The Engine Stalled

The Engine Stalled

The two dials that govern the subscription engine — contract value, which tells whether the book is refilling, and wallet retention, which tells whether each client is growing or shrinking — both turned the wrong way in 2025. Total Insights contract value grew 1% on a foreign-exchange-neutral basis, to $5.16 billion, and net wallet retention crossed below 100% for the first time since the pandemic year [1]. The machine did not break. It stopped refilling. What follows separates how much of that is Gartner's problem from how much is the category's.

Sources: FY2025 10-K, segment KPI table [2]; Q1 FY2026 earnings call [3].

Clients Stayed, Spend Shrank

The two dials from The Cash Machine — client retention on logos kept, wallet retention on dollars kept from those same clients [4] — pulled apart in 2025 in the diagnostic direction. Client retention held — 85% in the Global Technology Sales channel and 86% in Global Business Sales, both within a point of the prior year — while wallet retention dropped six to seven points, to 96% for GTS from 102%, and to 99% for GBS from 106% [5]. Enterprises are not walking out the door. They are renewing for less.

That distinction narrows the diagnosis considerably. A wave of departing logos would point to a product losing to a rival; flat logos with shrinking spend points to the expand half of land-and-expand going into reverse — fewer added seats, less new-topic enrichment, thinner upsell — while the installed base stays loyal. It shows up first in the leading gauge. Global Technology Sales contract value, the larger channel and the one selling to both technology users and technology vendors, finished the year at $3.91 billion, effectively flat against $3.91 billion a year earlier; Global Business Sales, selling to finance, HR, supply-chain and legal leaders, grew 3% to $1.25 billion [6]. The technology-facing node is where the flatness concentrates.

The wallet-retention series over time is what gives the moment its shape.

Source: reported segment KPI disclosures across Gartner's FY2019–FY2025 10-Ks; FY2025 and FY2024 figures per the FY2025 filing [7]. Gray line marks the 100% net-expansion threshold.

Two features matter. First, both channels spent the post-2020 recovery well above the 100% line — GBS wallet retention peaked at 115% in 2021, meaning retained clients were spending 15% more each year without a single new logo — and that is the tailwind that had powered mid-teens revenue growth. Second, the only prior dip below 100% was 2020, the pandemic air pocket, which fully reversed the following year. That precedent cuts both ways: it shows a sub-100% reading can be a cyclical trough that snaps back, and it also means 2025 is only the second time in a decade the flywheel has run in reverse. Which of the two 2025 resembles is the question the rest of this chapter, and the acts after it, are built to answer.

One Node, Not One Company

The most useful test of whether a stall is execution or condition is whether an independent competitor, run by different people on a different platform, is describing the same thing in the same words. Forrester — the closest pure-play, roughly one-sixteenth Gartner's size — is. On its Q4 2025 call, Forrester reported contract value down 6% and full-year revenue down 8%, with net retention at 87% [8]. Its CFO attributed the contract-value decline to "low wallet retention, primarily driven by lower enrichment numbers" [9] — enrichment being Forrester's word for the same expansion motion Gartner tracks in wallet retention. Two firms, the same leading metric, the same named cause: shrinking spend per retained client, not client flight.

The softness also has a clear address inside the information-services industry, and it is the syndicated-research node specifically. The FY2025 revenue-growth scoreboard makes the split unambiguous.

Source: reported financials, FY2025 company filings (peer comparison).

The diversified data, ratings and events majors kept growing — Informa up 13.7%, S and P Global up 7.9%, Thomson Reuters up 3.0% — while the research-and-advisory pure-plays shrank: Forrester down 8.2%, Information Services Group down 1.2%, Clarivate down 4.0%. Gartner, up 3.7%, sits at the top of the contracting cohort rather than the bottom of the growing one. A caution on that scoreboard: S and P Global, Thomson Reuters and Informa are adjacencies — ratings, legal-and-tax data, professional events — not head-to-head sellers of syndicated IT research, so they belong in the picture as evidence of where the money is still pooling in information services, not as direct rivals. The genuine like-for-like set is Forrester and ISG, with Clarivate a step further out in scientific and IP data. Read that way, the pattern is not that information services is soft; it is that the subscription IT-research niche that is Gartner's core is soft, and the broader data business around it is not.

That is the load-bearing distinction of this chapter. The headline revenue number — 3.7% growth, decelerating from 15.5% in 2021 — reads like a single company losing momentum. The retention and peer evidence says something more specific: a category-wide contraction in expansion spend, concentrated in the technology-facing node, that two independent operators diagnose identically.

The Leader in a Shrinking Peer Set

Inside that contracting node, the competitive result is lopsided in Gartner's favor, and the numbers make the point without adjectives. Gartner earned a 15.8% operating margin in 2025; Forrester ran an outright operating loss at negative 28.5%, and Information Services Group managed 7.3% [10]. Gartner is the only pure-play in the set still turning a healthy profit while growing, and its wallet retention — even at a diminished 96% to 99% — sits nine to twelve points above Forrester's 87%. When sub-scale competitors slip into losses in the same downturn that costs the leader a few points of expansion, that is scale economics doing what scale economics do.

Source: reported financials, FY2025 company filings; Gartner operating margin and operating income per the FY2025 10-K [11].

Two things are true at once, and holding them together is the whole point. Gartner is winning the peer race decisively — extending its lead as the only profitable, growing pure-play — and its own forward moat metrics eroded, with both retention lines below 100% and contract value flat. The erosion is not in the product's economics: the Insights gross contribution margin held at 77% in both 2025 and 2024, unchanged [12]. A Gartner seat still earns roughly the same contribution it always did. What softened is the rate at which clients add and expand those seats — a demand condition, not a unit-cost one.

It is worth being precise about what kind of moat that leaves, because management is unusually candid about it. Gartner's own 10-K states that "limited barriers to entry exist in the markets in which we do business," names "free sources of information that are available to our clients through the internet," and anticipates "more competition with increased adoption of AI services" [13]; its risk factors describe markets "characterized by intense competition," including "information available on the internet free of charge" [14]. The written barrier is low. The demonstrated one — 2,400-plus experts, more than 510,000 direct client interactions a year, a proprietary research corpus no entrant can assemble quickly — is why no one has displaced Gartner in decades even though the legal wall is thin. The moat is brand, scale and proprietary content, and the profit ladder above shows it is real. The question is whether the first force in years capable of making that low written barrier actually bite — free AI answers — is arriving, and that is the last thread this chapter opens rather than closes.

The Federal Air Pocket

Before AI, one isolable and quantifiable slice of the 2025 stall needs to be pulled out, because it behaves like a one-off and can be sized. Gartner's Insights contract value with the US federal government was about $126.0 million at the end of 2025, and less than half of the prior-year federal book was retained in 2025 as government spending cuts hit [15]. A book that more than halved on a base that small does not sink a $5.16 billion contract value on its own, but it is large enough to bend the headline retention figures — and, being a discrete cancellation wave rather than gradual client defection, it annualizes out of the comparison rather than compounding.

The first-quarter 2026 numbers let you measure how much. Excluding the federal government, contract value grew 3.5% in Q1 2026, against reported total contract value growth of 1%; ex-federal GTS wallet retention was 99% versus 97% reported, and the remaining federal book had shrunk to about $114 million [16]. Management sized the drag directly: federal was a roughly 250-basis-point headwind in the quarter, and it expects that business to be flat in 2026 and grow from there as the cancellations lap [17]. Strip the federal air pocket out and the expansion engine is close to flat-to-slightly-positive rather than contracting — which is the crux of the cyclical case: a chunk of the 2025 damage was a policy shock that is already visible reversing in the run-rate.

That is the ex-federal read on its own terms, as a demand fact. Whether management's framing of the federal exposure held up against its earlier statements — how small and diversified it was said to be before it halved — is a matter for the credibility record, and belongs to the reconciliation of what was said against what happened (Said Versus Did).

The Force Both Sides Claim

The one variable that could turn a cyclical trough into a permanent step-down is the same one both Gartner and Forrester are betting on for growth — which is exactly why it is unresolved. Gartner's 10-K is explicit about the threat: large language models "provide substantive content in search results, including AI-generated content, which, if expanded to the areas in which we operate, could reduce the need to enter our websites," and clients "may load our proprietary information into large language models, which could reduce the value of our offerings" [18]. Free, fluent, always-available answers are a substitute for the commodity end of paid research, and the company says so in writing.

The same technology is also the product both firms are selling hardest. Gartner launched AskGartner, an AI tool over its own proprietary corpus, in August 2025, and its defense is that its insights are "available exclusively from Gartner" and cannot be reconstructed from a public model [19] [20]. Forrester goes further in the opposite direction, casting disruption as fuel: "the more disruption, the faster our business model will grow," with unique users of its AI product up 55% year over year and its defense resting on proprietary data, original analysis, and trust [21]. The two closest players in the category read the identical evidence in opposite directions — one names AI a disintermediation risk, the other an accelerant — and both retreat to the same shelter: exclusive content and trusted human experts.

That is where the diagnosis lands and where it stops. The 2025 stall is category-wide, not a Gartner execution failure; a measurable piece of it is a federal cancellation wave already reversing; and the leader is gaining share even as its own expansion metric turns negative. What the evidence cannot yet settle is whether below-100% wallet retention is the pandemic pattern repeating — a cyclical trough that snaps back once federal laps and budgets thaw — or the leading edge of AI substituting for the paid seat, in which case the expansion flywheel does not simply restart. That structural question sits underneath everything the later acts weigh against the price. First, though, the record: five years of what management promised about this engine, set against what it delivered.