Reset A A Font size: Print

Court of Appeal, First District, Division 4, California.

ZL TECHNOLOGIES, INC., Plaintiff and Appellant, v. DOES 1-7, Defendants and Respondents; GLASSDOOR, INC., Real Party in Interest and Respondent.


Decided: July 19, 2017

Counsel for Plaintiff and Kerr & Wagstaffe LLP, Jasmine Kaur Singh, James Appellant, ZL Technologies, M. Wagstaffe, Michael von Loewenfeldt, and Anna Inc.: P. Chang. Counsel for Defendants and No appearance. Respondents, Does 1-7: Counsel for Real Party in Seubert French Frimel & Warner LLP, William Interest and Respondent, Joseph Frimel. Glassdoor, Inc.: Counsel for Amici Curiae, Public Citizen Litigation Group, Paul Alan Levy and Public Citizen, Inc. and Scott Michelman; Wilson Sonsini Goodrich & Rosati Twitter, Inc. on behalf of Real P.C., Corina I. Cacovean. Party in Interest and Respondent

Plaintiff and appellant ZL Technologies, Inc. (ZL) appeals from an order and judgment dismissing its complaint with prejudice for failure to serve defendants, and from an order denying its motion to compel compliance with the subpoena it served on real party in interest and respondent Glassdoor, Inc. (Glassdoor). ZL contends the trial court erred in denying its motion to compel, prohibiting it from identifying Doe defendants whom it contends anonymously defamed it on Glassdoor's website, and then improperly dismissed its action for failing to serve the same individuals, after denying it access to the information necessary to identify them. We agree and therefore shall reverse the judgment.


According to its complaint, ZL is a California corporation that provides email archiving, eDiscovery, and compliance software and support to businesses throughout the country. Glassdoor operates a website for job seekers on which people may anonymously post information and express opinions regarding current or past employers. Between September 2010 and June 2012, individuals representing themselves as current or former ZL employees posted seven anonymous reviews on Glassdoor's website criticizing ZL's management and work environment. On August 29, 2012, ZL filed a complaint against the individuals who posted the critical reviews, naming them as Doe defendants. The complaint alleged causes of action for libel per se in violation of Civil Code section 45, and online impersonation in violation of Penal Code section 528.5 to the extent any of the defendants was not actually a ZL employee. The following month, ZL served a subpoena on Glassdoor, requesting records identifying and providing contact information for defendants.

Glassdoor objected to the subpoena, among other things contending that: compulsory disclosure of defendants' identities would violate their free speech rights under the First Amendment of the United States Constitution (First Amendment), and their privacy rights under the California Constitution; the posted statements were “protected opinion, patently hyperbolic, not harmful to reputation,” or uncontested statements of fact; Glassdoor's reputation would be harmed if it disclosed defendants' identities; and, under Krinsky v. Doe 6 (2008) 159 Cal.App.4th 1154 (Krinsky), ZL was obligated to make a prima facie showing the statements were libelous before it could compel disclosure. The parties corresponded regarding Glassdoor's objections, but did not reach a resolution.

ZL then filed a motion to compel Glassdoor to comply with the subpoena (motion to compel). The trial court issued a tentative ruling denying the motion to compel, reasoning that defendants had a First Amendment right to remain anonymous, their critical reviews of ZL were “similar to that written on bathroom walls—anonymous, angry, opinionated, and not very reliable,” and it was “unclear” whether ZL, as defendants' former employer, might have alternatives for discovering their identities. After hearing argument from the parties the same day, the trial court took the matter under submission. The following day, it issued an order adopting the tentative ruling. The order recited the trial court's finding that ZL “failed to make a sufficient showing ․ the [defendants] engaged in wrongful conduct causing harm to [ZL].” “In the context of [Glassdoor's] website,” the order stated, defendants' reviews were “primarily opinion and would not be considered reliable by the average person.”

After the trial court issued its order, ZL explored independent methods for identifying defendants, without success. More than a year after ruling on the motion to compel, the trial court issued an order to show cause (OSC) why the case should not be dismissed given ZL's continued failure to serve defendants. At the subsequent OSC hearing, the trial court requested briefing about whether it should retain jurisdiction. ZL filed its motion to retain jurisdiction the following month. Contending it presented a prima facie case of libel (apparently by attaching copies of defendants' reviews to its complaint),1 had taken reasonable steps to identify defendants, and had no remaining alternatives for securing that information, ZL requested renewal of the subpoena to compel Glassdoor to identify defendants, so that it might serve the complaint on them. The trial court denied ZL's request. The following month, after a hearing on the matter, the court dismissed the action with prejudice in light of ZL's failure to serve the defendants.

This timely appeal followed. After the case was fully briefed, we received a request from Public Citizen and Twitter, Inc. to file a brief as amici curiae in support of Glassdoor, which we granted. ZL subsequently filed an answer to amici curiae's brief.


A. The Standard of Review on Appeal

On appeal, ZL challenges the trial court's order and judgment dismissing its complaint for failure to prosecute, and the court's underlying order denying its motion to compel Glassdoor to comply with the subpoena.

The trial court has discretion to dismiss an action for delay in prosecution if “[s]ervice is not made within two years after the action is commenced against the defendant.” (Code Civ. Proc., § 583.420, subd. (a)(1); see also Cal. Rules of Court, rule 3.1340 [requiring a noticed hearing on the issue].) When reviewing a discretionary dismissal, we must presume the decision of the trial court is correct, unless the party challenging the decision shows the trial court abused its discretion. (Howard v. Thrifty Drug & Discount Stores (1995) 10 Cal.4th 424, 443.) An appeal from a discovery order also “normally is reviewed under the deferential abuse of discretion standard. [Citations.]” (Krinsky, supra, 159 Cal.App.4th at p. 1161 [“ ‘the trial court has wide discretion in managing discovery issues' ”].)

The trial court's discretion is limited, however, by the applicable legal principles. (People ex rel. Dept. of Corporations v. SpeeDee Oil Change Systems, Inc. (1999) 20 Cal.4th 1135, 1144.) “Thus, where there are no material disputed factual issues, the appellate court reviews the trial court's determination as a question of law. [Citation.]” (Ibid.) Here, the trial court grounded its ruling denying ZL's motion to compel on legal conclusions, i.e., that ZL did not make a sufficient showing defendants engaged in wrongful conduct causing ZL harm, and that defendants' Glassdoor reviews of ZL were “primarily opinion” that an average person would not consider reliable. In reaching these conclusions, the trial court referenced the constitutional principle that an “author's decision to remain anonymous ․ is an aspect of the freedom of speech protected by the First Amendment.” (McIntyre v. Ohio Elections Com'n (1995) 514 U.S. 334, 342.)

“Thus, in this case, we need not defer to a trial court's resolution of disputed facts and inferences. Instead, we are concerned with the legal significance of the undisputed facts in the record.”2 (People ex rel. Dept. of Corporations v. SpeeDee Oil Change Systems, Inc., supra, 20 Cal.4th at p. 1144.) We, therefore, review the trial court's exercise of its discretion in denying the motion to compel compliance with the subpoena “as a question of law in light of the pertinent legal principles.” (Ibid.; see, e.g., California Farm Bureau Federation v. State Water Resources Control Bd. (2011) 51 Cal.4th 421, 436 [appellate courts independently review questions of law]; Krinsky, supra, 159 Cal.App.4th at p. 1161 [appellate courts independently review “whether a particular communication falls outside the protection of the First Amendment”].)3

B. The Test for Compulsory Disclosure of an Anonymous Speaker's Identity

“[T]his case presents a conflict between a plaintiff's right to employ the judicial process to discover the identity of an allegedly libelous speaker and the speaker's First Amendment right to remain anonymous.” (Doe 2 v. Superior Court (2016) 1 Cal.App.5th 1300, 1310 (Doe 2).) Neither the United States nor the state Supreme Court has established a standard for resolving this conflict. In California, however, after surveying case law from this and other jurisdictions, our colleagues in the Sixth Appellate District, in Krinsky, supra, “agree[d] with those courts that have compelled the plaintiff to make a prima facie showing of the elements of libel” to obtain compulsory disclosure of a defendant's identity. (Krinsky, supra, 159 Cal.App.4th at p. 1172.)

ZL and Glassdoor both cite Krinsky's requirement of a prima facie showing as providing the appropriate test in deciding ZL's motion to compel compliance with its subpoena. Amici curiae, however, urge us to go further than Krinsky, and to require application of the multifactor test the New Jersey appellate court articulated in Dendrite Intern. v. Doe No. 3 (2001) 342 N.J.Super. 134, 775 A.2d 756 (Dendrite). In particular, amici advocate requiring a final balancing test following any prima facie showing. For reasons discussed below, we concur with amici curiae about the need for courts to consider most of the Dendrite factors, but we decline to require the final balancing test, for the reason stated in Krinsky.

1. Krinsky's prima facie requirement

In Krinsky, the Court of Appeal commenced its analysis by acknowledging that “the constitutional right to publish anonymously” has long been recognized as “ ‘an aspect of the freedom of speech protected by the First Amendment.’ [Citations.]” (Krinsky, supra, 159 Cal.App.4th at pp. 1163-1164.) “ ‘Persecuted groups and sects from time to time throughout history have been able to criticize oppressive practices ․ either anonymously or not at all.’ [Citation.] ‘The decision in favor of anonymity may be motivated by fear of economic or official retaliation, by concern about social ostracism, or merely by a desire to preserve as much of one's privacy as possible.’ [Citations.]” (Id. at p. 1163.) The court went on to observe that “the relative anonymity afforded by the Internet” (id. at p. 1162), combined with the frequent use of pseudonyms in that forum, presented valuable opportunities for an unhindered exchange of views, but also created some risks. (Id. at pp. 1162-1164.)

On the one hand, the court stated, “[t]he use of a pseudonymous screen name offers a safe outlet for the user to experiment with novel ideas, express unorthodox political views, or criticize corporate or individual behavior without fear of intimidation or reprisal. In addition, by concealing speakers' identities, the online forum allows individuals of any economic, political, or social status to be heard without suppression or other intervention by the media or more powerful figures in the field.” (Krinsky, supra, 159 Cal.App.4th at p. 1162.) On the other hand, the court noted, the environment of “informal debate and criticism” that is created leads “many to substitute gossip for accurate reporting and often to adopt a provocative, even combative tone,” “ ‘heighten[ing] this sense that “anything goes.” ’ [Citation.]” (Id. at p. 1163.) “[C]riticism on the Internet is often so recklessly communicated that the harm to its targets, particularly in the financial arena, may extend far beyond what is covered by rules applicable to oral rhetoric and pamphleteering.” (Id. at p. 1164.) “When vigorous criticism descends into defamation,” Krinsky cautioned, “constitutional protection is no longer available.” (Ibid.; see, e.g., Ashcroft v. Free Speech Coalition (2002) 535 U.S. 234, 245-246 [the First Amendment's speech protections do not extend to defamation]; Gertz v. Robert Welch, Inc. (1974) 418 U.S. 323, 340 [“there is no constitutional value in false statements of fact”; “[n]either the intentional lie nor the careless error materially advances society's interest in ‘uninhibited, robust, and wide-open’ debate on public issues”].)

The court concluded that “[c]orporate and individual targets of these online aspersions may seek redress by filing suit against their unknown detractors.” (Krinsky, supra, 159 Cal.App.4th at p. 1164.) To serve their complaint, plaintiffs may then seek disclosure of those detractors' identities. When this occurs, the anonymous Internet speakers' First Amendment rights must be balanced against a libel plaintiff's right to prosecute its case. (Id. at p. 1165.) Most of the state and federal courts that had addressed the subject, Krinsky observed (including Dendrite, supra, 775 A.2d at pp. 768-772), had agreed this balancing necessitated a prima facie showing of the elements of libel. (Krinsky, supra, at p. 1171; see also id. at pp. 1167-1171 [discussing case law in other jurisdictions, including Dendrite].) Krinsky concurred with this approach and adopted the same requirement. (Id. at pp. 1171-1172.)

The court went on to define the prima facie standard as requiring evidence “ ‘that ․ 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 [the] fact sought to be established but need not eliminate all contrary inferences. [Citation.]’ [Citations.]” (Krinsky, supra, 159 Cal.App.4th at p. 1172, fn. 14.) To meet this standard in seeking compulsory disclosure of an anonymous Internet speaker's identity, the court added, “[a] plaintiff need produce evidence of only those material facts that are accessible to [it],” for example, evidence of the allegedly libelous statement, its falsity, and its effect on the plaintiff. (Id. at p. 1172.) Requiring a prima facie showing, Krinsky reasoned, “ensures that the plaintiff is not merely seeking to harass or embarrass the speaker or stifle legitimate criticism.” (Id. at p. 1171.)

Without directly discussing it, Krinsky also adopted the second factor of Dendrite's prima facie showing requirement, namely, the existence of a legally valid claim. (Dendrite, supra, 775 A.2d at p. 760 [To make a prima facie showing, a plaintiff must “establish[ ] that its action can withstand a motion to dismiss for failure to state a claim upon which relief can be granted ․ [and] produce sufficient evidence supporting each element of its cause of action”].)4 Explaining its decision to adopt the prima facie requirement, Krinsky stated, “When there is a factual and legal basis for believing libel may have occurred, the writer's message will not be protected by the First Amendment.” (Krinsky, supra, 159 Cal.App.4th at p. 1172, italics added.) In its analysis, the court specifically focused on whether the plaintiff had stated a legally valid defamation cause of action, as the defendant contended the speech was nonactionable opinion. (Id. at pp. 1173-1178.) Concluding the online speech, “while unquestionably offensive ․ , did not constitute assertions of actual fact and therefore [was] not actionable”—i.e., that the plaintiff had not pled a viable cause of action for defamation—the court ruled the subpoena to discover the defendant's identity should have been quashed. (Id. at p. 1179.)

As noted, although ZL opposes adoption of the entire Dendrite test, it does not challenge Krinsky's requirement that it make a prima facie showing. (Krinsky, 159 Cal.App.4th at pp. 1172-1179.) We agree with Krinsky's approach and adopt the same prima facie showing requirements, i.e., a plaintiff seeking compulsory disclosure of an allegedly libelous speaker's identity must state a legally sufficient cause of action against the defendant, and must make a prima facie showing of the elements of that cause of action.

2. Additional factors to be considered

Amici curiae urge us to go beyond Krinsky, by requiring application of the complete Dendrite test in deciding requests for compulsory disclosure of an Internet speaker's identity in a libel case. As recounted in Krinsky, in Dendrite “a corporation alleged defamation by multiple Doe defendants on [an Internet] message board and then sought expedited discovery in order to learn their identities. The New Jersey appellate court set forth a four-part test to ensure that plaintiffs do not use discovery to ‘harass, intimidate or silence critics in the public forum opportunities presented by the Internet.’ [Citation.] First, the plaintiff must make an effort to notify the anonymous poster that he or she is the subject of a subpoena or application for a disclosure order, giving a reasonable time for the poster to file opposition. The plaintiff must also set forth the specific statements that are alleged to be actionable. Third, the plaintiff must produce sufficient evidence to state a prima facie cause of action. If this showing is made, then the final step should be undertaken: to balance the strength of that prima facie case against the defendant's First Amendment right to speak anonymously. [Citation.] In Dendrite, the appellate court affirmed the trial court's denial of the discovery application, as the corporate plaintiff had failed to produce evidence” of one of the elements of libel and, thus, failed to make the requisite prima facie showing. (Krinsky, supra, 159 Cal.App.4th at p. 1167.)

Amici curiae also cite Doe v. Cahill (Del. 2005) 884 A.2d 451 (Cahill), another seminal case in which the Delaware Supreme Court adopted portions of the Dendrite test. Krinsky considered Cahill as well, noting the court there adopted a standard applicable to a plaintiff opposing summary judgment. (Krinsky, supra, 159 Cal.App.4th at p. 1169, citing Cahill, supra, 884 A.2d at p. 460.) Under that standard, Krinsky recounted, “the plaintiff ‘must support his defamation claim with facts sufficient to defeat a summary judgment motion.’ ․ [Citation.] The second Dendrite requirement, that the plaintiff set forth the exact statements alleged to be defamatory, was unnecessary [according to Cahill,] because those statements must be quoted in the plaintiff's complaint to avoid summary judgment. The fourth Dendrite step, the balancing of the defendant's First Amendment rights against the strength of the plaintiff's case, was also unnecessary [Cahill concluded,] because ‘[t]he summary judgment test [was] itself the balance. The fourth requirement add[ed] no protection above and beyond that of the summary judgment test and needlessly complicate[d] the analysis.’ ” (Krinsky, supra, 159 Cal.App.4th at p. 1169, quoting Cahill, supra, 884 A.2d at p. 461.) Cahill “did, however, endorse the first element of the Dendrite test,” Krinsky observed, requiring “that the plaintiff make reasonable efforts to notify the anonymous poster about the subpoena or request for a disclosure order and give the defendant a reasonable opportunity to respond. The Cahill court even required the plaintiff to publish that notice on the same message board where the allegedly defamatory statement appeared.” (Krinsky, supra, 159 Cal.App.4th at p. 1169.)

Even if this court, like Cahill, determines the final balancing test is unnecessary, amici curiae contend, the other Dendrite factors present the minimal protections required by the First Amendment. ZL urges us to reject Dendrite's notice and final balancing test factors, citing decisions from other states that, it maintains, required only a prima facie showing. (But see Independent News v. Brodie (Md. 2009) 966 A.2d 432, 456 [cited in ZL's brief, adopting the Dendrite test].)5 Having already adopted Dendrite's prima facie showing requirements, we consider the other three Dendrite factors in turn, below.

a. Notice

The first Dendrite requirement is that the plaintiff attempt to notify the anonymous Internet poster that he or she is the subject of a subpoena or application for a disclosure order, giving a reasonable time for the poster to file opposition. (Dendrite, supra, 775 A.2d at p. 760.) Two California appellate panels recently have adopted a notice requirement, citing Krinsky. (Glassdoor, Inc. v. Superior Court (2017) 9 Cal.App.5th 623, 634 (Glassdoor); Doe 2, supra, 1 Cal.App.5th at p. 1311.) Although the anonymous Internet speaker whose speech rights Krinsky addressed had received notice of the action through the sponsor of the message board where the speech appeared, the court agreed in principle that imposing a notice requirement “[did] not appear to be unduly burdensome.” (Krinsky, supra, 159 Cal.App.4th at p. 1171.)

The Krinsky court also “recognize[d], however, that an Internet Web site, chat room, or message board may no longer exist or be active by the time the plaintiff brings suit; consequently, it would be unrealistic and unprofitable to insist, as did the Cahill court, that a plaintiff ‘post a message notifying the anonymous defendant of the plaintiff's discovery request on the same message board where the allegedly defamatory statement was originally posted.’ [Citation.]” (Krinsky, supra, 159 Cal.App.4th at p. 1171.) “The posting requirement is also ‘more idealistic than practical,’ ” Krinsky continued, as “ ‘a wronged plaintiff is unlikely to want to keep a false assertion alive by inviting continued debate.’ [Citation.]” (Id. at p. 1171, fn. 11.) “Moreover, when ISPs and message-board sponsors ․ themselves notify the defendant that disclosure of his or her identity is sought, notification by the plaintiff should not be necessary. And ․ where the defendant is moving to quash the subpoena, the notification requirement benefits no one ․ [as] [o]bviously [the defendant] has already learned of the subpoena.” (Id. at p. 1171.)

Plaintiff cites the practical obstacles discussed in Krinsky, and also submits it should not be required to “exacerbate its own injury” by posting notice of its subpoena on Glassdoor where prospective employees would see it. If notice is required, it contends, the trial court should direct the subpoenaed party to provide it. Amici curiae appear to agree this is often the simplest course, as have state courts in other jurisdictions. (See, e.g., Solers, Inc. v. Doe (D.C.Ct.App. 2009) 977 A.2d 941, 954 (Solers) [“it often will be simpler and more effective to require the recipient of the subpoena (who likely knows the identity of the anonymous defendant, or at least knows how to contact him)” to notify defendants]; Pilchesky v. Gatelli (Pa.Super. 2011) 12 A.3d 430, 442 [“a reviewing court should inquire of the proprietor of the website to determine the most effective means of notification”].) We are persuaded by the foregoing authority that, to balance properly a plaintiff's reputational interests with the First Amendment rights of anonymous defendants, a court must ensure reasonable efforts are made to notify the defendants, permitting them an opportunity to respond, before disclosure of their identities may be compelled. (See also, e.g., The Mortgage Specialists, Inc. v. Implode-Explode Heavy Industries, Inc. (2010) 160 N.H. 227, 999 A.2d 184, 193 [endorsing the Dendrite test, including the notice requirement]; Mobilisa, Inc. v. Doe (2007) 217 Ariz. 103, 170 P.3d 712, 719 [endorsing the Cahill test, including the notice requirement].) As the District of Columbia's Court of Appeal did in Solers, supra, 977 A.2d at p. 955, “[w]e leave it to the trial court to determine in the circumstances of each case who should notify the anonymous defendant of the efforts to discover his identity,” while observing that Glassdoor here acknowledged it had email and ISP addresses for defendants.6

b. Specific statements

The second Dendrite requirement is that the plaintiff “identify and set forth the exact statements ․ that [it] alleges constitute[ ] actionable speech.” (Dendrite, supra, 775 A.2d at p. 760.) As Krinsky noted, the Cahill court considered this requirement unnecessary “because, under the law of that jurisdiction, the offending statements in a libel case ‘must be quoted in the plaintiff's complaint.’ [Citations.]” (Glassdoor, supra, 9 Cal.App.5th at p. 635, citing Krinsky, supra, 159 Cal.App.4th at p. 1169.) “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.’ [Citations.]” (Glassdoor, supra, 9 Cal.App.5th at p. 635.) Accordingly, adoption of the Dendrite test, in this regard, imposes no new burden for a defamation plaintiff.

c. Final balancing test

The final Dendrite requirement, “assuming the court concludes that the plaintiff has presented a prima facie cause of action,” is a balancing test, weighing “the defendant's First Amendment right of anonymous free speech against the strength of the prima facie case presented and the necessity for the disclosure of the anonymous defendant's identity to allow the plaintiff to properly proceed. ․ The guiding principle is a result based on a meaningful analysis and a proper balancing of the equities and rights at issue.” (Dendrite, supra, 775 A.2d at pp. 760-761.)

Amici curiae contend the test is comparable to that applied in deciding a petition for preliminary injunction, which is appropriate, they contend, because an order compelling disclosure amounts to an injunction. (See, e.g., People v. Brewer (2015) 235 Cal.App.4th 122, 135 [a mandatory injunction “may compel the performance of an affirmative act”]; Drakes Bay Oyster Company v. California Coastal Commission (2016) 4 Cal.App.5th 1165, 1171-1172 [a court deciding an application for a preliminary injunction “evaluates ․ the likelihood plaintiff will prevail on the merits at trial” and the relative interim harm to the parties of granting or denying the application].) Applying a balancing test, amici contend, would allow defendants to show disclosure of their identities “may expose them to significant danger of extra-judicial retaliation,” for example, in the form of negative economic consequences within their industry on being identified as employees who publicly criticized their employer. (See, e.g., Cahill, supra, 884 A.2d at p. 457 [“there is reason to believe that many defamation plaintiffs bring suit merely to unmask the identities of anonymous critics,” “ ‘the primary goal being to silence John Doe and others like him’ ”].) Against these considerations, amici submit, a trial court should balance the strength of plaintiff's case, the nature of plaintiff's allegations (e.g., the type of speech at issue, see, e.g., Kasky v. Nike, Inc. (2002) 27 Cal.4th 939, 946 [“commercial speech receives a lesser degree of constitutional protection”] ), the likelihood that the alleged libel significantly damaged plaintiff, and the extent of plaintiff's own responsibility “for the problems of which [it] complains.” (See also Mobilisa, supra, 170 P.3d at pp. 720-721 [adoption of a balancing test, among other things, allows consideration of a party's relative need for disclosure, and the availability of alternative discovery methods].)

ZL responds that application of the Dendrite balancing test is “wholly unjust” because it would allow denial of discovery, effectively precluding a plaintiff from proceeding with its case, even after it successfully shows a prima facie case of injury. This court should follow Krinsky, ZL contends, by ruling that, “[w]hen there is a factual and legal basis for believing libel may have occurred, the writer's message will not be protected by the First Amendment.” (Krinsky, supra, 159 Cal.App.4th at p. 1172; accord, Solers, supra, 977 A.2d at p. 956 [“a separate balancing test at the end of the analysis is not necessary”]; Cahill, supra, 884 A.2d at pp. 460-461 [separate balancing test unnecessary because requiring a plaintiff to support a defamation cause of action “with facts sufficient to defeat a summary judgment motion”—i.e., a prima facie showing—“is itself the balance”].) We concur with Krinsky that a further balancing should not be required “[w]here it is clear to the court that discovery of the defendant's identity is necessary to pursue the plaintiff's claim,”7 and the plaintiff makes a prima facie showing that a libelous statement has been made.8 (Krinsky, supra, 159 Cal.App.4th at p. 1172, italics added, citing Beauharnais v. People of State of Ill. (1952) 343 U.S. 250, 266 [“libelous utterances” are not constitutionally protected speech]; Gerawan Farming, Inc. v. Lyons (2000) 24 Cal.4th 468, 485 [First Amendment right to freedom of speech is not absolute]; Aguilar v. Avis Rent A Car System, Inc. (1999) 21 Cal.4th 121, 134 [same].)

C. Applying the Test to ZL's Subpoena

Having decided the proper analytic framework for deciding a motion to compel compliance with a subpoena in this context, we independently review the record to determine whether the trial court erred as a matter of law in denying ZL's motion.

1. Legally Valid Cause of Action

The trial court denied ZL's motion to compel compliance with the subpoena because it concluded defendants' critical reviews on Glassdoor qualified as protected opinion and, therefore, legally could not provide the basis for a defamation cause of action. ZL contends the trial court erred as a matter of law because the reviews contained statements that either directly asserted or clearly implied provably false factual assertions. Glassdoor disagrees, maintaining all of the statements were nonactionable opinions. Amici curiae concur that many of the alleged statements were protected opinions, although, as will be discussed, they concede a few statements did include factual assertions providing a basis for a legally valid defamation cause of action.

a. Background

i. Glassdoor's website

According to the declaration of Glassdoor's General Counsel and Senior Vice President, Allyson Willoughby, the website glassdoor.com is intended to “help job seekers make informed decisions about the companies at which they are considering working.” The website explains it “is a free jobs and career community that offers the world an inside look at jobs and companies. What sets us apart is our ‘employee generated content’—anonymous salaries, company reviews, interview questions, and more—all posted by employees, job seekers, and sometimes the companies themselves,” giving users “all the information [they] might need to make [their] next career decision. [¶] ․ [¶] No other career or jobs site offers such detailed information about specific jobs at specific companies. ․”

The “Terms of Use Agreement for Glassdoor Services” provides that users agree they will not post any “false or misleading” content, and users “represent ․ any information [they] provide in a ․ Company Review ․ is correct.” Glassdoor also advises it does not control “[c]ontent from other Users, advertisers, and other third parties” and “make[s] no guarantees about the accuracy ․ or quality of the information in such [c]ontent.”

A person posting a company review on Glassdoor is asked to rate the company as an employer “on a scale of one through five, both on an overall basis and with respect to a number of categories such as ‘Culture & Values' and ‘Compensation & Benefits.’ ” The person then “lists the ‘pros' and ‘cons' of working” for the company, and offers “ ‘advice to senior management.’ ”

ii. The Doe defendants' Glassdoor reviews of ZL

ZL bases its action on the following seven reviews that Doe defendants posted about it on Glassdoor's website:

1. The first review (first review), titled “Great Product, Very Poor Management,” allegedly posted on September 21, 2010 by a Doe defendant identified as a “Current Employee in San Jose, CA,” provided in full as follows: “Pros - Best product in the market. [¶] Broad responsibilities, rewards initiative and creativity, great exposure to the market, collegiate [sic] atmosphere, competent technical staff with good product experience. [¶] Cons - This is a lifestyle business for senior management. No transparency or accountability for their decisions. This company is great for hands-on experience, very poor for mentoring or professional development. Despite a great product, management lacks the experience [or] focus to drive growth. [¶] Advice to Senior Management - Bring in experienced middle management so the executive team can focus on strategy.”

2. The second review (second review), titled “Internecine strife, nepotism, and mismanagement = very unhappy employees,” allegedly posted on April 13, 2011 by a Doe defendant identified as “Former N/A,” provided in full as follows: “Pros - Great product, industry leader, moving ahead despite internal issues. Technical, marketing, and engineering staff do put a concerted effort into doing their best for the product and the customer. [¶] Cons - The company hires graduates from UPenn almost exclusively, and the practice is to hire people who are fresh out of school with no experience or degree in a related field (so as to allow justification for lower pay). Because of the ‘young blood’ employees, standards are set egregiously low for professionalism, communication, attire, adherence to organizational standards. No organizational chart, job title, or job descriptions exist in this company. It's every man/woman for themselves. If you like to engage with your job like it's a bloodsport in a Roman Coliseum, you'll love your job. Also, pay is 30-50% lower than industry standards in Silicon Valley. Benefits are deplorable. Absolutely no perks or incentives for employees other than fear of castigation or being fired. As a result, there is a high turnover rate, widespread low morale, and minimal room for mentorship or growth. No one who is any good (and not dependent on immigration law (work visas) stays. [¶] Advice to Senior Management - People talk about middle management as if it is a bad thing; at ZL, you see what happens when there is no middle management whatsoever. The CEO, CTO, and the CEO's wife (head of HR, imagine that) manage by implementing a culture of fear, disrespect, and mistrust. There is no management in this company of approx. 60 people other than the CEO & CTO. Hire some competent, efficient managers that are not blood related or academically affiliated to other staff. Support them in doing their job so you can do yours. Focus on the broad vision and hire new trustworthy folks to take over your more cumbersome roles – people with a minimum of 10-20 years experience to help set a more professional tone. Your company WILL NOT grow if you keep a vice [sic] grip on it; customer dissatisfaction will increase and employee turnover will continue to skyrocket as things are now.”

3. The third review (third review), entitled “An effective guide for the mismanagement of any organization,” allegedly posted on April 26, 2011 by a Doe defendant identified as “Former Employee,” provided in full as follows: “Pros - Working at ZL will enable you to take on projects and experience you otherwise would not have as a first-year employee elsewhere due to the high turnover and lack of senior employees. [¶] Collegiate/young atmosphere means that everyone under management generally gets a long [sic] well. [¶] First class talent pool despite poor management. [¶] Cons - Nontransparent – Management runs the company without transparency despite the extremely small size of the company. [¶] Nepotism – Management is composed of purely family and school-specific friends. [¶] No Professional Growth – Mentorship and professional development are non-existent. Advancement is not based on any measurable output or metrics. Feedback on performance never happens. Everything is completely subject to CEO's opinion. [¶] No Viable Strategy – Management constantly changes corporate strategy based on the smallest shred of evidence. [¶] Not scalable – No experienced managers to grow the company. Micromanagement means unreasonable demands with little to no support. Resources are scarce to close deals or increase pipeline. [¶] Disrespect – Management does not respect their employees. They belittle them in public, foster a negative atmosphere, and cast aspersions upon them. [¶] Average time for an experienced hire at ZL is 6 months. For new hires out of college, about a year – but only because that is their first job. No one stays if they can help it. [¶] Advice to Senior Management - Although management views that the problem is that they are not involved enough, it is actually the opposite. Management should learn how to trust others and hire experienced managers to develop, maintain, and motivate talent. Specifically on the business side, they need to get out of their own way. [¶] It would be understandable if this type of management had actually led to success, but it has not. Without change, the company will continue to stagnate.”

4. The fourth review (“fourth review”), entitled “A Real Nightmare I'm Trying to Forget,” allegedly posted on April 28, 2011 by a Doe defendant identified as “Former Business Associate in San Jose, CA,” provided in full as follows: “Pros – Solid team – The primary thing that got me and my colleagues going at ZL was the team we worked with. Despite management trying its best to suppress friendship and bonding, we shared information freely among one another, and never threw anyone under the bus. [¶] The ZL products themselves are of high quality, led by a gifted CTO and development team. [¶] If you can grab onto a key project and own it, you will be able to grow with it very fast. This is where a ‘start-up mode’ works to your advantage – you can manage projects from ground up to finish. [¶] Cons – Minimal potential for growth – Unfortunately, management tends to tell potential new hires that ‘1 year at ZL is worth 3 years in the “real world”, and that working in a big company is a waste because you are pigeon-holed and won't get to do anything of value’. While this might be true in a start-up that empowers its employees, unfortunately, it does not [sic] at ZL. Employees are typically given tasks in pieces that may or may not be used, and feedback from management that has no real management experience is ultimately useless. There are no annual reviews, or performance feedback meetings, unless you request them, and then they are usually a meeting of ‘what you did wrong’. [¶] Constantly harassed – ZL used to hire only immigrants with ‘visa issues'. It was clear that the management felt that they were most likely to stay longer, and tolerate more from management. If management did not like someone for any reason, rather than address the matter privately, they publicly humiliate the employee and reduce his/her importance in the company by changing a title, or taking away projects. [¶] Mom-and-pop environment – ZL is essentially run by a CEO and his wife (admin/HR/accounting). There is zero accountability and if you have a problem with anything – there is not much you can do about you [sic]. Also they nickle-and-dime [sic] employees in every way – from benefits to expense reports. The management also fosters unhealthy competition and backstabbing, which he views as ‘natural aggression’ amongst ‘high-flyers'. [¶] High turnover – ZL has been recruiting from different schools over the past few years – usually going after one school, building up a bad reputation, and then moving on to another. New hires quickly report back to the admissions or recruiting teams at their alumni, making it near to impossible for the company to continue recruiting at the same place year-after-year. People with visa issues aside, most new hires make the decision to leave within 3-6 months. [¶] Advice to Senior Management – Unless the ZL management is willing to overhaul its current strategy, or hire a layer of middle management (with full authority, not just title), the future is a little bleak. It will remain a ‘late-stage start-up’ till it runs into the ground. Management has to realize that you can no longer lead only with the stick, but with the carrot [sic], and that people must be managed with compassion and respect. A company is nothing without its people, and ZL is constantly in search of good people.”

5. The fifth review (fifth review), entitled, “Don't Work Here,” allegedly posted on March 15, 2012 by a Doe defendant identified as “Former Employee in San Jose, CA,” provided in full as follows: “Pros - The other employees were great – supportive and like a small community. The product was excellent. [¶] Cons - The CEO, Kon Leong, cannot effectively manage the organization and lacks the self awareness to realize it. No respect for other employees. Pitted other employees against one another and frequently attempted to foster a hostile and competitive environment among employees. CEO does not know how to get along with other employees and make them a functional part of the organization. Complete lack of mid-management; hires college kids to do the work of senior level employees. Most employees quit in less than a year. Kon Leong has no respect for anyone else and only thinks of sales as his bottom line. Marketing and other business initiatives are irrelevant and lacks [sic] direction. [¶] Advice to Senior Management - Give up and sell the business. The CEO doesn't know what he is doing and is making other people miserable, some of whom aren't properly informed about the environment prior to joining the organization. The company isn't growing and hasn't for years. There is no future and competitors are quickly taking over the space.”

6. The sixth review (sixth review), entitled “Stay as far away from this place as possible,” allegedly posted on March 20, 2012 by a Doe defendant identified as “Former Employee,” provided in full as follows: “Pros - Every last employee is incredibly sharp. Lots of professional mentorship and personal support from peers and superiors (besides upper management). [¶] Cons - Turnover rate is nearly 50% annually, and the only reason it's not higher is that 50% of employees work on the engineering side and aren't required to interact with the CEO. Those that do leave within one year, if not months. Why? There is absolutely no respect for employees. Before you commit to working at ZL, please strongly consider the following: [¶] Are you willing to be treated as though you are extremely replaceable? To work at ZL, you must be willing to accept the lack of respect to employees that come in the form of tangible benefits (401k not matched, lack of holiday bonuses, bizarre vacation policy, no team-strengthening initiatives or celebrations of successes to build a culture, nickel-and-dimed [sic] to the point where the CEO himself gets involved.) [¶] Are you willing to put up with utterly controlling behavior? To work at ZL, you must be willing to give up all desire to take initiative. Here, having ‘the ability to learn’ translates to ‘the ability to do only what management tells you to do.’ Only one person's idea is [the] right idea, and that is the CEO's. Furthermore, even seasoned sales and marketing veterans (who were brought in to amp up their respective departments) left ZL because upper management prohibited them from putting their knowledge to use. [¶] Are you willing to put up with public disparagement and humiliation? To work at ZL, you must be willing to endure personal attacks (both passive-aggressive and purely aggressive) by the CEO in front of your colleagues. This can come in the form of a company-wide email or a large meeting. [¶] Many salient points are not mentioned above, as other reviewers have touched on them. However, if you decide to work for ZL, you MUST be willing to be treated with complete disrespect. All in all, ZL's environment is unhealthy and makes one wonder if a work environment could possibly be worse. [¶] Advice to Senior Management - Despite the incredible product that the dedicated engineering team supports, ZL will go nowhere if the CEO is not replaced.”

7. Finally, the seventh review (seventh review), entitled “Tragic,” allegedly posted “around June 2012” by a Doe defendant identified as “Former Employee in San Jose, CA,” provided in full as follows: “Pros - There are some great people at ZL. Working with many of the energetic and lively people there was a pleasure. This company is very flat, everyone reports either to the CEO or the CTO, or the office manager who is the CEO's wife. The CTO is a competent engineering leader. [¶] Cons - Other than the CTO the senior leadership is not great. Pennywise and pound foolish is the perfect description of the culture and style that pervades here. In its entire history this company has never managed to keep any non-founding member of the executive team for more than 18 months, not a VP Marketing, not a VP Sales, not a General Counsel. Those execs are not all wrong. The 90% turnover, year over year in sales and marketing is also an indication of trouble emanating from the top. [¶] Advice to Senior Management - Start by being willing to take advice, especially from the people you hired because they are good at their jobs and have been successful.”

b. Relevant Legal Principles

Civil Code section 45 defines libel as “a false and unprivileged publication by writing ․ , which exposes any person to hatred, contempt, ridicule, or obloquy, ․ or which has a tendency to injure him in his occupation.” “A libel which is defamatory of the plaintiff without the necessity of explanatory matter ․ is said to be a libel on its face” (Civ. Code, § 45a), or “libelous per se” (see, e.g., McGarry v. University of San Diego (2007) 154 Cal.App.4th 97, 112). “A corporation can be libeled by statements which injure its business reputation” (Barnes-Hind, Inc. v. Superior Court (1986) 181 Cal.App.3d 377, 381), and “[a] corporation's reputation as an employer is ․ an important aspect of its business reputation” (Di Giorgio Fruit Corp. v. AFL-CIO (1963) 215 Cal.App.2d 560, 571).

“ ‘ “The sine qua non of recovery for defamation ․ is the existence of falsehood.” [Citation.] Because the statement must contain a provable falsehood, courts distinguish between statements of fact and statements of opinion for purposes of defamation liability. Although statements of fact may be actionable as libel, statements of opinion are constitutionally protected. [Citation.]’ [Citation.] That does not mean that statements of opinion enjoy blanket protection. [Citation.] On the contrary, where an expression of opinion implies a false assertion of fact, the opinion can constitute actionable defamation. [Citation.]” (Summit Bank v. Rogers (2012) 206 Cal.App.4th 669, 695-696.) “ ‘The key is not parsing whether a published statement is fact or opinion, but “whether a reasonable fact finder could conclude the published statement declares or implies a provably false assertion of fact.” [Citations.]’ [Citation.]” (Bently Reserve LP v. Papaliolios (2013) 218 Cal.App.4th 418, 427 (Bently Reserve).)

“[I]t is a question of law for the court whether a challenged statement is reasonably susceptible of [a defamatory] interpretation.” (Bently Reserve, supra, 218 Cal.App.4th at p. 428.) In deciding the question, “courts use a totality of the circumstances test. [Citation.] ‘[A] court must put itself in the place of an average reader and determine the natural and probable effect of the statement ․’ [Citation.] Thus, a court considers both the language of the statement and the context in which it is made. [Citations.]” (Id. at p. 427.)

Use of “hyperbolic, informal” (ComputerXpress, Inc. v. Jackson (2001) 93 Cal.App.4th 993, 1013 (ComputerXpress)), “ ‘crude, [or] ungrammatical’ language, satirical tone, [or] vituperative, ‘juvenile name-calling’ ” provide support for the conclusion that offensive comments were nonactionable opinion. (Bently Reserve, supra, 218 Cal.App.4th at pp. 429-430.) Similarly, overly vague statements (ComputerXpress, supra, at p. 1013), and “ ‘generalized’ comments ․ ‘lack[ing] any specificity as to the time or place of’ alleged conduct may be a ‘further signal to the reader there is no factual basis for the accusations.’ ” (Bently Reserve, supra, at p. 431, citing Chaker v. Mateo (2012) 209 Cal.App.4th 1138, 1149-1150 (Chaker) [claims the plaintiff “pick[ed] up streetwalkers and homeless drug addicts and [was] a deadbeat dad” were nonactionable].) On the other hand, if a statement is “factually specific,” “earnest” (Bently Reserve, supra, at p. 433), or “serious” in tone (id. at p. 431), or the speaker “represents himself as ‘unbiased,’ ” “ ‘having specialized’ ” (ibid.) or “ ‘first-hand experience,’ ” or “ ‘hav [ing] personally witnessed ․ abhorrent behavior’ ” (id. at p. 428), this may signal the opposite, rendering the statement actionable (id. at pp. 428-429, 431, 433).

As noted, courts must also consider the context of the allegedly defamatory statements, “ ‘examin[ing] the nature and full content of the particular communication, as well as the knowledge and understanding of the audience targeted by the publication.’ [Citation.]” (Bently Reserve, supra, 218 Cal.App.4th at p. 427.) Courts will consider, for example, whether the statements were posted anonymously on an Internet website, as “ ‘the culture of ․ most electronic bulletin boards ․ encourages discussion participants to play fast and loose with facts,’ ” and the use of anonymity or pseudonyms, “ ‘is a cue to discount [the authors'] statements accordingly.’ [Citation.]” (Summit Bank, supra, 206 Cal.App.4th at pp. 696-697.) “However, the mere fact speech is broadcast across the Internet by an anonymous speaker does not ipso facto make it nonactionable opinion and immune from defamation law.” (Bently Reserve, supra, at p. 429; see id. at pp. 422, 433 [the defendant's negative Internet review of an apartment building, using a pseudonym, “could reasonably be understood as conveying [provable] facts”].) Rather, a defendant's anonymity, the name of the Internet forum, the nature, language, tone, and complete content of the remarks all are relevant. (Id. at pp. 430-431.)

“ ‘[N]ot every word of an allegedly defamatory publication has to be false and defamatory to sustain a libel action ․ “[T]he test of libel is not quantitative; a single sentence may be the basis for an action in libel even though buried in a much longer text ․” ’ [Citation.]” (Balzaga v. Fox News Network, LLC (2009) 173 Cal.App.4th 1325, 1338.)

c. Analysis

Looking at the totality of the circumstances in this case, with one exception discussed below, a reasonable fact finder could conclude all of defendants' Glassdoor reviews contained statements that declared or implied provably false assertions of fact, providing a legally sufficient basis for a defamation cause of action. (Bently Reserve, supra, 218 Cal.App.4th at p. 427.) Both the language and the context of the reviews support this conclusion.

We begin with the shared context of the Glassdoor website. Although, as noted, the fact that statements were published on the Internet using a pseudonym may suggest an average reader should “view them with a certain amount of skepticism” (Summit Bank, supra, 206 Cal.App.4th at p. 696), the name and professed purpose of the Glassdoor website, and the formatting of reviews posted there, created a setting in which the audience reasonably might anticipate receiving a mix of content, including factual statements and opinion. Unlike the Craigslist website “ ‘Rants and Raves,’ ” where offensive speech was posted in Summit Bank, supra, 206 Cal.App.4th at p. 677, for example, the website name “Glassdoor” does not immediately suggest a culture in which users will post content that is “fast and loose with facts” (id. at p. 696), or reflect an “anything goes” mindset (Krinsky, supra, 159 Cal.App.4th at p. 1163). Rather, the name suggests what the website provided, a transparent or open door, a means of ingress or access, for those seeking inside information about jobs and companies. Glassdoor advised readers that its website offered “an inside look at jobs and companies” in the form of “ ‘employee generated content’—anonymous salaries, company reviews, interview questions, and more”—providing job seekers “all the information [they] might need,” “detailed information,” to make “career decision[s].” Although the term “company reviews” suggested employees would share opinions, the website's emphasis on providing “information” also strongly intimated factual material would be included.

We are unpersuaded by Glassdoor's suggestion that the statements included on its website explaining it did not compose reviews, control their content, or guarantee their accuracy, signaled to all readers the reviews should be understood exclusively as opinion, and not relied upon as conveying facts. The suggestion is at odds with the website's noted emphasis on providing “information.” That some of this information would be factual in nature is further underscored by the “Terms of Use Agreement for Glassdoor Services,” which affirmed users had agreed not to “post any ․ false or misleading” content and represented that any information they provided was “correct.”

The Glassdoor-provided structure and content of company reviews also assist in creating the impression that users will provide balanced reviews, combining fact and opinion. Reviewers are asked to list the “ ‘pros' ” and “ ‘cons' ” of employment with a specific company, content that one might reasonably expect would include factual statements. Then, reviewers are asked to rate the company in several categories and provide “ ‘advice to senior management,’ ” endeavors that necessarily will convey opinions.

The language of defendants' reviews supports the conclusion that some statements were intended as factual assertions. While informal in some respects, including use of contractions and simply stated opinions (for example, “Best product in the market,” “Don't Work Here,” “bizarre vacation policy,” “senior leadership is not great”), the language generally adhered to grammatical rules. Each review listed positive points about ZL in the “pros” section (“rewards initiative and creativity,” “[t]he product was excellent,” “[l]ots of professional mentorship and personal support,” “[t]here are some great people at ZL”, “[s]olid team”, “[f]irst class talent pool”), suggesting an intent to present a balanced and unbiased perspective. Although the reviews contained criticism and also some hyperbole and exaggeration (e.g., referring to work at ZL as a “bloodsport” or a “[n]ightmare”), the language was neither vituperative nor crude and included no “ ‘juvenile name-calling.’ ”9 Rather, after identifying themselves as current or former employees or associates, presumably drawing upon first-hand experience, and using language appropriate to a more professional work environment, defendants shared their general assessment of ZL, for example, stating opinions about employee benefits (e.g., “[b]enefits are deplorable”), the work environment (“[m]om-and-pop environment,” management “fosters unhealthy competition and backstabbing”), and company culture (“[p]ennywise and pound foolish” describes “the culture and style that pervades here,” “standards are set egregiously low for professionalism, communication, [and] attire”) that would be relevant for potential job applicants.

In addition to opinion statements, however, with the exception of the first review, discussed further below, each of the remaining six review (the six reviews) included one or more specific factual assertions that could be damaging to a business's reputation, for example, that the company purposefully hired inexperienced personnel, paid below industry standards, publicly disparaged staff, and had high staff turnover rates. Without attempting an exhaustive list, we note the following factual assertions: from the second review, that ZL had a “practice” of “hir[ing] people who are fresh out of school with no expertise,” “no organizational chart, job title, or job description exist[ed] in the company,” and employee pay was “30-50% lower than industry standards in Silicon Valley”; from the third review, that ZL management was “composed of purely family and school-specific friends,” there were “[n]o experienced managers to grow the company,” and management “belittle [d] [employees] in public ․ cast[ing] aspersions upon them”; and from the fourth review, that “most new hires ma[de] the decision to leave [ZL] within 3-6 months,” implying departures in that time frame.

Additionally, in the fifth review, a Doe defendant asserted ZL “hire[d] college kids to do the work of senior level employees” and “[m]ost employees quit in less than a year,” suggesting a practice of hiring inexperienced staff, and assigning them tasks beyond their skill level, causing or contributing to high employee turnover. In the sixth review, a Doe defendant made a direct factual assertion about the company's rate of employee turnover—“nearly 50% annually”—and implied a causal factor, namely, dislike for ZL's CEO—stating, “the only reason [the turnover rate was] not higher [was] that 50% of employees work[ed] on the engineering side and [were not] required to interact with the CEO.” The same review also asserted the CEO publicly disparaged employees (“Are you willing to put up with public disparagement and humiliation?” “To work at ZL, you must be willing to endure personal attacks ․ by the CEO in front of your colleagues”).10 Finally, the seventh review asserted that, “[i]n its entire history,”11 ZL “never managed to keep any non-founding member of the executive team for more than 18 months, not a VP Marketing, not a VP Sales, not a General Counsel. ․ [There is a] 90% turnover, year over year in sales and marketing.” All of the enumerated statements conveyed specific factual assertions, capable of being proved true or false, providing support for a defamation cause of action. (See Balzaga v. Fox News Network, LLC, supra, 173 Cal.App.4th at p. 1338 [“ ‘ “a single sentence may be the basis” ’ ” for a libel action].)12

Defendant's first review is the exception. Although written in a serious tone by a person who self-identified as a “current” ZL employee, the brief review was comprised entirely of generalized comments, most of which undisputedly communicated nonactionable opinions, e.g., “[b]est product in the market,” “competent technical staff,” “management lacks the experience [or] focus to drive growth.” ZL points to only one sentence in this review that it contends conveyed a factual assertion supporting a defamation cause of action: “No transparency or accountability for [senior management's] decisions.” The assertion stated a generalized conclusion, however, without tying it to specific conduct, or a time and place (Bently Reserve, supra, 218 Cal.App.4th at p. 431 [these factors suggest a statement of protected opinion] ), and was preceded and followed by other broadly worded statements of opinion. Viewed in context, the passage read, “This is a lifestyle business for senior management. No transparency or accountability for their decisions. This company is great for hands-on experience, very poor for mentoring or professional development.” Given this context, we conclude the statement is best understood as nonactionable opinion. (See, e.g., Chaker, supra, 209 Cal.App.4th at pp. 1149-1150; Carver v. Bonds (2005) 135 Cal.App.4th 328 (Carver).)

Glassdoor maintains that defendants' use of informal and exaggerated language signaled their reviews were not to be understood as conveying facts. As examples, it notes, among other things, (1) the March 15, 2012 review's pronouncement that ZL's CEO had “no respect for anyone else and only thinks of sales as his bottom line” and (2) the March 20, 2012 review's statements warning job seekers to “[s]tay as far away ․ as possible” from ZL, that they would be “treated with complete disrespect” if hired, and ZL's work environment was “unhealthy ․ mak[ing] one wonder if [it] could possibly be worse.” The reviews were also angry in tone, Glassdoor maintains, suggesting that defendants were simply venting about their experiences and that, taken in context, their remarks could not be reasonably seen as factual.

We are unconvinced. While the six reviews may include exaggerated or angry statements, these facts alone do not compel the conclusion they conveyed no factual assertions. Nor do the cases cited by Glassdoor support such a conclusion here. In Krinsky, for example, in concluding the defendant's Internet remarks about the plaintiff—for example, that she “ ‘[had] fat thighs, a fake medical degree, ․ and poor feminine hygiene’ ”—could not “be interpreted as asserting or implying objective facts,” the court considered the entire context of the defendant's multiple, lengthy online “diatribes,” observing they “convey[ed] scorn and contempt,” were “sarcastic” and “derisive” in tone, engaged in “juvenile name-calling,” used language that was “crude, ungrammatical,” “vulgar and insulting,” and could only be considered an “irrational, vituperative expression of contempt.” (Krinsky, supra, 159 Cal.App.4th at pp. 1175-1177.) Defendants' Glassdoor reviews of ZL do not match this description or even approach it in terms of vitriol. They were grammatical, contained no crude or vulgar statements, were relatively succinct, and focused entirely on ZL's work environment. Neither their context nor their language compels the conclusion they should be discounted as fact-free diatribes.

Similarly, in Summit Bank, supra, the defendant posted multiple “free-flowing diatribes” on a Craigslist website entitled “Rants and Raves.” (Summit Bank, supra, 206 Cal.App.4th at p. 699.) The posts contained spelling errors, poor grammar, and crude and juvenile name calling, complaining, among other things, that defendant's “screwed up Bank” had paid no dividend that year, and “[t]he bitch CEO that runs this Bank thinks that the Bank is her personel [sic] Bank to do with it as she pleases. Time to replace her and her worthless son.” (Id. at p. 679.) In concluding the statements constituted nonactionable opinion (id. at p. 700), the court reasoned the name of the electronic bulletin board alone should have “predisposed [readers] ․ to view [the defendant's statements] with a certain amount of skepticism,” and to understand they likely would “present one-sided viewpoints rather than assertions of provable facts.” (Id. at p. 696.) Additionally, the court observed, the poor spelling and grammar used in the posts, combined with the defendant's “colloquial epithets” strongly suggested he was communicating “his own unsophisticated, florid opinions,” which were not “to be understood as assertions of fact.” (Id. at p. 699.) As we have already described, the context and language employed in defendants' Glassdoor reviews is of a different character.

Two other cases that Glassdoor cites are distinguishable because, unlike in the six reviews discussed above, the speech in the cited cases conveyed “insults [that] are generalized” and “lack[ing] any specificity” (Chaker, supra, 209 Cal.App.4th at p. 1149); see also ComputerXpress, supra, 93 Cal.App.4th at p. 1013 [vague assertions that the plaintiff “only cared about selling stock and not about generating a genuine business”] ), and either “were replete with explicit statements of opinion” (ComputerXpress, supra, 93 Cal.App.4th at p. 1013 [for example, “ ‘IMO [in my opinion],’ ‘what I think is a fraud,’ ‘I firmly believe’ ”] ), or had an “overall thrust”—“[the plaintiff was] a dishonest and scary person”—that was “on its face nothing more than a negative, but nonactionable opinion” (Chaker, supra, 209 Cal.App.4th at p. 1149). In contrast, the six reviews about ZL included, for example, specific factual assertions that the company had inexperienced, underpaid staff, and a high rate of employee turnover, with the latter allegedly caused or exacerbated by the CEO's practice of publicly disparaging staff.

Finally, Glassdoor contends the fact its website contained numerous reviews of ZL, both positive and negative, would indicate to a reasonable reader that ZL's review page was simply a forum for former employees to share conflicting opinions about the company. Again, we disagree. While Glassdoor did invite former employees to share “content” that naturally would include opinions, for example, by rating companies on various topics, it also touted itself as a source of “all the information,” “detailed information,” users might need to make career decisions. A reader would reasonably anticipate this “information” would include facts. Nor did the volume or content of Glassdoor's ZL review page suggest “an ongoing, free-wheeling and highly animated exchange” among reviewers, as apparently was true in the cases Glassdoor cites. (See, e.g., Global Telemedia Intern., Inc. v. Doe 1 (C.D.Cal. 2001) 132 F.Supp.2d 1261, 1267 [1,000 messages a week posted on the Internet chatroom in which the defendants were repeat posters, randomly exchanging views with other individual investors]; see also Krinsky, supra, 159 Cal.App.4th at p. 1175 [“many messages” were posted to electronic bulletin board in two months generating a “[h]eated discussion” about the plaintiff's “credentials and ‘credibility’ ”].) The ZL review page received only 11 reviews over an approximate 30-month period (September 2010 to February 2013). The record contains no suggestion any one individual made repeat posts or that posters engaged in an animated exchange of viewpoints. Rather, each reviewer appears to have provided a complete independent evaluation. In the context of Glassdoor's ZL review page, at a minimum, the highlighted remarks from the six reviews were fairly interpreted as conveying factual assertions, capable of being proved true or false. The trial court, therefore, erred as a matter of law in deeming their content to be entirely protected opinion. The six reviews did include factual assertions providing a legally sufficient basis for ZL's defamation cause of action.

2. Prima Facie Showing

Glassdoor and amici curiae contend the trial court's order nonetheless should be affirmed because ZL did not satisfy the evidentiary component required to make a prima facie showing. Specifically, they submit, ZL did not present evidence the reviews' factual assertions were false.

“As a general matter, a defamation claim does not require a plaintiff to plead or prove falsity or malice.” (Industrial Waste and Debris Box Service, Inc. v. Murphy (2016) 4 Cal.App.5th 1135, 1156; see, e.g., Brown v. Kelly Broadcasting Co. (1989) 48 Cal.3d 711, 753, fn. 37 (Brown) [Generally “ ‘[t]he burden of proof with respect to the issue of truth or falsity is on the defendant’ ”]; Ringler Associates Inc. v. Maryland Casualty Co. (2000) 80 Cal.App.4th 1165, 1180 [“It is the defendant's burden to ․ show the truth of the [offensive] statements”]; Hawran, supra, 209 Cal.App.4th at p. 293 [“truth is a complete defense” to defamation].) That is “because of the relatively high value we place on individual dignity and reputation.” (Melaleuca, Inc. v. Clark (1998) 66 Cal.App.4th 1344, 1360.) “ ‘[T]he individual's right to the protection of his own good name “reflects no more than our basic concept of the essential dignity and worth of every human being—a concept at the root of any decent system of ordered liberty.” ’ [Citation.]” (Brown, supra, 48 Cal.3d at p. 744)

When the speech involves a matter of public concern, however, a private-figure plaintiff 13 must prove the falsity of the offensive speech. (Brown, supra, 48 Cal.3d at p. 747, citing Philadelphia Newspapers, Inc. v. Hepps (1986) 475 U.S. 767, 777.) Additionally, if a plaintiff in a defamation case involving a matter of public concern seeks presumed damages, because the statements are libelous per se (Civ. Code § 45a), it must prove actual malice, i.e., that defendant knew the complained-of speech was false or acted with reckless disregard of whether it was false. (Brown, supra, at p. 747, citing Dun & Bradstreet, Inc. v. Greenmoss Builders, Inc. (1985) 472 U.S. 749, 756; see also Khawar v. Globe Internat., Inc. (1998) 19 Cal.4th 254, 273-275.) Evidence of falsity will be relevant to any determination of malice (see, e.g., Philadelphia Newspapers, Inc. v. Hepps, supra, 475 U.S. at p. 778), and is likely to be accessible to a defamation plaintiff. (See Krinsky, supra, 159 Cal.App.4th at p. 1172.)

ZL, Glassdoor, and amici curiae disagree about whether defendants' reviews of ZL involved a matter of public concern. Although the trial court's order denying ZL's motion to compel did not decide whether the reviews involved a matter of public concern, or that evidence of falsity was required, it did state a finding ZL “failed to make a sufficient showing ․ [of] wrongful conduct causing [it] harm.” The court then stated its legal conclusion that defendants' reviews were “primarily opinion” and denied ZL's motion. (Italics added.) By using the word “primarily,” amici curiae contend, the trial court implied an additional basis for its ruling, i.e., that ZL was obligated, but failed, to present evidence of falsity. ZL argues it was not so obligated because the speech did not involve a matter of public concern and therefore the burden will be on the defendant to prove falsity.

As we have discussed above, ZL is theoretically correct. Generally, the burden of proving falsity is not on the plaintiff if the alleged defamation does not involve a public figure or a matter of public concern. But the procedural posture of this case requires a different analysis, regardless of whether a matter of public concern is involved. The issue before us is not who has the burden of proof for purposes of liability, but whether there is reason to believe the lawsuit has sufficient merit to require the unmasking of the Doe defendants in the face of First Amendment and privacy rights.

As we have already discussed, an author's decision to remain anonymous is an aspect of freedom of speech that is protected by the Constitution. (See, Krinsky, supra, 159 Cal.App.4th at pp. 1163-1164.) We must also give weight to an anonymous speaker's right to protect his or her privacy interest, which is safeguarded by our state constitution. (Cal. Const., art. I, § 1.) “This express right is broader than the implied federal right to privacy. [Citation.] The California privacy right ‘protects the speech and privacy rights of individuals who wish to promulgate their information and ideas in a public forum while keeping their identities secret,’ and ‘limits what courts can compel through civil discovery.’ [Citation.]” (Digital Music News LLC v. Superior Court (2014) 226 Cal.App.4th 216, 228 (Digital Music News LLC).) So, for example, in a civil action, “ ‘the party seeking discovery [of matters protected by privacy interests] must demonstrate a compelling need for discovery, and that compelling need must be so strong as to outweigh the privacy right when these two competing interests are carefully balanced.’ [Citation.] A discovery proponent may demonstrate compelling need by establishing the discovery sought is directly relevant and essential to the fair resolution of the underlying lawsuit. [Citations.]” (Id., at p. 229.)

This weighing process is usually invoked when a party is seeking the identity of third parties (see, e.g., Johnson v. Superior Court (2000) 80 Cal.App.4th 1050, 1071 [identity of sperm donor] ), or private information about a plaintiff (see, e.g., Lantz v. Superior Court (1994) 28 Cal.App.4th 1839, 1853–1857 [plaintiff's medical records] ), and we have not found any case in which this analysis was conducted when seeking the identity of a defendant. Indeed, the need to know the identity of the defendant in order to pursue a claim will usually be deemed “essential to a fair resolution of [a] lawsuit.” (Digital Music News LLC v. Superior Court, supra, 226 Cal.App.4th at p. 230, citing Krinsky among other cases.) We are faced, however, with an increasing number of cases involving this new species of “Cybersmear” or “CyberSLAPP” (depending on whether one is a plaintiff or defendant),14 and “there is reason to believe that [at least some] defamation plaintiffs bring suit merely to unmask the identities of anonymous critics,” “ ‘the primary goal being to silence John Doe and others like him.’ ” (Cahill, supra, 884 A.2d at p. 457.) We do not here suggest that this is ZL's motivation; we are merely attempting to craft a means, consistent with due process, of filtering out those cases that are being filed primarily—or solely—as an instrumentality for identifying an anonymous speaker. Some minimal precautions should be undertaken to protect the right of a speaker to put ideas into the public marketplace without fear of harassment or retaliation.

We therefore conclude that, whether or not the defendant bears the burden of proving falsity in a particular action, the constitutional protections weigh in favor of requiring the plaintiff to make a prima facie evidentiary showing of the elements of defamation, including falsity, before disclosure of a defendant's identity can be compelled. This is congruent with the analysis in Krinsky and is tempered by the caveat that a plaintiff need only “produce evidence of those material facts that are accessible to [it].” (Krinsky, supra, 159 Cal.App.4th at p. 1172.) Further, in cases where the statements are libelous per se, as defined in Civil Code § 45a,15 actual damage or injury need not be proven at all (Hawran, supra, 209 Cal.App.4th at p. 290) and therefore would not be required in the preliminary evidentiary showing, unless the alleged libel involves a matter of public concern, in which case the plaintiff would be required to produce evidence of malice (Brown, supra, 48 Cal.3d at p. 747).16

The burden placed upon a plaintiff in these circumstances is neither heavy nor unfamiliar. The anti-SLAPP statute provides similar protections in lawsuits arising out of a defendant's exercise of the right to speak or petition. (Code Civ. Proc. § 425.16.) In such cases, a plaintiff can be required, at the outset, to “ ‘ “demonstrate that the complaint is both legally sufficient and supported by a sufficient prima facie showing of facts to sustain a favorable judgment if the evidence submitted by the plaintiff is credited.” ’ [Citation.].” (Jarrow Formulas, Inc. v. LaMarche (2003) 31 Cal.4th 728, 741.) We think the same prophylactic conditions are an appropriate and measured counterbalance to a defendant speaker's constitutional rights to privacy and anonymous speech.


The judgment is reversed. The trial court is ordered to vacate its ruling denying the motion to compel, and the case is remanded for further proceedings consistent with this decision.


1.   There is no indication in the record, or in ZL's appellate briefing, that ZL relied on any other evidence apart from the complaint allegations in seeking to establish a prima facie case of libel and ZL's counsel confirmed at oral argument ZL had not submitted other evidence below.

2.   Although ZL and defendants, if served, might disagree about whether the individual statements underlying this action were true or false, the dispute on appeal focuses exclusively on legal issues (e.g., whether the statements addressed a matter of public concern or qualified as protected opinion). The relevant facts—the content of the allegedly defamatory statements—are not in dispute.

3.   ZL does not contend on appeal that the trial court abused its discretion in dismissing the action if it was correct in denying the motion to compel. Accordingly, our review focuses on the propriety of that discovery ruling.

4.   In California, a demurrer serves the purpose of a motion to dismiss. (The Swahn Group, Inc. v. Segal (2010) 183 Cal.App.4th 831, 844.) A demurrer tests the legal sufficiency of the pleading. (Ferrick v. Santa Clara University (2014) 231 Cal.App.4th 1337, 1341.)

5.   In other cases that ZL cites, the anonymous defendants were aware of the efforts to compel disclosure of their identities and appeared through counsel to oppose those efforts, eliminating the need for those courts to consider at least one factor of the Dendrite test. (See Hadley v. Doe (Ill. 2015) 34 N.E.3d 549, 552; Cooley Law School v. Doe 1 (Mich.App. 2013) 833 N.W.2d 331, 336; Matter of Cohen v. Google, Inc. (2009) 887 N.Y.S.2d 424, 425.)

6.   The record contains no indication whether Glassdoor notified defendants in this case.

7.   In this case, it is clear ZL needed information identifying defendants to serve them with its complaint. (See, e.g., Solers, supra, 977 A.2d at p. 955 [although a final balancing test is not required, a court should ensure the requested identifying information “is important to the litigation,” for example, because “the anonymous speaker is the defendant and the litigation cannot proceed” until the defendant is served with process].) We reject Glassdoor's suggestion that ZL should be obligated to exhaust every conceivable option for identifying defendants before it can obtain the information through a subpoena. (See, e.g., id. at pp. 955-956 [“When the other elements of the test have been satisfied, we see little point in requiring the plaintiff to travel more circuitous trails in search of Doe's identity”].)

8.   If, as here, the statement is alleged to be libelous per se because the defamatory meaning is plain on its face (see Civ. Code, § 45a), then injury generally will be presumed and evidence of injury will not be required to make a prima facie case. (Hawran v. Hixson (2012) 209 Cal.App.4th 256, 290 (Hawran).)

9.   The single exception—a review that deemed ZL's CEO “crazy”—is not cited as providing a basis for this action and was so brief that its impact on the overall tenor of ZL's Glassdoor page was insignificant.

10.   Amici curiae agree the alleged statements in the sixth review about nearly 50% annual turnover and employees being subjected to “public disparagement” would be sufficient to meet a motion to dismiss, or demurrer, standard.

11.   According to its complaint, ZL had been in existence for about 13 years by the time defendant posted this review.

12.   Amici curiae agree statements contained in the second review (about inexperienced employees and a low pay rate), and in the fifth and seventh reviews (about a high employee turnover rate) were factual assertions, but they contend ZL did not adequately plead their falsity, precluding reliance on those statements to show a legally valid defamation cause of action. We disagree. Although the paragraphs of the complaint that presented those allegations did not specifically allege their falsity, the complaint elsewhere generally alleged the falsity of all review statements enumerated in the complaint, and that sufficed.

13.   There has been no suggestion ZL should be considered a public figure. (See, e.g., Reader's Digest Assn. v. Superior Court (1984) 37 Cal.3d 244, 254 [to be a “ ‘public figure’ ” a “plaintiff must have undertaken some voluntary act through which he seeks to influence the resolution of ․ public issues”]; Vogel v. Felice (2005) 127 Cal.App.4th 1006, 1021 [“A public figure or public official who seeks to recover damages for a defamatory statement bears the burden of proving that the challenged statement was false”].)

14.   See, e.g., Moore, The Challenge of Internet Anonymity: Protecting John Doe on the Internet (2009) 26 J. Marshall J. Computer & Info. L. 469, 471 [Despite downsides, “there has been a plethora of defamation, copyright infringement, and other litigation against anonymous Internet users”].

15.   Civil Code section 45a provides, in pertinent part, “[a] libel which is defamatory of the plaintiff without the necessity of explanatory matter, such as an inducement, innuendo or other extrinsic fact, is said to be a libel on its face. Defamatory language not libelous on its face is not actionable unless the plaintiff alleges and proves that he has suffered special damage as a proximate result thereof.”

16.   Our review of the record indicates the trial court did not adjudicate whether the allegedly defamatory statements can be considered libelous per se nor whether they involve matters of public concern, and we express no views as to those questions.

Rivera, Acting P.J.

We concur: Reardon, J. Streeter, J.

Copied to clipboard