Opinion for the Court filed by Circuit Judge TATEL.

TATEL, Circuit Judge:

To facilitate speedy consideration of two mergers of major cable companies, the Federal Communications Commission ordered the merger applicants to submit certain proprietary documents for review and, central to this case, proposed to make them available for examination by other players in the cable industry on an expedited schedule. Concerned that those documents would reveal information about their own dealings, petitioners — several other large entertainment companies — asked the Commission to reconsider. The Commission refused. Because, for the reasons set forth in this opinion, we find the Commission's action both substantively and procedurally flawed, we grant the petition for review and vacate the order.

I.

The Communications Act of 1934 requires the Commission to review cablecompany mergers. 47 U.S.C. § 310. The heart of that mandate, section 310(d), prohibits any merger unless it serves "the public interest, convenience, and necessity." To assess a particular merger, the Commission has long required the parties to submit information about their business. In the context of cable-company mergers, that information usually includes key affiliate contracts and negotiation documents. To help it better understand those materials, the Commission has on occasion asked third parties — usually people with insight into the specific industry — to review and comment on them. See Examination of Current Policy Concerning the Treatment of Confidential Information Submitted to the Commission, 13 F.C.C.R. 24816, 24831 (1998) (¶ 21) ("Confidential Information Policy") ("In recent years, the Commission has ... permitt[ed] limited disclosure for a specific public purpose.").

The issues in this case arise in the context of two proposed mergers: AT & T seeks to join forces with DirecTV, and, until recently, Comcast wanted to combine with Time Warner Cable and Charter Communications. (On April 24, Comcast and Time Warner dropped their merger bid. See Shalini Ramachandran, Comcast Kills Time Warner Cable Deal, WALL STREET JOURNAL, Apr. 24, 2015, available at http://goo.gl/vPG1hh.) Because the original merger applicants made up five of the world's seven largest video-programming distributors, the Commission requested that they submit for review certain documents that it believed would help it evaluate these important corporate marriages.

We can best explain why the Commission needs such documents with a hypothetical. Suppose DirecTV is eager for its customers to have access to ESPN — ESPN being the most-watched cable channel among the key 18-49 demographic group. See Rick Kessell, ESPN No. 1 in Cable Ratings for 2014, VARIETY. COM, available at http://goo.gl/QQG2Bp. If DirecTV wishes to offer ESPN to its subscribers, it will have to negotiate a price with Disney, which owns the channel. Likewise, AT & T will have to reach its own deal with Disney if it wants to offer its customers the same sports package DirecTV does. And when DirecTV and AT & T ask for permission to merge, this information — what kind of a deal DirecTV agrees to with Disney, and how AT & T's compares — could help the Commission understand what the market would look like if the two cable companies combined.

Petitioners — CBS, Viacom, Disney, and several other content producers — have no complaint about the Commission's decision to review that information. In fact, they seem quite eager for the Commission to take a hard look at the proposed merger, and they agree that those contracts and negotiating documents are important to the process. They worry, however, that the Commission plans to show that information to third parties and that their own proprietary documents have gotten caught up in the dragnet.

For example, suppose the Commission discloses the DirecTV and AT & T contracts described earlier. Although the decision to make this kind of proprietary business material available to outsiders is not always popular, the Commission maintains, it is the price of doing business. If two companies want to merge, they must prove that the merger is in the public interest, and to do so they often have to release some information. If the Commission gives third parties access to information about the merger applicants' dealings with ESPN and Disney, however, more than just the applicants will be affected. For instance, by disclosing AT & T's contracts with Disney, the Commission will necessarily be disclosing Disney's contracts with AT & T. It would therefore be a simple matter for, say, Fox to peruse those documents, figure out what Disney charges for ESPN, and then price its own sports channel accordingly. Not having signed up for that exposure, petitioners think it unfair and, more important for our purposes, unlawful. Specifically, they argue that such disclosure is precluded by the Trade Secrets Act, which prohibits disclosing sensitive business information unless "authorized by law," 18 U.S.C. § 1905, as well as the Commission's own regulations and internal policies, which provide that a "persuasive showing as to the reasons for inspection will be required," 47 C.F.R. § 0.457(d)(1), that disclosure must serve a "compelling public interest," that the benefits of disclosure must outweigh the costs, and that the underlying documents must be "necessary" to the review process. Confidential Information Policy at 24820-21, 24824 (¶¶ 5, 8).

The Commission has been sensitive to those concerns. Indeed, recognizing that its disclosure decisions could have significant collateral consequences, the Commission has long worked to ensure that confidential materials are as protected as possible — while also serving the public's interest in meaningful merger review — by using protective orders. According to the Commission, such orders "can provide the benefit of protecting competitively valuable information while permitting limited disclosure for a specific public purpose." Id. at 24831 (¶ 21); see also News Corp.-Liberty Media Corp., 22 F.C.C.R. 12797, 12798-804 (2007) (same).

In petitioners' view, however, the Commission has not done enough to protect their information, which has come to be known as Video Programming Confidential Information, or VPCI. Throughout the merger-review process, the Commission permitted third parties to access highly confidential information, including VPCI. Although the protective orders contained certain safeguards to mitigate the risk of unauthorized disclosure, those protections centered on the merger applicants. For instance, the orders permitted only outside counsel and outside consultants not involved in "Competitive Decision-Making" — that is, negotiating or advising on contracts between a company and one of the merger applicants — to access VPCI, and it allowed only the merger applicants to object to disclosure. See, e.g., Joint Protective Order, Joint App. 246-47, 249-50 (¶¶ 2, 7) (April 4, 2014). Petitioners filed comments contesting the Commission's decision to disclose VPCI, contending that the only effective way to address their concerns would be for the Commission to forgo disclosure of these materials entirely and instead to review them in secret.

In response to those concerns, the Commission's Media Bureau sought public comment on possible additional procedural protections. After considering those comments, on October 7, 2014 — and here the dates are relevant to the procedural issue before us — the Bureau issued an order that, though continuing to permit disclosure to any potential commenter not engaged in Competitive Decision-Making, adopted additional procedures to protect third parties from competitive harm. Among other steps, the order expanded the definition of "Competitive Decision-Making" to include outside counsel and consultants working for entities in competition with any party having an interest in VPCI, not just those in competition with the merger applicants. It also allowed such interested third parties, not just the merger applicants, to object to disclosure. See Modified Joint Protective Order, Joint App. 135-36 (¶¶ 2, 7-8, 10) (October 7, 2014) ("October Bureau Order").

That last protection is critical. The October Bureau Order — like, petitioners submit, all such orders before it — would have prevented disclosure of confidential information "[u]ntil any objection is resolved by the Commission and, if appropriate, by any court of competent jurisdiction." Id. ¶ 8. By guaranteeing agency and judicial review, the order would have prevented hasty and potentially ill-conceived disclosure decisions. Petitioners no doubt applauded that review provision, but it turned out to have an exceedingly short half-life.

Soon after the Bureau issued its October Order, individuals — mostly representatives of the merger applicants and other entities interested in the outcome of the proposed mergers — began filing requests for access to VPCI. Concerned that some of those individuals also happened to represent direct competitors, petitioners, on October 15, filed objections to all these requests. Because the Bureau believed that this broad challenge would effectively stall VPCI disclosure and therefore delay the merger-review process until the Commission and a court could adjudicate each of those objections, it announced on November 4 that it would reconsider the October Order. In its decision on reconsideration, the Bureau reaffirmed that VPCI "must be part of the record available to commenters, subject to the multiple protections in the... Protective Order[] that minimize any risk of competitive harm as a result of the production." Order on Reconsideration, Joint App. 36 (¶ 17) (November 4, 2014) ("November Bureau Order"). The Bureau also amended the protective order in one respect central to the issues before us: it truncated the process for challenging and reviewing VPCI-access requests. Under the new order, individuals seeking to view VPCI would be allowed access just five days after the Bureau — not the Commission or a court — rejects any objections. Id. at 45 (¶ 36).

According to the Bureau, with that decision it sought to balance the "opportunity for the consideration of legitimate objections" with the need to "proceed[] with the merger review in a timely manner." Id. By contrast, petitioners pointed out in a request for further review that as a result of the latest order, within five days of a Bureau decision granting access to VPCI, any objection must be filed with the Commission, and, even if the Commission issues a decision in time, the objection must be raised in a court in an emergency proceeding. Again, all within five days — a requirement that two commissioners strongly criticized. See Order, Joint App. 3-4 (Dissenting Statement of Commissioner Pai), 5 (Dissenting Statement of Commissioner O'Reilly) (November 10, 2014) ("November Commission Order").

In an order issued November 10, the Commission denied petitioners' application for review and, "for the reasons stated by the Media Bureau," affirmed the amended protective order. November Commission Order ¶ 1. To give the parties time to seek judicial review of that order, the Commission delayed access to VPCI for seven calendar days. Id. ¶ 3. Taking the Commission up on that offer, petitioners sought review in this court, arguing both that the decision to disclose VPCI at all was unlawful and that the five-day process was inconsistent with past agency practice. At the same time, they sought an emergency stay pending review. In granting that petition, this court's Special Panel noted that although the stay precluded third-party access to VPCI, "[t]he agency has access to the relevant documents at issue in this matter and can continue to evaluate the proposed merger during the stay." Order, CBS Corporation v. Federal Communications Commission, No. 14-1242 (Nov. 21, 2014).

The dispute here, then, boils down to the following: May the Commission disclose petitioners' confidential information to third parties and may it do so on a timeline so swift as to effectively preclude judicial review? We consider these issues in turn.

II.

We begin with petitioners' substantive challenge to the Commission's decision to disclose their VPCI, and we look first to the relevant text. See 5 U.S.C. § 706(2)(A), (E) (agency's application of statute and implementing regulations shall be "set aside" if "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law"). In this case, that text consists of the Trade Secrets Act, the Commission's regulations, and its Confidential Information Policy.

Partly a response to government mishandling of confidential business information, the Trade Secrets Act makes it criminal for government officials to publish such information unless disclosure is "authorized by law." 18 U.S.C. § 1905. In this case, the Commission's regulations provide the necessary authority: although "[t]rade secrets ... are not routinely available for public inspection," the Commission may, despite the Act's near-categorical protection, disclose private information upon a "persuasive showing as to the reasons" for doing so. 47 C.F.R. § 0.457(d)(1), (2).

This case presents two questions: what exactly does that persuasive showing entail, and has the Commission made its case? Before considering those questions, however, we need to address an antecedent issue: must the Commission make a persuasive showing in cases like this one at all? We ask that question because the regulations say only that "a persuasive showing as to the reasons for inspection will be required in requests for inspection." 47 C.F.R. § 0.457(d)(2). The regulations say nothing about what should happen where, as here, the Commission decides to disclose confidential documents on its own, not in response to a request. The Confidential Information Policy — a document the Commission wrote ostensibly to clarify the contours of its approach to the kind of private documents at issue here — is no more helpful. Framing the persuasive-showing mandate almost entirely in the passive voice, the Policy allows disclosure if "`a persuasive showing' is made," Confidential Information Policy at 24822 (¶ 8) (emphasis added), and requires the Commission to determine that a "`persuasive showing' has been made to warrant disclosure," id. at 24821 (¶ 6) (emphasis added). But made by whom — the requester or the Commission? And what happens if there is no requester and the Commission decides to disclose confidential information on its own initiative? The Policy provides no answers.

Notwithstanding this confusion, the Bureau concluded in its November Order that it must make the persuasive-showing finding whether or not someone requested disclosure. Specifically, it observed that section 0.457(d) "permit[s] disclosure" of confidential information "on a `persuasive showing' of the reasons in favor of its release." November Bureau Order ¶ 23. Commission counsel admitted as much at oral argument. See Oral Argument Recording at 22:02 ("I think we're in a world where the persuasive-showing standard applies."). For purposes of this opinion, we shall therefore assume that the persuasive-showing standard applies to the Commission's disclosure of petitioners' documents and, because there is no requester, the Commission itself must satisfy the requirement. See Thomas Jefferson University v. Shalala, 512 U.S. 504, 512, 114 S.Ct. 2381, 129 L.Ed.2d 405 ("We must give substantial deference to an agency's interpretation of its own regulations.").

With this issue out of the way, we turn to the first of our two questions: What must the "persuasive showing" look like? To answer that question, we return to the Confidential Information Policy. In places the Policy appears to make clear exactly what is required: The Commission permits disclosure where it "has identified a compelling public interest in disclosure," and "the rules also contemplate that the Commission will engage in a balancing of the interests favoring disclosure and nondisclosure." Confidential Information Policy at 24822 (¶ 8). But even if the Commission finds that the public interest and the balance of equities favor disclosure, it will "not automatically authorize ... release of such information." Rather, "the Commission has adhered to a policy of not authorizing the disclosure of confidential financial information on the mere chance that it might be helpful, but insists upon a showing that the information is a necessary link in a chain of evidence that will resolve an issue before the Commission." Id. at 24823 (¶ 8) (internal quotation marks omitted).

If that were all the Confidential Information Policy had to say on the matter, we could stop there: in order to justify disclosure to third parties, the Commission would have to demonstrate that "necessary link." But in Paragraph 17, which the Bureau never mentioned in its November Order, the Policy casts some doubt on that requirement: "Because [the Commission] believe[s] that a case-by-case determination is most appropriate ... [it] decline[d] to adopt a blanket rule requiring the requester to demonstrate that access is `vital' to the conduct of a proceeding [or] necessary to the `fundamental integrity' of the Commission process at issue." Id. at 24829 (¶ 17).

What, then, must the Bureau do to justify disclosing confidential business information? In its November Order interpreting the Confidential Information Policy, the Bureau acknowledged that its "persuasive showing" must include "identif[ying] a compelling public interest in disclosure" and that "[t]he rules also contemplate that the Commission will engage in a balancing of the interests favoring disclosure and nondisclosure." November Bureau Order ¶ 23. Although the Bureau lifted these requirements wholesale from the first two sentences of Paragraph 8 of the Policy, it inexplicably failed to include the next sentence, which, recall, makes up a major part of that same paragraph: The Commission will "not automatically authorize... release of such information" just because disclosure is in the public interest or because the information will be helpful to the process. Instead, the information must serve as a "necessary link in a chain of evidence." Confidential Information Policy at 24822-23 (¶ 8). The Bureau believed the first half of that paragraph binding. It has given no reason why the second half — the necessary-link requirement — should not also control. In these circumstances, and notwithstanding Paragraph 17's enigmatic and unexplained language, we conclude that Paragraph 8's necessary-link finding is an unavoidable component of the persuasive showing the regulations require.

To sum up: to make the persuasive showing necessary to disclose petitioners' confidential documents, the Commission must explain (1) why disclosure is in the public interest, (2) why it is a good idea on balance, and (3) why the information serves as a "necessary link in a chain of evidence." The Bureau's November Order easily clears the first two bars. The benefits to the public are obvious: third-party review of VPCI documents will ensure a sounder decision. If "a large number of... documents [were excluded] from review by commenters," "it would deprive the commenters of the opportunity to argue that the documents have significance in ways that are not apparent to the Commission." November Bureau Order ¶ 16. This different perspective on materials that the Commission is considering facilitates informed decision making. At the same time, petitioners have offered no evidence that the Commission will countenance disclosure of the kind described earlier — say, allowing executives at Fox to see Disney's contracts. And we have reason to doubt that they could make that case, as the governing Protective Orders limit VPCI access to "outside counsel of record" and "outside consultants" who are not "involved in Competitive Decision-Making." See October Bureau Order at 2. The new order also confirms that anyone who obtains access to VPCI may use it "solely for the preparation for and conduct of [the merger] proceeding." Id. ¶ 6. The risks involved in disclosure thus appear minimal. Accordingly, disclosure would serve the public's interest in a thorough review process, and the benefits outweigh the harms.

But the Commission falters at the last requirement: the confidential information must be necessary to the Commission's review process. In its Order, the Bureau concluded that VPCI is "highly relevant... to the pending transactions" — even "central." November Bureau Order ¶ 23. In normal parlance, however, "relevant" and "central" are not the same as "necessary." Something is "relevant" if it merely "ha[s] significant and demonstrable bearing on the matter at hand." WEBSTER'S NEW COLLEGIATE DICTIONARY 976 (Henry Bosley Woolf, ed., 1977). A piece of information is "central" if it is "of cardinal importance." Id. at 181. By contrast, something is "necessary" only if it is "absolutely needed" or "required." Id. at 767.

We think this linguistic distinction makes the best sense of all relevant texts. The Trade Secrets Act exists for an important reason — Congress has decided that confidential business information should be private unless there's good cause to disclose it — and the Commission recognizes as much: its regulations acknowledge that "[t]rade secrets ... [are] not routinely available for public inspection," 47 C.F.R. § 0.457(d), and the Confidential Information Policy makes clear that disclosure will not be "automatic[]" but will instead be proper only in limited circumstances, Confidential Information Policy at 24822-23 (¶ 8). By contrast, because corporate business documents will almost always be relevant to a merger between two industry participants, allowing the Commission to disclose confidential information based on mere relevance would mean that such information would, subject to the governing protective orders, be routinely available for inspection. We must read the statute and the Commission's precedents to avoid that construction if we are to be faithful to Congress's plan and to the Commission's own historical approach.

Consistent with that goal, this court has affirmed the relevant/necessary dichotomy in a nearly identical situation. In Qwest v. Federal Communications Commission, we decided that a general desire to permit broad public participation, or even an interest in a more effective decision-making process, must yield when sensitive information will be disclosed to competitors. See 229 F.3d 1172, 1180-84 (D.C.Cir.2000). The question in that case, much like the question here, was whether the Commission could release private business data to supplement its own audit procedures. The Commission argued, as it does here, that it would get "useful information about the accuracy and validity of the audits" if "commenters were allowed to examine how those general procedures were actually implemented." Id. at 1183. We held that although "broad[] comment [might have] greatly assist[ed] the Commission in resolving the issues" before it, the Confidential Information Policy suggests that "assistance" is not enough. Instead, disclosure is proper only if the information disclosed is absolutely necessary to the process. In that case, we thought it "unclear why" that information was necessary, observing that the audit methodology could be "evaluated in theoretical terms as applied to hypothetical situations or to a composite of raw data without identifying an individual['s] sensitive commercial information." We also noted that "[o]ther ways of avoiding the release of raw audit data to competitors might be equally effective for the Commission's purposes." Id. Crucially, the petitioners had failed even to offer an alternative, yet it was enough for the court that "on the basis of the record, [it could not] tell that other ways would not be equally effective." Id.

So too here. We have no doubt that petitioners' VPCI, as well as commenter analysis of it, would be helpful to the Commission's evaluation of the proposed mergers since those documents "contain[] information that is central to the contracting parties' ... business strategies." October Bureau Order ¶ 13. And review of those business strategies is in turn essential to the Commission's merger-review process, as "[a] critical issue" in that process is how each proposed transaction "will alter the incentives and abilities of the resultant companies as they bargain with [programmers]." November Bureau Order ¶ 11. The private documents at issue here thus "provide what is likely the best evidence available to test the validity of allegations as to how incentives and abilities ... vary with size, integration, and other characteristics that the transactions would alter." Id. Are the documents relevant? Absolutely. Important? Sure. Central? Probably. The Commission would thus be derelict if it failed to consider VPCI as it evaluates the proposed mergers.

But to justify disclosure, the information must be "necessary" to the Commission's review process. Otherwise, Congress and the Commission have decided, the risk to the affected businesses will not be worth it. And we simply have no idea whether VPCI is necessary to that process. It might be, for example, that, as in Qwest, other information — or information in another, less compromising form — could be sufficient to analyze the merger. Nowhere does either the Bureau or the Commission make the jump from useful or relevant or central to necessary.

In short, by failing to explain why VPCI is a "necessary link in a chain of evidence that will resolve an issue before the Commission," the Commission has failed to overcome its — and Congress's — presumption against disclosure of confidential information. We shall therefore vacate the Commission's Order.

With that conclusion established, we turn to one more interpretive issue that deserves mention. Although the Confidential Information Policy makes it apparent that the Commission must show that something is a "necessary link in a chain of evidence," the key passage is susceptible to two interpretations. The Commission will "not authorize disclosure of confidential financial information on the mere chance that it might be helpful, but insists upon a showing that the information is a necessary link in a chain of evidence." Confidential Information Policy at 24823 (¶ 8). But what does "it" refer to? And which "information" must be necessary: the information gleaned from third parties or the confidential information itself? Looking at the two clauses of the sentence together, we read the "information" referred to in the second clause to describe the confidential information at issue — after all, the word "information" in the first clause refers to those confidential documents — and, consequently, we take the "it" to refer to that same confidential information. So, replacing the pronoun with "that confidential information," we believe the Commission meant to say this: it will "not authorize disclosure of confidential financial information on the mere chance that [that confidential information] might be helpful, but insists upon a showing that the [confidential] information is a necessary link in a chain of evidence." In other words, we understand the Commission to be saying that it will not allow outsiders to view confidential information unless the information itself is necessary to the evaluation process. This makes sense. In order to vindicate the goals of the Trade Secrets Act, the Commission will refuse to disclose confidential documents unless it has a good reason to do so — namely, that it would benefit from third-party comment on information that is necessary to the review process.

Although this understanding has guided our analysis, we note that even if the Commission meant to say in Paragraph 8 of the Confidential Information Policy that disclosure — not the information itself — must be a "necessary link," the agency still failed to make its case. This is because in adopting the governing protective orders, the Commission made no effort to explain how disclosure of VPCI to any and every qualifying entity that might file a comment in this proceeding is necessary to the process. The potential commenter's willingness to sign the protective order does not, as the Commission would have it, answer this question by itself.

In reaching these conclusions, we emphasize that we have done our best to make sense of the confusing and often contradictory materials in light of the Commission's own stated understanding of them. Given this, we take no position on what the Commission should do next. When it reconsiders its disclosure order, the Commission is free to clarify its current policy or to amend it. It may, for instance, explain who must make the required "persuasive showing"; what must be a "necessary link in a chain of evidence" — the confidential information itself or third-party comments on it; and whether "necessity" is the standard at all.

III.

This brings us to petitioners' procedural challenge. When an agency departs from past practice, it "must provide a reasoned analysis indicating that prior policies and standards are being deliberately changed, not casually ignored." Ramaprakash v. Federal Aviation Administration, 346 F.3d 1121, 1124 (D.C.Cir.2003) (internal quotation marks omitted). It must, in short, explain why it has changed its policy. Until November 2014, when the Bureau ordered disclosure of petitioners' sensitive business information, the Commission always allowed aggrieved parties to seek review at the agency level and, if necessary, by a court — all before the Bureau could disclose the information. Indeed, as the Commission has made clear, it has long recognized that confidential information should remain confidential until the merits of a disclosure decision have been fully resolved. See Confidential Information Policy at 24832, 24856-57.

That practice makes sense given that review can be effective only if it occurs before confidential information is disclosed to third parties. "Disclosure followed by appeal after final judgment is obviously not adequate in such cases — the cat is out of the bag." In re Papandreou, 139 F.3d 247, 251 (D.C.Cir.1998); see also Ruckelshaus v. Monsanto Co., 463 U.S. 1315, 1317, 104 S.Ct. 3, 77 L.Ed.2d 1417 (1983) (once trade secrets are disclosed, they cannot "be made secret again if the judgment below ultimately is" reversed).

The November Bureau Order, however, eliminates pre-disclosure review. If the Bureau orders disclosure, it may now make the documents in question available to third parties "five business days after" it resolves any "objection ... in favor of the person seeking access," even if neither the Commission nor a court has had an opportunity to weigh in. November Bureau Order ¶ 8. So under the new protective orders, aggrieved parties like petitioners have only five days to challenge the Bureau's decision to disclose their information, and if they fail to convince the Commission or a court either to stay or to overturn that decision, they are out of luck.

The Order thus amounts to a substantive and important departure from prior Commission policy. According to petitioners, the Commission has failed entirely to acknowledge this difference, much less to explain it. For its part, the Commission insists that it has met its obligation to explain a departure from prior policy, pointing to Paragraph 36 of the Bureau's November 4 Order. Because that paragraph represents the Commission's entire explanation, we think it worth quoting almost in full. The Bureau begins by announcing that it is "amend[ing] paragraph 8 of the Modified Joint Protective Orders to remove any doubt about whether a party is able to suspend indefinitely another party's (or every other party's) effective participation in the proceeding simply by filing an objection." November Bureau Order ¶ 36. After reciting the new five-day rule and before doing some housekeeping, the Bureau stated that "this approach provides an appropriate balance between providing ample opportunity for the consideration of legitimate objections and proceeding with the merger review in a timely manner." Id.

In our view, the Commission has failed to make its case. For starters, although the Commission concedes that the Bureau has changed the governing protective orders, the Bureau acknowledged nowhere in its Order that the new rule departs from longstanding practice. The Commission insists that by adding the five-day rule to the protective order, the Bureau did acknowledge that it was breaking from precedent. That is, the Bureau acknowledged the departure — by departing. This, of course, is completely insufficient. An agency must "provide a reasoned analysis indicating that prior policies are being deliberately changed." Ramaprakash, 346 F.3d at 1124 (emphasis added) (internal quotation marks omitted).

In the alternative, the Commission argues that by acknowledging it was altering the protective orders, the Bureau recognized the larger policy shift. But admitting to a technical change in the governing documents is a far cry from acknowledging a fundamental departure from longstanding policy. Instead, it seems like the old policy is being "casually ignored." Id.

On this issue, then, the Commission begins in a deep hole. Worse still, it offers an exceedingly thin rationale for the new rule. By suggesting that the five-day rule prevents a party from getting in the way of another party's effective participation in the process, we take the Commission to be saying that the Bureau was concerned about a large number of protests gumming up the works, and that a five-day time limit to raise such concerns would avoid that consequence. We see two serious problems with this rationale.

First, the Commission never explains how the old rule, which precluded disclosure until a court had a chance to weigh in if necessary, actually slowed things down. As the Special Panel observed in its order granting a stay in this case, petitioners' objections do not prevent the Commission from accessing VPCI and conducting its review of the proposed mergers. In fact, the stay prevented the Commission neither from restarting its internal clock for completing its review of the proposed transactions — at least for a time, see Public Notice (DA 14-1739) (Dec. 3, 2014), available at http://goo.gl/hvNlDX — nor from issuing additional data requests to the merger parties, nor from setting new schedules requiring all submissions in both merger proceedings to be filed by mid-January 2015, id.

The Commission's claim that the Bureau adopted the shortened review procedure because petitioners have "abus[ed] the objection process" by challenging every one of the 266 requests for access to VPCI rests on a similarly flawed premise. The vast majority of those challenges — some 230 — were general challenges to the disclosure of VPCI. Because neither the Commission nor the court had yet ruled on the propriety of disclosing VPCI when petitioners filed their objections, how could anyone think that petitioners abused the process by bringing those concerns to the agency's attention? In other words, given the Commission's failure to act on petitioners' intra-agency appeal and emergency stay request, the objection process represented the only administrative avenue open to petitioners to protect their right to meaningful pre-disclosure review. And finally, because, as the Commission itself pointed out in another context, few complex challenges will remain once the global challenge is resolved, the review process is unlikely to get bogged down even without the expedited, five-day rule.

Moreover, even were speed a potential concern, the Commission has failed to explain why expedited review is so important here given that it has followed the old rule through dozens of merger reviews over the last fifteen years. See In the Matter of Examination of Current Policy Concerning the Treatment of Confidential Information Submitted to the Commission, 14 F.C.C.R. 20128 (1999) (¶ 4) (providing for pre-disclosure judicial review even though "disclosure [could] be delayed pending the appeals process"). So why not continue to allow pre-disclosure review? Companies like petitioners often oppose disclosure, and given the competitive stakes it seems safe to assume that companies opposing disclosure will challenge it in court. Are there usually fewer challengers? Fewer challenges? Nothing in either the Bureau's Order or the Commission's brief sheds any light.

We conclude with a cautionary observation. Although petitioners emphasized their departure-from-past-practice argument, they suggest a substantive concern as well: "the Operative Protective Orders fail to give Petitioners a meaningful opportunity to ensure" that they "will not be harmed by disclosure." Petitioners' Br. 17. We share petitioners' apprehension about a process that puts tremendous pressure on the Commission, the parties, and this court to get their ducks in a row in a short time. We say this not to prejudge the question, but simply to emphasize to the Commission that should it choose to retain the five-day rule, it must not only come forward with a "reasoned analysis" for this dramatic break from the past, but also explain why speed is so important as to justify limiting one of the fundamental principles of administrative law — judicial review. See Mach Mining, LLC v. Equal Employment Opportunity Commission, 575 U.S. ___, 135 S.Ct. 1645, ___, ___ L.Ed.2d ___ (2015), slip op. at 8 (reiterating and explaining the "strong presumption favoring judicial review of administrative action").

IV.

For the foregoing reasons, we grant the petition for review and vacate the Commission's order.

So Ordered.