On February 3, 2026, Gartner filed its year-end earnings release with the SEC. Full-year 2025 revenue: $6.5 billion, up 4 percent. Contract value, the annualized worth of its subscription agreements, closed the year at $5.2 billion. Healthy numbers for any company. They are also the numbers of a company whose flagship product has been examined closely by American courts twice and classified the same way both times: as opinion.
That word is not my characterization. It is the holding that won Gartner both cases, and Gartner deserved to win them. What interests me is the asymmetry those holdings expose. The publisher's successful legal defense rests on the ratings being subjective judgment that no reasonable reader would take as fact. The buying side of the industry cites those same ratings in RFP gates and board decks as if they were audited fact. Somebody has the classification wrong, and it is not the courts.
What does a legacy analyst subscription actually buy?
Start with the shape of the business, because the shape explains the product. Research subscriptions are Gartner's engine: contract value splits into $3.9 billion of Global Technology Sales and $1.2 billion of Global Business Sales. In the fourth quarter of 2025, the research subscription business, reported as the Insights segment, produced $1.28 billion of revenue at a 77 percent contribution margin. Most software companies would take that margin profile without asking a single question.
Individual pricing is never published, so use marketplace data and label it as such. Vendr, a SaaS purchasing platform that publishes anonymized figures from deals it has negotiated, puts the median Gartner contract at about $76,000 a year, with deals in its dataset running from roughly $27,000 to over $308,000. The same page prices analyst advisory time at $3,000 to $7,000 per hour and conference passes at $5,000 to $8,000 per attendee. These are estimates from observed deals, not Gartner's rate card, and any given contract will vary. The order of magnitude is the point.
Now the cadence. Gartner's own Magic Quadrant FAQ states that Magic Quadrants are reviewed annually, with each report's refresh date noted on its planned research list. So the flagship comparative artifact inside a five- to six-figure subscription updates once a year, in markets where the products being compared ship monthly.
None of this is scandal. It is a precise description of a product category: periodic, subscription-gated, comparative opinion, produced at scale and at enviable margins.
What have courts actually said about the Magic Quadrant?
Twice, vendors unhappy with their placement have tested the question in court. Both outcomes need stating in full, because allegations travel further than verdicts.
ZL Technologies, an email-archiving vendor, sued Gartner in 2009 over its Niche Players placement, alleging defamation and related claims. The federal court for the Northern District of California dismissed the case, ZL Technologies v. Gartner, 709 F. Supp. 2d 789 (2010), holding that the placement was non-actionable opinion. The overall tenor of a Magic Quadrant report, the court reasoned, negates any impression that a vendor's dot position asserts objective fact.
NetScout went further. After Gartner's 2014 report for the network performance monitoring market designated it a Challenger rather than a Leader, NetScout sued in Connecticut, alleging the placement reflected a pay-to-play scheme tied to consulting purchases. Gartner won in the trial court in 2017, and won on the evidence: NetScout's own expert could not connect ratings to consulting spend without disregarding the two vendors that bought the most consulting, both of which ranked comparable to or lower than NetScout. The Connecticut Supreme Court affirmed in January 2020, holding that every challenged statement was nonactionable opinion. Gartner was cleared of pay-to-play on a full record, and it publicizes the ruling with justified satisfaction.
Read the two holdings side by side, though, and notice what the winning defense is. A defamation claim needs a false statement of fact. The Magic Quadrant survived both suits because, as a matter of law, it makes no statement of fact for a court to test. That is a legitimate defense, and a revealing one. In my view, those two opinions are the most authoritative product documentation the analyst industry has ever received.
Why buyers treat opinion-class research as evidence
Because opinion is convenient, and I mean that structurally rather than as a jab. An opinion cannot be wrong in an auditable way. It carries no error rate to inspect and no falsifiable claims to re-check, and on an annual cycle it holds still long enough to cite. Evidence behaves worse. It arrives with dates and named sources, and it lives with the permanent risk of being contradicted by something newer. A procurement team under deadline will reach for the thing that holds still.
Here is a hypothetical, and only that. A procurement lead at a mid-market insurer writes a shortlist rule: Leaders quadrant only. Sentrix, a fictional vendor, shipped the exact capability the insurer needs eight months ago, well after the report's data cutoff, so the snapshot carries none of it and Sentrix never reaches a demo. Meridian, an equally fictional Leader, earned part of its position on a capability it has since deprecated. Both errors share a root: nobody re-checked a claim after the report's as-of date, because the artifact was treated as settled evidence rather than dated opinion. What other input to a seven-figure purchase decision would survive the sentence "legally, this is opinion"?
What separates evidence-class intelligence from opinion-class research
A working definition. Evidence-class intelligence ties every capability claim to a named, inspectable source. It grades each source for independence, because a vendor's own datasheet and a third-party confirmation are not the same class of support. It refreshes on the market's release rhythm rather than a publisher's production calendar. Above all, it is falsifiable: a newer fact can overturn a grade, and the record shows when that happened.
By that definition, OmniAxis is an evidence-class system. It re-checks vendor capability claims on a managed schedule, weekly on the top plan, rather than once a year, and each claim carries a grade for the independence of the support behind it. A vendor's marketing-claim score sits beside its evidence-backed score, so the gap between what a vendor says and what independent material actually supports is itself visible, dated data. The grades are falsifiable by design: when a newer fact overturns one, it moves with that fact, not with the next annual publication.
Keep the analyst subscription if the inquiry hours earn their keep; a sharp analyst on the phone can be worth the rate. But relabel the quadrant in your head, because two courts have already done the labeling for you. It is opinion. Cite it, price it, and gate your shortlists on it accordingly.