ORDER

The opinion filed July 12, 2010 and appearing at 611 F.3d 653 , is withdrawn and replaced with the accompanying opinion.

OPINION

The proceeding before us is but a short chapter in an acrimonious and long-running business dispute between Quixtar, Inc. (“Quixtar”), successor to the well-known Amway Corporation, and Signature Management TEAM, LLC (“TEAM”). Quixtar sued TEAM, claiming that TEAM orchestrated an Internet smear campaign via anonymous postings and videos disparaging Quixtar and its business practices. As part of the discovery process, Quixtar sought testimony from Benjamin Dickie, a TEAM employee, regarding the identity of five anonymous online speakers who allegedly made defamatory comments about Quixtar. Dickie refused to identify the anonymous speakers on First Amendment grounds. The district court ordered Dickie to disclose the identity of three of the five speakers.

The Anonymous Online Speakers seek a writ of mandamus directing the district court to vacate its order regarding the identity of the three speakers. Quixtar cross-petitions for a writ of mandamus directing the district court to order Dickie to testify regarding the identity of the anonymous speakers from the remaining two sources. Because neither party has established that it is entitled to the extraordinary remedy of mandamus, we deny both petitions.

Background

Quixtar is a multilevel marketing business that distributes consumer products such as cosmetics and nutritional supplements through Independent Business Owners (“IBOs”). TEAM provides business training and support materials and has sold its products, including motivational literature and educational seminars, to Quixtar IBOs. TEAM was founded by two Quixtar IBOs, Orrin Woodward and Chris Brady. As IBOs, their contracts with Quixtar included post-termination non-competition and non-solicitation provisions. Disagreement regarding contract compliance and enforceability came to an impasse in August 2007, when both Woodward and Brady were terminated as IBOs, and they joined a class action against Quixtar.

TEAM and Quixtar became embroiled in several lawsuits across the country. In this suit, Quixtar asserts claims against TEAM for tortious interference with existing contracts and with advantageous business relations, among other claims. The tortious interference claims are premised on Quixtar’s contention that TEAM used the Internet to carry out a “smear campaign” with the objective and effect of inducing Quixtar IBOs to terminate their contracts at Quixtar and join a competing multilevel marketing company affiliated with TEAM.

During discovery in this suit, Quixtar took the deposition of Dickie, TEAM’S Online Content Manager. Dickie refused to answer questions regarding the identity of certain anonymous online speakers. In response, Quixtar brought a motion to compel Dickie to testify regarding his knowledge of the authors of statements from five different online sources: the “Save Us Dick DeVos” blog, the “Hooded Angry Man” video, the “Q’Reilly” blog, the “Integrity is TEAM” blog, and the “IBO Rebellion” blog. According to Quixtar, statements contained in these five fora support its claims of tortious interference, including comments such as: “Quixtar has regularly, but secretly, acknowledged that its products are overpriced and not sellable”; “Quixtar refused to pay bonuses to IBOs in good standing”; Quixtar “terminated IBOs without due process”; “Quixtar currently suffers from systemic dishonesty”; and “Quixtar is aware of, approves, promotes, and facilitates the systematic noncompliance with the FTC’s Amway rules.” Quixtar believes that the anonymous speakers of these statements are actually TEAM officers, employees, or agents.

After reviewing the specific statements from each source, the district court ordered Dickie to testify regarding his knowledge of the identity of the anonymous online speakers from three of the sources: “Save Us Dick DeVos,” the “Hooded Angry Man” video, and the “Q’Reilly” blog. The Anonymous Online Speakers from those sources filed this petition for a writ of mandamus in an effort to block Dickie’s testimony. Quixtar opposes the petition and cross-petitions for a writ of mandamus directing the district court to order Dickie to reveal the speakers from the remaining two sources — the “Integrity is TEAM” blog and the “IBO Rebellion” blog.

Analysis

I. Anonymous Speech and the First Amendment

First Amendment protection for anonymous speech was first articulated a half-century ago in the context of political speech, Talley v. California, 362 U.S. 60, 64-65 , 80 S.Ct. 536 , 4 L.Ed.2d 559 (1960), but as the Supreme Court later observed, the Talley decision harkened back to “a respected tradition of anonymity in the advocacy of political causes.” McIntyre v. Ohio Elections Comm’n, 514 U.S. 334, 343 , 115 S.Ct. 1511 , 131 L.Ed.2d 426 (1995). Undoubtedly the most famous pieces of anonymous Anerican political advocacy are The Federalist Papers, penned by James Madison, Alexander Hamilton, and John Jay, but published under the pseudonym “Publius.” Id. at 344 n. 6, 115 S.Ct. 1511 . Their opponents, the Anti-Federalists, also published anonymously, cloaking their real identities with pseudonyms such as “Brutus,” “Centinel,” and “The Federal Farmer.” Id. It is now settled that “an author’s decision to remain anonymous, like other decisions concerning omissions or additions to the content of a publication, is an aspect of the freedom of speech protected by the First Amendment.” Id. at 342 , 115 S.Ct. 1511 .

Although the Internet is the latest platform for anonymous speech, online speech stands on the same footing as other speech — there is “no basis for qualifying the level of First Amendment scrutiny that should be applied” to online speech. Reno v. Am. Civil Liberties Union, 521 U.S. 844, 870 , 117 S.Ct. 2329 , 138 L.Ed.2d 874 (1997). As with other forms of expression, the ability to speak anonymously on the Internet promotes the robust exchange of ideas and allows individuals to express themselves freely without “fear of economic or official retaliation ... [or] concern about social ostracism.” McIntyre, 514 U.S. at 341-42 , 115 S.Ct. 1511 .

The right to speak, whether anonymously or otherwise, is not unlimited, however, and the degree of scrutiny varies depending on the circumstances and the type of speech at issue. Given the importance of political speech in the history of this country, it is not surprising that courts afford political speech the highest level of protection. Meyer v. Grant, 486 U.S. 414, 422, 425 , 108 S.Ct. 1886 , 100 L.Ed.2d 425 (1988) (describing the First Amendment protection of “core political speech” to be “at its zenith”). Commercial speech, on the other hand, enjoys “a limited measure of protection, commensurate with its subordinate position in the scale of First Amendment values,” Bd. of Trustees of SUNY v. Fox, 492 U.S. 469, 477 , 109 S.Ct. 3028 , 106 L.Ed.2d 388 (1989), as long as “the communication is neither misleading nor related to unlawful activity.” Central Hudson Gas & Elec. Corp. v. Public Serv. Comm’n of N.Y., 447 U.S. 557, 564 , 100 S.Ct. 2343 , 65 L.Ed.2d 341 (1980). And some speech, such as fighting words and obscenity, is not protected by the First Amendment at all. See, e.g., Chaplinsky v. New Hampshire, 315 U.S. 568, 571-72 , 62 S.Ct. 766 , 86 L.Ed. 1031 (1942).

II. Petition by Anonymous Online Speakers

In this case, our decision is guided by the interplay of these bedrock First Amendment principles with the standards governing our review of petitions for writs of mandamus. We have repeatedly emphasized that “[t]he writ of mandamus is an ‘extraordinary’ remedy limited to ‘extraordinary’ causes.” Burlington N. & Santa Fe Ry. Co. v. U.S. Dist. Court, 408 F.3d 1142, 1146 (9th Cir.2005) (quoting Cheney v. U.S. Dist. Court, 542 U.S. 367, 380 , 124 S.Ct. 2576 , 159 L.Ed.2d 459 (2004)). This limit on our mandamus power is particularly salient in the discovery context because “the courts of appeals cannot afford to become involved with the daily details of discovery,” although “we have exercised mandamus jurisdiction to review discovery orders raising particularly important questions of first impression, especially when called upon to define the scope of an important privilege.” Perry v. Schwarzenegger, 591 F.3d 1147, 1157 (9th Cir.2010) (internal quotation marks and citation omitted).

In evaluating mandamus petitions, we are guided by the practically enshrined Bauman factors:

Id. at 1156 (citing Bauman v. U.S. Dist. Court, 557 F.2d 650, 654-55 (9th Cir.1977)). We do not require every factor to be satisfied, and “the absence of the third factor, clear error, is dispositive.” Burlington, 408 F.3d at 1146 . Ultimately, mandamus is discretionary and “even where the Bauman factors are satisfied, the court may deny the petition.” San Jose Mercury News, Inc. v. U.S. Dist. Court, 187 F.3d 1096 , 1099 (9th Cir.1999).

A. Standards Guiding Courts in Balancing Discovery and the Right to Anonymous Speech

This case is not the first time we have considered the relationship between the First Amendment and compelled discovery in the context of a petition for mandamus. See, e.g., Perry, 591 F.3d at 1165 (granting a petition for mandamus regarding a discovery order compelling disclosure of political campaign information). Perry involved the efforts of a party in the same-sex marriage suit in California to obtain internal campaign communications relating to the campaign strategy and advertising of the proponents of a ballot proposition. Focusing on First Amendment associational rights, we held that the district court erred in determining that “the First Amendment privilege, as a categorical matter, does not apply to the disclosure of internal campaign communications.” Id. at 1161 . We concluded that permitting discovery “would likely have a chilling effect on political association,” and that plaintiffs had “not shown a sufficient need for the information.” Id. at 1165 .

Although we emphasized that our holding was “limited to private internal campaign communications concerning the formulation of campaign strategies and messages,” id. at 1165 n. 12, the structure of the analysis is instructive. We first considered whether the proponents — the opponents of disclosure — made a prima facie case of arguable First Amendment infringement and then shifted the burden to plaintiffs to “demonstrate a sufficient need for the discovery to counterbalance that infringement.” Id. at 1164 .

The Perry decision l'ested on the importance of political association and political expression, and it did not involve anonymous speakers. Indeed, we have not previously considered the First Amendment claims of an anonymous, non-party speaker on the Internet in the context of commercial contractual relationships like those at issue here. Nor have we considered such a challenge in the discovery context.

Two circuit courts have, however, addressed analogous situations in published opinions. The issue has also been raised in a number of state and federal trial courts, and more cases are percolating through the system. In 1998, the Sixth Circuit considered a government agency’s motion to compel a newspaper to answer a subpoena identifying an anonymous advertiser. NLRB v. Midland Daily News, 151 F.3d 472 (6th Cir.1998). Just last year, the Fourth Circuit considered whether to uphold an order allowing a deposition of an anonymous speaker in a securities fraud class action. Lefkoe v. Jos. A. Bank Clothiers, Inc., 577 F.3d 240 (4th Cir.2009).

In both of these cases, the courts held that the anonymous speech at issue was commercial speech, but declined to establish or follow any particular standard, other than the general and long-standing precepts governing commercial speech. The Sixth Circuit, in Midland Daily News , noted that as long as commercial speech is about lawful activity and is not misleading, it is protected. 151 F.3d at 475 (citing Central Hudson Gas & Elec. Corp., 447 U.S. at 566 , 100 S.Ct. 2343 ). The court affirmed the district court’s denial of the National Labor Relations Board’s (“NLRB”) motion to compel the identification of the anonymous advertiser, because it was not the “least extensive means” the NLRB could use. Id. In Lefkoe , the Fourth Circuit reiterated that commercial speech enjoys only limited First Amendment protection and held that “the Doe Client’s claimed First Amendment right to anonymity [wa]s subject to a substantial governmental interest in disclosure so long as disclosure advance[d] that interest and [went] no further than reasonably necessary.” Id. at 248-49 . The court highlighted the balance between discovery under Federal Rule of Civil Procedure Rule 26 and protection of anonymous speech: “the substantial governmental interest in providing Jos. A. Bank a fair opportunity to defend itself in court is served by requiring the Doe Client to reveal its identity and provide the relevant information. Rule 26 explicitly expresses this interest.” Id.

This issue has arisen not infrequently in trial courts; the paucity of appellate precedent is not surprising because discovery disputes are not generally appealable on an interlocutory basis and mandamus review is very limited. The many federal district and state courts that have dealt with this issue have employed a variety of standards to benchmark whether an anonymous speaker’s identity should be revealed.

To begin, a few courts have declined to adopt a new or different standard to accommodate anonymous speech. See, e.g., Klehr Harrison Harvey Branzburg & Ellers v. JPA Dev., No. 0425, 2006 WL 37020 , at *8 (C.P.PhilaJan. 4, 2006) (noting that “the grafting of new tests onto existing rules threatens to compromise the values protected by other constitutional provisions, including due process, equal protection, and the right to a trial by jury”).

A number of courts have required plaintiffs to make at least a prima facie showing of the claim for which the plaintiff seeks the disclosure of the anonymous speaker’s identity. See, e.g., Doe I v. Individuals, 561 F.Supp.2d 249 (D.Conn.2008); Highfields Capital Mgmt., LP v. Doe, 385 F.Supp.2d 969 (N.D.Cal.2005); Sony Music Entm’t, Inc. v. Does 1-40, 326 F.Supp.2d 556 (S.D.N.Y.2004). The lowest bar that courts have used is the motion to dismiss or good faith standard. See, e.g., Columbia Ins. Co. v. seescandy.com, 185 F.R.D. 573 (N.D.Cal.1999); In re Subpoena Duces Tecum to America Online, Inc., No. 40570, 2000 WL 1210372 (Va.Cir.Ct. Jan. 31, 2000) (reversed on other grounds, America Online, Inc. v. Anonymous Pub licly Traded Co., 261 Va. 350 , 542 S.E.2d 377 (2001)).

Other courts have relied on a standard that falls somewhere between the motion to dismiss and the prima facie standards. In Doe v. 2TheMart.com, 140 F.Supp.2d 1088 (W.D.Wash.2001), the court drew from seescandy.com and America Online, but recognized that a higher standard should apply when a subpoena seeks the identity of an anonymous Internet user who is not a party to the underlying litigation. See id. at 1095 (noting that identification is only appropriate where the compelling need for discovery outweighs the First Amendment right of the speakers because litigation may continue without disclosure of the speakers’ identities); accord Sedersten v. Taylor, No. 09-3031-CV-S-GAF, 2009 WL 4802567 (W.D.Mo. Dec. 9, 2009); Enterline v. Pocono Med. Ctr., 3:08-CV-1934, 2008 WL 5192386 (M.D.Pa. Dec. 11, 2008).

The district court in this case applied the most exacting standard, established by the Delaware Supreme Court in Doe v. Cahill, 884 A.2d 451 (Del.2005). The Ca-hill standard requires plaintiffs to be able to survive a hypothetical motion for summary judgment and give, or attempt to give, notice to the speaker before discovering the anonymous speaker’s identity. Id. at 461 . The court in Cahill therefore required that the city councilman plaintiff “ ‘submit sufficient evidence to establish a prima facie case for each essential element’ ” of his defamation claim. Id. at 463 (quoting Colgain v. Oy Partek Ab (In re Asbestos Litig.), 799 A.2d 1151, 1152 (Del.2002)). The court pointed to its “concerní] that setting the standard too low will chill potential posters from exercising their First Amendment right to speak anonymously,” id. at 457, and reasoned that “the summary judgment standard more appropriately balances a defamation plaintiffs right to protect his reputation and a defendant’s right to speak anonymously.” Id. at 462.

Interestingly, in each of these cases, the initial burden rests on the party seeking discovery and requires varying degrees of proof of the underlying claim. In Perry , however, we evaluated the First Amendment political associational rights separately from the underlying claims and adopted a “heightened relevance standard” requiring plaintiffs to “ ‘demonstrate[ ] an interest in obtaining the disclosures ... which is sufficient to justify the deterrent effect ... on the free exercise ... of [the] constitutionally protected right of association.’ ” 591 F.3d at 1164 (quoting NAACP v. Alabama, 357 U.S. 449, 463 , 78 S.Ct. 1163 , 2 L.Ed.2d 1488 (1958) (omissions and alterations in Perry)).

With this broad array of standards in mind, we consider the Anonymous Online Speakers’ petition for mandamus.

B. No Clear Error

We begin with the premise that a district court “has wide latitude in controlling discovery” and that decisions governing discovery are highly fact-intensive. White v. City of San Diego, 605 F.2d 455, 461 (9th Cir.1979).

The district court here appropriately considered the important value of anonymous speech balanced against a party’s need for relevant discovery in a civil action. It also recognized the “great potential for irresponsible, malicious, and harmful communication” and that particularly in the age of the Internet, the “speed and power of internet technology makes it difficult for the truth to ‘catch up’ to the lie.”

Against this backdrop, the district court applied Cahill , which elevates the bar to disclosure to the highest level. Because Cahill involved political speech, that court’s imposition of a heightened standard is understandable. In the context of the speech at issue here balanced against a discretionary discovery order under Rule 26, however, Cahill’s bar extends too far. As in Perry and as recently illustrated by the Supreme Court in Doe v. Reed, we suggest that the nature of the speech should be a driving force in choosing a standard by which to balance the rights of anonymous speakers in discovery disputes. See Perry, 591 F.3d at 1160-61 ; Doe v. Reed, — U.S. —, 130 S.Ct. 2811, 2817-18 , 177 L.Ed.2d 493 (2010). For example, in discovery disputes involving the identity of anonymous speakers, the notion that commercial speech should be afforded less protection than political, religious, or literary speech is hardly a novel principle. See Lefkoe, 577 F.3d at 248 (inasmuch as the speech in question is of a commercial nature it “enjoys less First Amendment protection”). The specific circumstances surrounding the speech serve to give context to the balancing exercise.

By contrast with Cahill , this case does not involve expressly political speech but rather speech related to the non-competition and non-solicitation provisions of Quixtar’s commercial contracts with its IB Os. We need not, however, decide if the speech at issue here constitutes commercial speech under the Supreme Court’s definition in Central Hudson. See 447 U.S. at 561-62 , 100 S.Ct. 2343 . Even if the speech was commercial, the district court’s choice of the Cahill test did not constitute clear error.

The clear error standard is highly deferential and is only met when “the reviewing court is left with a ‘definite and firm conviction that a mistake has been committed.’ ” Cohen v. U.S. Dist. Court, 586 F.3d 703 , 708 (9th Cir.2009) (citations omitted). The district court' weighed appropriate considerations and, given the decision to disclose the speakers’ identities even under the strictest test outlines in Cahill , there was no clear error. If there was error at all, it was an error with no consequence. Cf. Sinclair v. TubeSockTedD, 596 F.Supp.2d 128 (D.D.C.2009) (declining to adopt a standard because plaintiffs claim would fail under either the Cahill or Dendrite standard).

We decline to consider the other four Bauman factors, because we conclude that the third factor, whether the district court’s order was clearly erroneous, is dis-positive. Burlington, 408 F.3d at 1146 . We deny the anonymous speakers’ petition for writ of mandamus. We leave to the district court the details of fashioning the appropriate scope and procedures for disclosure of the identity of the anonymous speakers. On this point, we note that the parties have a protective order in place that provides different levels of disclosure for different categories of documents to various recipients, such as disclosure for “Attorneys’ Eyes Only.” Second Amended Protective Order at 3, Quixtar v. Sig nature Management Team, 566 F.Supp.2d 1205 (D.Nev.2008) (No. 437). A protective order is just one of the tools available to the district court to oversee discovery of sensitive matters that implicate First Amendment rights. See Perry, 591 F.3d at 1164 (noting that a protective order can ameliorate the harms of disclosure).

III. Cross-Petition by Quixtar

In its cross-petition, Quixtar seeks reversal of the district court’s order denying the motion to compel testimony from Dickie regarding the identity of the anonymous authors of the “Integrity is TEAM” and the “IBO Rebellion” blogs. The cross-petition suffers from a fundamental error — Quixtar fails to present any foundation for its request for mandamus relief. Quixtar’s cross-petition lacks even a citation to our opinion in Bauman , which established the factors we consider to evaluate a writ of mandamus. Quixtar’s cross-petition falls into the category of a garden variety discovery dispute: it offers no extraordinary circumstance that merits exercising our mandamus power.

Conclusion

Neither party has shown that it is entitled to relief. We deny both the Anonymous Online Speakers’ petition and Quixtar’s cross-petition for writ of mandamus.

PETITION AND CROSS-PETITION DENIED.

. Not only is the mandamus standard difficult to meet as a practical matter, only in the rare case will we consider interlocutory review of discovery disputes under the collateral order doctrine. See Mohawk Industries, Inc. v. Carpenter, - U.S. -, 130 S.Ct. 599, 606 , 175 L.Ed.2d 458 (2009) (noting that courts have generally denied pre-trial review of discovery disputes). In Perry , we reserved as a close question "whether Mohawk should be extended to the First Amendment privilege." 591 F.3d at 1156 . As in Perry , we need not decide that question here because in both petitions, the parties rely on mandamus jurisdiction under the All Writs Act, 28 U.S.C. § 1651 (a). . A similar issue arose in a related case pending in the Circuit Court for the County of Kent in Michigan. On May 11, 2010, that court issued an opinion denying the Anonymous Online Speakers’ motion to quash Dickie’s deposition, during which he would presumably reveal the names of the persons who made anonymous Internet postings about Quixtar. In allowing the deposition to proceed, the court directed that only counsel may be present at the deposition, and the deposition transcript will be "for attorney eyes only.” If either party believes the presence of a non-attorney is necessary, the court noted that it would entertain such a motion. The court also noted that in the absence of a decision from this court, it would consider a motion by either party to strike portions of the transcript and/or remove the "for attorney eyes only” condition. Indep. Bus. Owners Ass’n Int’l v. Woodward, No. 07-08513-CZ (Kent County Cir. Ct. (Mich.) May 11, 2010).