super 3 [ 12 A.3d 430 , 437, fn. 9] [dictum; standing not raised and not subject to determination sua sponte]; Trawinski v. Doe (N.J.App.Div., June 3, 2015) 2015 WE 3476553, p. *5; In re Subpoena Duces Tecum to America Online, Inc. (2000) 52 Va.Cir. 26 (AOL), revd. on another ground in America Online, Inc. v. Anonymous Publicly Traded Co. (2001) 261 Va. 350 [ 542 S.E.2d 377 ]; In re Verizon Internet Services (D.D.C. 2003) 257 F.Supp.2d 244 , 257-258 (Verizon), revd. on another ground in Recording Industry Assn. of America, Inc. v. Verizon Internet Services, Inc. (D.C. Cir. 2003) 351 F.3d 1229 , 1239.)

B. The Matrixx Decision

MZ contends that a contrary rule was adopted by this court in Matrixx Initiatives, Inc. v. Doe (2006) 138 Cal.App.4th 872 [ 42 Cal.Rptr.3d 79 ] (Matrixx). But that case did not involve the standing of a publisher or service provider. The person attempting to assert the rights of the anonymous online speaker there was a deponent who denied any connection to the offending posts, even though one of them had been traced to a hedge fund he managed. (Id. at p. 876.) This made him a “third part[y] in a lawsuit that may have

Glassdoor is not an avowed stranger to the speaker, as was the objector in Matrixx. It is the acknowledged publisher of the speech at issue. Such a publisher has a strong interest in protecting the right of its users to speak anonymously. Thus the court in AOL, supra, 261 Va. 350 [ 542 S.E.2d 377 ], observed that the service provider would be harmed by disclosure of the user’s identity because “ ‘[i]f [it] did not uphold the confidentiality of its subscribers . . . one could reasonably predict that [its] subscribers would look to [its] competitors for anonymity.’ ” (Matrixx, supra, 138 Cal.App.4th at p. 880, quoting AOL, supra, 52 Va. Cir. 26 , 32.) Similarly, failure by the service provider in Verizon to protect its users’ anonymity would diminish its “ ‘ability to maintain and broaden its customer base.’ ” (Matrixx, supra, 138 Cal.App.4th at p. 880, fn. omitted, quoting Verizon, supra, 257 F.Supp.2d 244 , 258.) Another court found that a newspaper had standing to defend a poster’s anonymity where ‘“preventing [it] from asserting the First Amendment rights of anonymous commentators” on its Web site would ‘“compromise the vitality of the newspaper’s online forums, sparking reduced reader interest and a corresponding decline in advertising revenues.” (Enterline v. Pocono Medical Center (M.D.Pa. 2008) 751 F.Supp.2d 782 , 786.)

The situation here is the same as in the cases distinguished by Matrixx. As Glassdoor’s corporate counsel declared, its business model ‘“relies on maintaining its users’ anonymity. The reliability of the information on glassdoor.com would likely decrease if litigants could readily obtain users’ identities, because users would fear retaliatory litigation based on the information they posted.” This would naturally tend to harm Glassdoor’s interests, because its usefulness to potential readers depends on the degree to which posters feel able to frankly recount their employment experiences without fear of adverse consequences. In the case of a current employee, such consequences can be severe, up to and including termination of employment. Even a past employee may be exposed to retaliation by, for instance, unfavorable references. Anonymity may provide a would-be poster’s only real protection against such consequences. By providing it, Glassdoor creates an opportunity for users to safely provide content of interest to other users. In exchange Glassdoor

Doe could of course engage counsel to appear on his behalf, but there is no reason to believe that this would yield better representation than Glassdoor will provide. Indeed, one consequence of denying standing to a publisher in Glassdoor’s position would be to cast upon the anonymous speaker the potentially prohibitive cost of defending the right to anonymity. MZ’s approach would thus result in what Justice McAdams characterized as sending mixed signals to would-be anonymous speakers: “The good news . . . [is that] your message will be protected by the First Amendment and your identity will be protected by the court quashing a third party subpoena .... The bad news: it may cost you tens of thousands of dollars to preserve your anonymity.” Tendler v. www.jewishsurvivors.blogspot.com (2008) 164 Cal.App.4th 802 , 810 [ 79 Cal.Rptr.3d 407 ] (conc. opn. of McAdams, J.) (Tendler).) There is no basis to assume that the typical online commenter has access to that kind of money. If not, and unless they can interest some

Even where the anonymous speaker can afford to pay for a defense, the prospect of doing so can only inhibit the speech at issue. Most content providers like Doe are unlikely to receive, and do not expect to receive, any economic reward for the content they provide. MZ’s rule would require them to decide whether to engage in an activity creating a significant risk of substantial pecuniary harm while offering no prospect of material reward. The prudent decision is to refrain from posting. From this perspective the publisher may have a greater interest in the right of anonymity than its contributors do, for they have the option of simply declining to speak—a decision that directly injures the publisher’s business. If the publisher is prepared to ameliorate this inhibiting effect by stepping into its contributors’ shoes when their anonymity is threatened, we see no sound reason to forbid it. Denial of that right would serve neither the purposes of prudential standing requirements nor the broader interests of a society devoted to the free flow of ideas and information.

Of course the right to speak anonymously is not an unalloyed good. Anonymity can facilitate various kinds of harmful speech, including defamation, wrongful disclosure of private information, and malicious disinformation. But we are here concerned with the threshold question whether jus tertii standing is justified by the inhibitory effect of burdening anonymous speakers with the cost of preserving their anonymity before any assessment of wrongfulness has been made. Indeed, in its present posture it cannot even be assumed that a suit such as this one was filed in the reasonable &<?//<?/that the speech was wrongful, or with the reasonable hope of prevailing. As Justice McAdams observed in Tendler, supra, 164 Cal.App.4th at page 812 (conc. opn. of McAdams, J.), “[sjome requests” for disclosure of a poster’s identity ‘“will be based on a legitimate right to discover the source of libelous statements or business disinformation schemes; but some will be solely for the purpose of silencing a critic by harassment, ostracism, or retaliation.” We would go further and suggest that some attacks on anonymity may be mounted for their in terrorem effect on potential critics. Here, for instance, such a suit might be intended to deter other past or present employees from criticizing MZ—an effect that would violate the public policy manifested in California statutory law, which prohibits employer restrictions on, or punishment for, speech regarding conditions of employment. (Lab. Code, § 232.5.) The message conveyed to such would-be speakers by the rule MZ advocates is the one articulated by Justice McAdams: that they may be entitled to engage in anonymous criticism, but only if they are prepared to pay a significant sum—or, as he says, ‘“tens of thousands of dollars”—to prevent disclosure of their identities. (Tendler, supra, at p. 810 (conc. opn. of

We conclude that Glassdoor has standing to assert John Doe’s interest in maintaining his anonymity as against Machine Zone’s efforts to compel Glassdoor to identify him.

Both parties treat our decision in Krinsky, supra, 159 Cal.App.4th 1154 , as providing the proper analytical framework for determining whether MZ is entitled to discover Doe’s identity. As we there observed, disputes of this kind are governed by a tripartite standard of review. We review the trial court’s order independently insofar as it rests on undisputed facts. (Id. at p. 1161.) We also determine independently the scope of First Amendment protection to be accorded to the speech upon which the plaintiff’s claim is predicated. (Krinsky, at pp. 1161-1162.) We defer to the trial court’s findings of fact on nonconstitutional questions so long as they are supported by substantial evidence. (Ibid.)

In

Krinsky, an officer of a Florida corporation brought suit for libel and interference with advantage against the anonymous authors of “scathing” posts on a financial message board. (Krinsky, supra, 159 Cal.App.4th at p. 1158.) To determine their identities, she subpoenaed records from the service provider who hosted the board. The trial court ordered disclosure, and one of the anonymous posters petitioned this court for a writ. After reviewing various approaches adopted by other courts in similar situations, we concluded that the plaintiff had to satisfy two requirements to overcome the defendant’s constitutional right to preserve his or her anonymity. First, if the defendant has not received notice of the attempt to lift the shield of anonymity, the plaintiff must make reasonable efforts to provide such notice. (Id. at p. 1171.) Second, the plaintiff must “make a prima facie showing that a case for defamation exists” (ibid.; see id. at p. 1172), by “setting forth evidence that a libelous statement has been made” (id. at p. 1172, fn. omitted). We described the required quantum of evidence as “ ‘that which will support a ruling in favor of [the plaintiff] if no controverting evidence is presented. [Citations.] It may be slight evidence which creates a reasonable inference of fact sought to be established but need not eliminate all contrary inferences. [Citation.]’ ” (Id. at p. 1172, fn. 14, quoting Evans v. Paye (1995) 32 Cal.App.4th 265 , 280, fn. 13 [ 37 Cal.Rptr.2d 915 ].)

Although the present case does not sound in libel, we see no reason to doubt that the same principles apply. In any action predicated on anonymous

In addi

tion to the requirements laid out in Krinsky, we believe it is necessary to require that a plaintiff seeking to discover the identity of an anonymous speaker first clearly specify the statements claimed to be actionable, state the actionable meanings assertedly conveyed by them, and set forth, if necessary, evidence sufficient to sustain a finding that the statements were capable of conveying those meanings. As we noted in Krinsky, at least one court had adopted a rule requiring the plaintiff to “set forth the specific statements that are alleged to be actionable.” (Krinsky, supra, 159 Cal.App.4th at p. 1167, citing Dendrite Internat. Inc. v. John Doe No. 3 (2001) 342 N.J.Super. 134 [ 775 A.2d 756 ]; see also Quixtar Inc. v. Signature Management Team, LLC (D.Nev. 2008) 566 F.Supp.2d 1205 , 1216, fn. omitted [“Plaintiff should not be afforded discovery regarding the identity of any anonymous author where the exact statement at issue has not been put into evidence”].) Another court, we observed, had deemed such a requirement superfluous because, under the law of that jurisdiction, the offending statements in a libel case “must be quoted in the plaintiff’s complaint.” (Krinsky, supra, at p. 1169, citing Doe v. Cahill (Del. 2005) 884 A.2d 451 , 461.) In defamation cases California follows a similar pleading rule, under which “the words constituting an alleged libel must be specifically identified, if not pleaded verbatim, in the complaint.” (Kahn v. Bower (1991) 232 Cal.App.3d 1599 , 1612, fn. 5 [ 284 Cal.Rptr. 244 ]; see Vogel v. Felice (2005) 127 Cal.App.4th 1006 , 1017, fn. 3 [ 26 Cal.Rptr.3d 350 ]; Gilbert v. Sykes (2007) 147 Cal.App.4th 13 , 31-32 [ 53 Cal.Rptr.3d 752 ] [description of allegedly defamatory statements was “a paradigm of vagueness, and does not even come close to the specificity required to state an actionable libel claim”]; 5 Witkin, Cal. Procedure (5th ed. 2008) Pleading, § 739, p. 159 [“the complaint should set the matter out verbatim, either in the body or as an attached exhibit”].)

However, we were not called upon in Krinsky to consider whether the defendants had been adequately apprised of the statements to which the plaintiff objected. We therefore did not discuss such a requirement beyond

We do not believe a plaintiff in MZ’s position is entitled to compel the disclosure of an anonymous poster’s identity without first clearly identifying, on the record, the specific statements claimed to have given rise to liability. It is not sufficient—if in fact it is true—that MZ identified the challenged statements in communications with opposing counsel. It is the court, not counsel, that must determine whether a prima facie showing of actionable statements has been made. It is impossible to perform such a task without knowing the exact statements on which liability is predicated. Moreover, if it is not obvious from the face of the statements that they indeed conveyed an actionable meaning, the plaintiff must clearly specify the meaning it contends was conveyed by them, and any extrinsic facts necessary to lend them that meaning—the equivalents, respectively, of what are known in defamation as the “innuendo,” or defamatory gist or sting, and the “inducement,” or factual context necessary to render facially neutral words defamatory. (See 5 Witkin, Summary of Cal. Law (10th ed. 2005) Torts, § 555, pp. 811, 812.)

MZ’s entire showing of a disclosure of confidential information consisted of the following averments by one of its in-house attorneys: “[T]he review disclosed the scope of Machine Zone’s investment in proprietary technology Machine Zone is developing for use outside of its public product, [Game of War], The review disclosed information concerning the development stage of that technology. It also disclosed the number of personnel working on that project. The review also contains certain quotes and statements attributed to

These averments were too vague and conclusory to satisfy MZ’s burden of making a prima facie case of a breach of the nondisclosure agreement. They are far less specific than averments that have elsewhere been found insufficient to sustain the analogous contention that a defendant has misappropriated a trade secret. (See In re Providian Credit Card Cases (2002) 96 Cal.App.4th 292 , 305 et seq. [ 116 Cal.Rptr.2d 833 ]; Cypress Semiconductor Corp. v. Maxim Integrated Products, Inc. (2015) 236 Cal.App.4th 243 , 263 [ 186 Cal.Rptr.3d 486 ] (Cypress).) The lack of specificity cannot be justified, as MZ suggested below, by the supposed sensitivity of the information in question. As counsel for Glassdoor points out, MZ could have sought leave to file such materials under seal, as it did with the review itself. Any notion that a specification of objectionable statements was too sensitive to be entrusted to a sealed court record is belied by counsel’s professed communication of such a specification to opposing counsel by e-mail.

The vagueness with which MZ framed its claims in the face of Glassdoor’s repeated demands for specificity is redolent with the possibility that greater specificity might disclose not valuable secrets but a lack of merit in the claims themselves. In Cypress, we described approvingly another case holding that ‘“a party was not entitled to describe a trade secret so vaguely as to amount to an open-ended work in progress.” (Cypress, supra, 236 Cal.App.4th at pp. 270-271, citing Perlan Therapeutics, Inc. v. Superior Court (2009) 178 Cal.App.4th 1333 , 1350 [ 101 Cal.Rptr.3d 211 ].) Yet that is precisely what MZ did here.

After not revealing the substance of its claims, MZ added insult to injury by accusing Glassdoor of ”fail[ing] to understand the confidential information at issue” and ”miss[ing] the point” MZ had itself not identified. At the hearing on the motion to compel, after counsel for Glassdoor noted that the only specific technology mentioned in the review was a matter of public record, counsel for MZ responded, ”[T]here are things that are in that post that are confidential that are different than what he was just talking about. Many things .... [T]he fact that they don’t appreciate what is confidential in there doesn’t mean it’s not our confidential information.”

The trial court tolerated this approach apparently because, once MZ moved to place the entire review under seal, both attorneys were barred from

This was an intolerable mode of proceeding. In the first place, the request to seal the review—and the order granting it—swept far too broadly. Much of Doe’s review obviously implicates no confidential information, describing such conditions of employment as workload, office sizes, perquisites of employment, and “work-life balance.” MZ correctly concedes that employees are entitled by statute to publicize their complaints about such matters. (See Lab. Code, § 232.5.) The record therefore provided no basis to seal the review in its entirety, nor for the trial court’s boilerplate findings that the sealing order was “narrowly tailored” and that “[n]o less restrictive means exist[ed] to achieve and further Machine Zone’s interests in the confidential information.” (See Cal. Rules of Court, rule 2.551(a); H.B. Fuller Co. v. Doe (2007) 151 Cal.App.4th 879 , 894-899 [ 60 Cal.Rptr.3d 501 ]; NBC Subsidiary (KNBC-TV), Inc. v. Superior Court (1999) 20 Cal.4th 1178 , 1216-1218 [ 86 Cal.Rptr.2d 778 , 980 P.2d 337 ]; Press-Enterprise Co. v. Superior Court of Cal. (1984) 464 U.S. 501 , 510 [ 78 L.Ed.2d 629 , 104 S.Ct. 819 ] [findings in support of closed proceedings must identify “overriding interest” justifying the closure].) Nor did the court appear to take note of the requirements that a sealing order: “(A) Specifically state the facts that support the findings', and [¶] (B) Direct the sealing of only those documents and pages, or, if reasonably practicable, portions of those documents and pages, that contain the material that needs to be placed under seal. All other portions of each document or page must be included in the public file.” (Cal. Rules of Court, rule 2.550(e)(1), italics added.)

The combination of MZ’s refusal to specify its claims and its grossly overbroad motion to seal rendered it almost impossible for Glassdoor to effectively defend against the subpoena other than by protesting, as it did, that the very vagueness of MZ’s showing precluded an order compelling the requested discovery. A party to litigation is entitled to adequate notice of the

B. Alleged Disclosures

1. Existence of Platform Team

Although MZ failed in the trial court to adequately specify the allegedly actionable statements (see pt. IV., ante), it has been more forthcoming in its filings in this court, in which it states that the offending statements are those referring to MZ’s “platform team.” Doe made three such statements:

“Terrible work-life balance, except for the platform team, which do not know what to work on.”

“The senior management lost directions. The company has invested heavily in the platform team (there are 70-80 engineers). However, after one year, nothing has been done by that team.”

“The CEO said in the team meeting: I don’t expect products and revenue from the platform team. I only want you can show demos. The platform is only for attracting investments from VCs.” (Italics added.)

According to MZ, these statements revealed that MZ was in the process of developing a technology which was then unknown to the world but which it has now unveiled as “a real time data transmission platform branded RTplatform™.” Prior to its “rebrand[ing],” it asserts, its “public-facing business had been solely mobile gaming apps. But on April 4, 2016, Machine Zone publicly announced that it had developed a stand-alone real-time platform technology that enables the exchange of data between billions of endpoints worldwide virtually simultaneously.” It is this new “real-time platform technology,” according to MZ, that Doe prematurely disclosed by his statements about MZ’s “platform team.”

We do not cite these web pages as affirmative evidence of the understanding a reader would actually form based on Doe’s use of the term “platform.” Rather they serve to highlight MZ’s critical failure to make any showing on that subject. The record provides no basis whatever for a finding that Doe’s mere allusions to a platform team would tell the public anything more than that a group of MZ workers were charged with developing the infrastructure for future games, or perhaps for some application or family of applications of unknown type. We fail to see how this vague information could be news, let

In sum, the record contains no support for a finding that Doe’s mere references to a “platform team” would tell the public anything it had not heard from MZ’s own CEO.

2. Size of Team

MZ emphasizes Doe’s statement that “[t]he company has invested heavily in the platform team (there are 70-80 engineers).” While it may be possible that this was confidential business information, MZ again failed to present any evidence to that effect. This is more than a technical point, for it is also possible that this or equivalent information was accessible to the public by, for instance, monitoring MZ’s job listings. (See Cypress, supra, 236 Cal.App.4th 243 , 250-251, 253, 263.) It is conceivable that the size of the team, coupled with some other knowledge, would tell readers something that was not yet publicly known. However, if that is the case, MZ failed to demonstrate it. Accordingly, MZ failed to make a prima facie showing that Doe breached the nondisclosure agreement by alluding to the number of engineers on the platform team.

The gist of Doe’s account was that whatever the platform team was working on, the CEO told its members to focus their efforts on creating demos to assist in raising venture capital. Doe may have intended to accuse the platform team of generating—and the CEO of ordering it to generate— “vaporware,” i.e., “[a] piece of software or other product for use in computing which, despite being publicized or marketed, either does not exist or has not (yet) been developed commercially.” (Oxford English Dictionary Online <http://www.oed.com/view/Entry/243248> [as of Mar. 10, 2017]; see 2 New Shorter Oxford English Diet. (3d ed. 1993) p. 3546 [“software that as yet exists only in the plans of publicity material of its developers”].) This is consistent with Doe’s statements—not cited by MZ as violations of the disclosure agreement—that “Management spreads unreal information to both outside VC’s and employees,” that everyone is working too hard “except for the platform team, which do not know what to work on,” that “senior management” has “lost direction[],” and that MZ was guilty of “telling the investors and employees the unreal information” and of “cheating.”

Subsequent events suggest that Doe’s interpretation of MZ’s actions and motives was simply wrong. MZ was indeed developing a new product, which it launched—together with a repackaging of its corporate image and mission—a mere 10 months after Doe accused it of accomplishing nothing of substance. The question thus remains: Insofar as an employee’s statement about an employer’s internal activities is untrue, can it ever violate a nondisclosure agreement? We think the answer is obviously negative. The essence of “information,” as that term is used in this context, is “Knowledge communicated concerning some particular fact, subject, or event; that of which one is apprised or told; intelligence, news.” (7 Oxford English Diet. (2d ed. 1989) p. 944, italics added.) Similarly, “confidential” means imparted in confidence, i.e., with the expectation that the matter communicated will not be further disclosed. “Confidential” is derived from “confide,” which means to “[tjell someone about a secret or private matter while trusting them not to repeat it to others.” (Oxford Living Dictionaries Online <https://en.oxforddictionaries.com/definition/confide> [as of Mar. 10, 2017], italics added.) The agreement here clearly used these terms in this sense; the confidentiality provision opens with the recital, “In the course of my Company employment, I will learn of or have disclosed to me various ‘Confidential Information.’ ”

Premo, J., and Grover, J., concurred.