Opinion for the Court filed by Circuit Judge WILLIAMS.
Zotos International, Inc. manufactures a number of hair care products, using an ingredient whose identity it contends is a trade secret — hereafter referred to as the “secret ingredient.” Regulations promulgated by the Food and Drug Administration under § 5(c)(3)(B) of the Fair Packaging and Labeling Act (the “Act”), 15 U.S.C. § 1454(c)(3)(B) (1982), require that cosmetics bear labels listing their ingredients in descending order of predominance, but permit a manufacturer to withhold the name of any ingredient whose identity is a trade secret.1 21 C.F.R. §§ 20.61, 701.3(a), 720.8 (1987).
In 1976 Zotos sought, and was denied, trade secret status for the secret ingredient. Zotos appealed the denial in federal district court, attacking both the merits of the determination and the constitutionality of the FDA’s procedures. The district court invalidated the procedures on due process grounds and remanded the matter for further proceedings. Zotos International, Inc. v. Kennedy, 460 F.Supp. 268 (D.D.C.1978) (“Zotos I”).
In response to the court’s opinion in Zotos I the FDA revised its procedures for handling trade secret applications. See 21 C.F.R. § 720.8 (1987). Applying these revamped procedures to Zotos’s resubmitted application, the FDA again determined that the ingredient was not a trade secret. Zotos again sought judicial relief. The district court referred the matter to a magistrate,2 who rejected Zotos’s claims. The district court affirmed on the basis of the magistrate’s report. On appeal to this court, Zotos presses three attacks: (1) the denial is arbitrary and capricious; (2) the FDA’s revamped procedures are facial violations of the due process clause of the Fifth Amendment; and (3) the FDA’s procedures are unconstitutional as applied in this case. We agree with Zotos that the denial is arbitrary and capricious and remand to the agency for further proceedings. We do not reach the due process claims.
I. Arbitrary and Capricious Character of the FDA Decision
Appellant’s substantive attack on the FDA decision proceeds under § 10(e) of the Administrative Procedure Act, 5 U.S.C. § 706 (1982), authorizing the court to set aside an agency decision that is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” Id. § 706(2)(A). (The substantial evidence test is inapplicable as no statute prescribes a hearing on the record, see id. § 706(2)(E)). The standard is a narrow one, but the agency must have examined the relevant data and “articulate[d] a satisfactory explanation for its action including ‘a rational connection between the facts found and the choice made.’ ” Motor Vehicle Manufacturers Ass’n v. State Farm Mut. Ins. Co., 463 U.S. 29, 43, 103 S.Ct. 2856, 2866, 77 L.Ed.2d 443 (1983) (citations omitted). We must, of course, give deference to the agency’s expertise in technical matters, FPC v. Florida Power & Light Co., 404 U.S. 453, 463, 92 S.Ct. 637, 643-44, 30 L.Ed.2d 600 (1972), but where its position “is so implausible that it could not be ascribed to a difference in view or the product of agency expertise,” we must re*352verse, State Farm, 463 U.S. at 43, 103 S.Ct. at 2866-67. Our inquiry into the facts is to be “searching and careful.” Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 416, 91 S.Ct. 814, 823-24, 28 L.Ed.2d 136 (1971).
For protection of trade secrets under the Fair Packaging and Labelling Act, the FDA has adopted the definition offered by the original Restatement of Torts, § 757 comment b (1939).3 Thus
any formula, pattern, device, or compilation of information which is used in one’s business and which gives him an opportunity to obtain an advantage over competitors who do not know or use it
qualifies as a trade secret. 21 C.F.R. § 20.61(a) (1987). And whether an ingredient meets this standard turns on an analysis of six factors that the Restatement identified as relevant:
(1) The extent to which the identity of the ingredient is known outside petitioner’s business;
(2) The extent to which the identity of the ingredient is known by employees and others involved in petitioner's business;
(3) The extent of measures taken by the petitioner to guard the secrecy of the information;
(4) The value of the information about the identity of the claimed trade secret ingredient to the petitioner and to its competitors;
(5) The amount of effort or money expended by petitioner in developing the ingredient; and
(6) The ease or difficulty with which the identity of the ingredient could be properly acquired or duplicated by others.
21 C.F.R. § 720.8(b) (1987).
Zotos discussed each of the six factors in making its case for trade secret status, claiming that they uniformly pointed to such a finding. Joint Appendix (“J.A.”) at 4-10, 13-17, 28-52. The FDA accepted Zotos’s representations “as true” with respect to the first five factors, but rejected its position on the sixth.4 J.A. at 54. It found that a knowledgeable chemist could “easily” identify the ingredient’s chemical composition by “reverse engineering,” i.e., working back from one of Zotos’s finished products to its constituent elements. See J.A. at 54.
In finding reverse engineering readily feasible, the FDA acknowledged that an inquiring chemist who had to start the process from scratch would face an overwhelming number of variables. It found, however, that the chemist could, and would, use certain publicly available information as a shortcut. The FDA’s explanation cited the particular documents it believed a literature search would reveal, and explained step-by-step how a knowledgeable chemist would employ them. J.A. at 53-55.
Looked at in isolation, this finding may well meet the standards of reasoned decisionmaking. How a chemist would go about analyzing a product is a highly technical matter in which the agency is expert. Zotos acknowledges the literature references relied upon by the FDA. It contests the *353ease with which the probing chemist would draw the inferences suggested by the FDA, but, given our hopeless ignorance of the subject matter, it would be most difficult for us to find the FDA’s rejection of these arguments arbitrary if the parties’ dispute were only a scientific one. See Baltimore Gas & Elec. Co. v. Natural Resources Defense Counsel, 462 U.S. 87, 103, 103 S.Ct. 2246, 2255, 76 L.Ed.2d 437 (1983).
The ease with which a competitor can identify Zotos’s secret ingredient, however, is not a purely scientific matter. In applying the six-factor Restatement test, the FDA has accepted Zotos’s contentions that exclusive possession of the ingredient is extremely valuable5 and that the secret ingredient’s identity is known only to Zotos. See J.A. at 54. The totality of the FDA’s findings here create a condition of cognitive dissonance: if the identity of the ingredient is of great value, and is readily ascertainable, why has no one bothered to identify it in the 20-plus years that Zotos has profited from its exclusive use? 6 FDA has offered no solution to this mystery, nor even, so far as appears, acknowledged its existence.
Of course in any case where an ingredient’s identity remains secret, we may infer that competitors have not found it worthwhile to engage in the research activities necessary for an identification. It does not follow either that all such items qualify as trade secrets — a rule which would make the first Restatement factor dispositive — or that the agency bears a special burden in establishing ease of discovery. But the FDA in adopting its trade secret regulations has implicitly found that the statute does not allow it to readily force the disclosure of highly valuable information; the fourth Restatement factor focuses on value to the manufacturer and its competitors. Implicit in this test is some sort of balance between the manufacturer’s interest in secrecy and the public’s in disclosure. Where the former is acknowledged to be very great, and the secrecy has been successfully maintained over a substantial period, any finding of easy discoverability must be reconciled with the fact of non-discovery.
“We obviously cannot affirm a decision based on ... different and inconsistent answers to the same fundamental questions.” Mid-Tex Electric Coop. v. FERC, 773 F.2d 327, 353 (D.C.Cir.1985). Therefore we reverse the denial of trade secret status and remand to the FDA to resolve the discrepancies in its findings.7
II. Procedural Claims
As the Act does not require the FDA to determine trade secret status by means of a hearing “on the record,” the proceeding falls into the APA’s largest gap: it is an informal adjudication, a category for which no procedures are specified. Compare 5 U.S.C. §§ 553, 554, 556 & 557 (specifying procedures for rulemakings (formal and informal) and for formal adjudications).
In Zotos I, the court found that the FDA’s procedures in the first proceeding violated elementary notions of due process. The core of that violation lay in the agency’s reaching a final determination without ever having afforded Zotos a clear enough glimpse into its analysis to enable Zotos to respond. As a result of Zotos I, the agency established a new, two-step process, involving a request by the petitioner for trade secret status, a tentative decision including an explanation of the grounds relied upon, an opportunity for the petitioner to respond, and a final determination. See 21 C.F.R. § 720.8 (1987).
Zotos here raises a considerable number of constitutional objections to the procedures, in both their general articulation and *354their application here. It suggests that the decision was made by an unidentified person (evidently on the assumption that the true authors of the tentative and final decisions were persons other than those named), and that this denied it an opportunity to question the authors’ credentials; that the “true” decisionmaker at both the tentative and final phases may be one and the same person; and that the FDA relied in its final decision on assertions never previously revealed.
In light of our finding that the FDA’s decision was arbitrary and capricious, we believe it would be inappropriate for us to address Zotos’s constitutional claims. See Ashwander v. TVA, 297 U.S. 288, 345-48, 56 S.Ct. 466, 482-83, 80 L.Ed. 688 (1936) (Brandéis, J., concurring); Meredith Corp. v. FCC, 809 F.2d 863, 872 (D.C.Cir.1987). A thorough FDA effort to resolve the substantive conflicts that we have identified seems likely to dispose of the objections to the agency’s application of its procedures. The facial objections may well also become moot, either through the communications between Zotos and the agency in the proceedings on remand or (possibly) through a different outcome.
It is so ordered.