Fit

Fit — Gartner, Inc. (IT)

Does not fit the framework (P1 not met)

Gartner does not fit the framework. The year-10 durability gate (P1) is not met — four jurors across two model families voted not_met with a trimmed-mean p_year10 of 0.60 and a spread of 0.05 — and under the framework's rules that gate forces the verdict on its own; nothing offsets it. Confidence is high: two model families agreed, the trial was order-stable, and load-bearing spreads were at most 0.15. No exclusion was triggered, the watchlist_only flag is off, and the name-mask probe raised no prior-driven-risk flag.

Universe and exclusions — unsoftened

Gartner clears both universe lines. It is a Delaware-incorporated company whose common stock is listed on the New York Stock Exchange under the symbol IT [1], a primary US listing (U1 met, unanimous). USD market capitalization computes to $10.599B (75,605,000 shares × $140.19 close on 2026-07-17), above the $10B scale line (U2 met, unanimous) — but the cushion is only about $0.60B (~6%), and it rests on a stock that has fallen 74.6% from its 2024 peak, so a further modest leg down would put the name out of universe.

No exclusion was hit. X1 (auto OEM) — Gartner is a research and advisory firm selling Insights, Conferences, and Consulting to 13,000+ enterprises, with zero vehicle-manufacturing revenue [2]; not a hit. X4 (consensus-saturated story) — the stock trades at ~1.6x sales and ~10x forward EPS after a 74.6% de-rating on a majority-hold sell side (3 buy / 9–10 hold), the antithesis of the extreme-multiple euphoria X4 tests for; not a hit. X2 (promotion pattern) and X3 (structural decline) are evidential and are treated in the pillar ledger below; both resolved not_met. S1 (China dependence) is a sensitivity flag only and was not raised: China is not separately disclosed, and any China exposure sits inside an "Other International" line that is 11.9% of revenue and spans all of Asia-Pacific and Latin America [3].

Note the four exclusions (X1, X2, X3, X4) show a label split — three jurors recorded not_met and one recorded not_hit — but that is wording, not a factual dispute: every juror found the exclusion did not apply.

Pattern match

On surface features this is the shape of an event-driven dislocation with a high-yield entry: a dated adverse catalyst (two guidance-day crashes), a ~77% peak-to-trough drawdown, an adjusted FCF yield sitting within a rounding error of the balance-sheet bar, and an earnings numerator that held while the multiple collapsed. That is the closest resemblance among the framework's constructive setups.

It fits none of the four patterns as a qualifying entry, because the setup fails on the pattern's own core checks rather than its surface. The year-10 durability gate is not met; the temporary-vs-permanent diagnosis (P5) came back leaning permanent (p_temporary 0.38); and the adjusted yield is below its bar today. A qualifying dislocation requires the impairment to be diagnosed as temporary with the franchise intact for a decade — here both of those load-bearing checks fail, so the resemblance does not convert into a fit. This is framing only; it does not alter the tally's verdict.

The pillar ledger

Source: deterministic fit tally (per-criterion aggregate verdicts, trimmed-mean probabilities, and spreads).

Year-10 durability (P1) — the gate that decides it

Not met (unanimous, both families; p_year10 0.60, spread 0.05). The revenue disqualifier is false: FY2025 revenue grew 3.7% to $6,497.2M with no consecutive-decline years, off a model that is asset-light, high-margin, and customer-funded — 78% of revenue is 77%-gross-margin Insights subscriptions that clients prepay, with contract liabilities of $2,841.6M and capex only 1.8% of revenue [4]. What puts the year-10 leg in genuine doubt is that the forward engine has stalled: total Insights contract value grew only ~1% FX-neutral, GTS contract value was flat ($3,910M vs $3,911M), and wallet retention fell below the 100% net-expansion line for the first time in the window — GTS 96% (from 102%), GBS 99% (from 106%) [5]. The 10-K itself concedes "Limited barriers to entry exist in the markets in which we do business" [6] and newly names LLMs that "could reduce the need to enter our websites" [7]. Operating margin compressed from 20.9% (FY2023) to 15.8% (FY2025) on flat contract value and a $150M Digital Markets goodwill impairment [8].

Strongest surviving counter-fact: Gartner is winning the peer race decisively — FY2025 revenue +3.7% at 15.8% operating margin against the closest pure-play, Forrester, at −8.2% and a −28.5% operating margin [9] — and part of the FY2025 retention hit is an isolable US-federal/DOGE one-off (Insights federal CV ~$126M, less than half retained), with ex-federal CV still growing ~3.5% in Q1 FY2026. That is why the temporary-vs-permanent diagnosis is genuinely two-sided rather than settled against durability — but the gate resolves on genuine doubt, and the doubt is present. Full treatment: The Engine Stalled and The Cash Machine.

FCF consistency (P2)

Met (unanimous). Hand-computed adjusted FCF (operating cash flow − capex − SBC − trailing five-year average acquisitions) held a tight band across FY2021–FY2025: a ~$1,034M five-year average, coefficient of variation ~12.4%, no negative years, and a max/min of only 1.37x [10]. The deterministic fcf_stability feature was not_computable (the cash-flow feed lacked capex/SBC/acquisitions), so every figure was rebuilt from the primary cash-flow statements in the FY2021 [11] and FY2022 [12] 10-Ks.

Strongest surviving counter-fact: the low-volatility record is a modern-regime phenomenon roughly five years long, not a full cycle — stretching the window to FY2019–FY2025 raises the CV to ~31%, because FY2019 adjusted FCF was only ~$321M in the CEB-integration era [13]. Full treatment: The Cash Machine.

Dislocation + yield (P3)

P3a identifiable event — met (unanimous). Two dated adverse events anchor the fall. On Aug 5, 2025 Gartner cut full-year guidance and named US-federal/DOGE weakness as the largest headwind; the stock fell 27.6% in one session on ~4.5M shares [14]. On Feb 3, 2026 the Q4 print introduced a weak 2026 guide against a "challenging selling environment" and the stock fell 20.9% on 7.31M shares — the single highest-volume day of the entire fall [15].

Strongest surviving counter-fact: the cuts reflected real deceleration — Q2 CV growth of only 5% and wallet retention below the net-expansion line [16] — and roughly half of the peak-to-trough loss came as event-free drift before Aug 5, so the setup is a hybrid (a slow multiple derating plus two fear-driven guidance legs), not a clean single-trigger dislocation.

P3b capitulation — met (unanimous; reference line, not a gate). Traded volume during the fall reached 7.17x the trailing median, far above the 2x reference line, centered on the two guidance-day crashes. The only caveat: heavy volume characterizes the event legs, not the ~50% of the decline that came as lower-volume drift.

P3c yield vs bar — not met (unanimous). Net-debt/EBITDA of ~1.0x classifies the balance sheet as moderate, setting the bar at 10% (the deterministic feature returned "unknown" because EBITDA was missing, so this was hand-classified). Current adjusted FCF yield computes to 9.61%, about 39 bps below the 10% bar; the three-year-average adjusted yield reaches 9.96%, ~4 bps short [17]. Strongest surviving counter-fact: the yield is high only because market cap fell ~77% (from ~$45B at the peak, ~2.7% yield) to $10.6B; FCF itself fell 15% in FY2025, so the yield "jumped toward the bar" via price, not via growing distributable cash.

P3d forward path — met (unanimous; p_forward_bar 0.62, spread 0.01). Consensus puts unadjusted FCF yield at 11.3–12.6% through FY2028 and adjusted yield crossing 10% by FY2028, so the underwrite is consensus-backed [18]. Strongest surviving counter-fact: consensus assumes FCF essentially flat FY2025–FY2027 with the decisive step-up deferred to FY2028, the last year before coverage drops out — and the most recent actual FCF fell 15%, so the flat-or-better path is already under pressure. Full treatment: Priced For Impairment.

Balance sheet + self-help (P4)

P4a outlast + allocation — met (unanimous). All $3,005.0M of debt is unsecured senior notes laddered 2028–2035 with no maturity before the $800M 2028 notes, the $1.0B revolver is fully undrawn, and net leverage is under ~2x, so debt service does not compete with repurchases [19] [20]. Strongest surviving counter-fact: the direction of pressure is the opposite of the criterion's fear — FY2025 buybacks of $1,991.1M exceeded operating cash flow of $1,290.4M and were part-funded by $799.9M of new 2031/2035 notes, lifting net debt from $526.8M to ~$1.26B and thinning equity to $319.9M [21].

P4b repurchase engine — met (unanimous). Executed cash buybacks totaling ~$6.03B over FY2021–FY2025 drove diluted shares down 16.0% (90.017M → 75.605M), with a record $1,991.1M in FY2025 and $535M more in Q1 FY2026; repurchases outrun SBC issuance roughly 10x, and the feature flags share_count_trend.rising = false [22] [23]. Strongest surviving counter-fact: the record FY2025 repurchase was debt-funded and executed into a falling knife — Q4-2025 open-market buys averaged $239.06/share against a ~$140 price by mid-2026 — so per-share value creation depends on whether the ~75% drawdown was mispricing or a rerating.

P4c dividend cover — not applicable (unanimous). Gartner pays no dividend; the return case runs entirely through buybacks [24]. Full treatment: The Buyback Bet.

Diagnosis (P5)

Not met (unanimous; p_temporary 0.38, spread 0.08, trial-carried). The probability that the impairment is temporary was produced by an adversarial trial — two opposing cited briefs, three independent judges — which returned a mean p_temporary of 0.393 and a tally value of 0.38, with contested = false. The earnings numerator held: adjusted operating income was roughly flat (FY2025 $1,025.7M reported + $150.0M goodwill impairment = $1,175.7M, vs FY2024 $1,156.3M), and operating cash flow fell only ~13% even as reported net income fell 42% to $729.2M on one-off distortions [25]. But a DCF-lite on FY2025 FCF of $1,175.2M at a 10% discount rate shows the market already capitalizes a near-permanent diagnosis: price (~$139 no-growth NPV) sits at the no-growth value and only ~$30/share above the permanent-erosion case (~$110), against a temporary NPV of ~$212 [26].

Strongest surviving counter-fact: the operating driver that broke is contract-value growth, which decelerated 16/12/8/8/1% over five straight years while wallet retention fell below 100% in both salesforces — broader than the isolable federal book and consistent with structural pressure, not a removable one-off [27]. Full treatment: Priced For Impairment.

Instrument context (I1)

Not verifiable (unanimous; facts only, never blocks a pillar). Gartner is an S and P 500 large-cap with an active listed-options market and a dated implied-volatility term structure (30-day IV ~65.8%, 180-day IV ~56.9% as of 2026-07-29, put-call open-interest ratio 0.595, per AlphaQuery), corroborated by run-dated realized 30-day volatility of ~53%. The 30-day IV sits in the elevated 60–70 reference band. The criterion is not_verifiable because the exact long-dated LEAPS expiration list and per-contract open interest/spreads could not be directly read — Barchart, Market Chameleon, OptionCharts, and Nasdaq option-chain pages all bot-walled — so LEAPS existence is inferred from large-cap status plus the observed IV term structure, not a directly read 2027/2028 chain. Full treatment: Priced For Impairment.

What a 3x-in-3-years would require

The tally records: "Re-rating math unavailable because the applicable bar or normalized adjusted FCF is missing." The bar, normalized_adjusted_fcf, implied_market_cap_at_bar, and upside_to_bar_pct fields are all null, because the deterministic adjusted-FCF-yield feature was not_computable (no annual FCF/capex/SBC series; EBITDA missing for FY2025). The framework's price-at-bar-yield test therefore cannot be rendered from a canonical feature.

The closest available arithmetic is the DCF-lite the Damage Math pillar carried (FY2025 FCF $1,175.2M, 10% discount rate):

Source: derived from reported financials (FY2025 Consolidated Statements of Cash Flows) [28].

A 3x in three years from $140.19 implies roughly $421/share. Even the temporary-diagnosis NPV (~$212, ~1.5x, near the sell-side high of ~$203) falls well short of that, and the permanent case (~$110) is below today's price. On the modeled paths the framework's 3x target is not reached on any scenario — it would require FCF to compound materially faster than the temporary case assumes and consensus to concede a full re-acceleration of contract value to double digits. For base-rate context on how deep this drawdown is and how such episodes have recovered, see Priced For Impairment. This is the framework's target test stated as arithmetic, not a recommendation.

Contested and undetermined

Nothing was marked contested; the flags list is empty. The one criterion that could have been contested — the P5 diagnosis — was pressed by the adversarial trial and returned contested = false (three judges at p_temporary 0.36 / 0.38 / 0.44; order-stability gap 0.05), landing at 0.38.

Nothing resolved to cannot-determine, so there is no named missing datapoint that flipped a verdict. One criterion, I1 (instrument context), is not_verifiable — options existence and IV term structure are established from a dated source, but a directly read far-dated option chain could not be pulled (vendor bot-walls); I1 never blocks a pillar and only drives the watchlist overlay, which is off here.

Provenance

Field	Value
Jury seats	a = claude, b = claude, c = codex, d = codex; masked = claude
Model families	claude, codex
Gate agreement	P1 unanimous not_met across both families (spread 0.05)
Load-bearing spreads	≤ 0.15 (P1 0.05; P3d 0.01; P5 0.08)
Trial order-stability	temporary-first mean 0.36; permanent-first mean 0.41; gap 0.05
Name-mask probe	max probability gap 0.02; no gate criteria flipped → prior_driven_risk = false
Skeptics	38 claims triaged; 17 adversarially pressed → 14 survived, 2 weakened, 0 refuted, 1 unverifiable, 21 triaged-only

Source: deterministic fit tally (provenance block) and the adversarial trial tally.

The verdict was pressed reasonably hard and held: two independent model families reached the same gate call without the tie-breaking help of any single juror, and re-ordering the trial's reading sequence moved the diagnosis probability by only 0.05. The name-mask probe — re-running the jury with the company's identity hidden — changed no gate verdict and moved probabilities by at most 0.02, so the call is driven by the evidence, not by a prior about the name.

The falsifier ledger

These are the standing what-would-change-this conditions carried in the tally. The first five are the framework's generic templates (each a direction on a level already underwritten); the remainder are Gartner-specific, most embedding their own threshold and window (a fiscal-quarter deadline, a contract-value or retention level, a margin floor). Conditions explicitly tagged "Toward temporary" / "Toward permanent" state the direction they would push the diagnosis.

Data gaps

What the run could not answer, from the tally's list:

• China exposure is not measurable. China revenue and assets are not separately disclosed in the FY2025 10-K; geography is reported only as US and Canada / EMEA / Other International, so any China dependence can be bounded (below the 11.9% "Other International" line covering all Asia-Pacific plus Latin America) but not measured.
• The cyclical-vs-structural crux is not resolvable from disclosures. Whether the contract-value/wallet-retention stall is cyclical (US federal/DOGE plus budget pullback) or structural (AI/LLM disintermediation) cannot be settled from filings — no price/volume or churn-reason cohort split is disclosed. This is the P1/P5 crux and was carried into the trial, not settled in the pillars.
• Adjusted-FCF features are not_computable. features.adjusted_fcf, adjusted_fcf_yield, fcf_stability, and balance_sheet_class all returned not_computable (no annual FCF/capex/SBC series; EBITDA missing for FY2025), so the year-10 adjusted-FCF leg, the yield-vs-bar test, and the balance-sheet class all rest on figures rebuilt by hand from the primary cash-flow statements rather than canonical features.
• No matched before/after consensus across the first, largest leg of the drawdown. The estimates-momentum feed only reaches back 180 days (to 2026-01-29), so the estimate path over the Nov-2024 peak to early-2026 leg — where most of the ~74.6% decline occurred — is not in the data; the peak-to-current damage is measured on price, not on a matched consensus revision.
• yield_baseline.per_year is empty, so the fortress "stable-yield-that-suddenly-jumped-to-the-bar" capitulation signature cannot be corroborated from the features file.
• Instrument granularity. The exact long-dated LEAPS expiration list and per-contract open interest / bid-ask spreads could not be directly observed (vendor option-chain pages bot-walled); options existence and IV term structure are established from a dated source plus large-cap status, but a directly read far-dated chain is unavailable.
• Covenant cushion is inferred, not measured. The exact maximum-leverage-ratio threshold in the 2024 Credit Agreement is not disclosed on the read pages, so covenant headroom is inferred from the wide margin (net leverage under ~2x), not measured to the limit.
• Insider-selling window is short. The Form-4 feed only spans 2025-10-01 to 2026-07-01, so a longer-run view of CEO Hall's 10b5-1 selling cadence for the X2 insider prong could not be sourced; within the window there was no Hall open-market selling.

Checked and unremarkable

All eight synthesis scouts returned load_bearing — none was routine or empty — so their findings feed the pillars above rather than this list: accounting/cash-quality, business-economics, capital-allocation, competition-moat, history/track-record, industry, people/governance, and valuation-expectations.

Playbook version

Playbook fcf-dislocation, version 4 (spec playbook/fit-spec, version 2), as frozen for this run.