ServiceTitan, Inc. [TTAN] — Research

Tier-2 memo · as of 2026-07-29 · spot $78.40 (2026-07-28 close) · framework v1.5.1 This memo issues no position verdict. It scores Criteria and outputs an analysis. The book decides.

0. Headline: the screen's scale was wrong by 2.7x, and the correction inverts its own conclusion

The Tier-1 screen recorded TTAN at a $2.775bn market cap on 35,397,085 shares, giving EV/Sales of 2.42x and a valuation margin of +18.9pp. Every one of those numbers is wrong, because the share count is a pre-IPO cover-page figure.

Where the share count came from. TTAN's XBRL CommonStockSharesOutstanding tag has exactly two observations, both from the Q3 FY2025 10-Q filed 2025-01-14 — a filing covering the quarter ended 2024-10-31, i.e. six weeks before the December 2024 IPO: 34,185,388 (2024-01-31) and 35,397,085 (2024-10-31). The company has not re-tagged that concept since. Post-IPO, the only share concepts it tags are WeightedAverageNumberOfSharesOutstandingBasic / ...DilutedSharesOutstanding, which read 95,003,992 for the quarter ended 2026-04-30. The scanner took the newest observation of the tag it was looking for and got a pre-IPO number with no staleness signal attached — the exact failure mode the brief warned about.

The dual-class dimension compounds it. TTAN has Class A (1 vote), Class B (10 votes) and Class C (0 votes, none issued). Dimensional XBRL tagging does not survive SEC aggregation, so the aggregated concept cannot express A + B. The authoritative figure is the 10-Q cover page, read directly:

"As of May 31, 2026, the registrant had 82,746,425 shares of Class A Common Stock … 12,651,154 shares of Class B Common Stock … and no shares of Class C Common Stock … outstanding." — Form 10-Q for the quarter ended 2026-04-30, filed 2026-06-05

Scale cross-check. Q1 FY27 net loss $(22,818)k ÷ 95,003,992 weighted shares = $(0.24) — which is exactly the filed basic-and-diluted EPS of $(0.24). Against the screen's 35.397m shares the same net loss implies $(0.64), which appears in no filing. The verified count ties; the screen's does not.

The net-cash error is separate and additive. The scanner used current cash ($421.5m at 2026-04-30) less a stale debt balance ($104.0m at 2025-01-31). TTAN voluntarily repaid the ~$107.0m term loan in full in January 2026 prior to the Second Amendment to its credit agreement, and the 10-Q states "As of April 30, 2026, no loans were outstanding under the Amended Credit Agreement." TTAN is debt-free with a $250m undrawn revolver to January 2031.

What this does to the screen's own conclusion is the point. Re-running the screen's own assumptions (terminal margin 11.2%, exit 23.5x EV/EBIT, 5y, WACC 10%) on the corrected share count:

The screen's PASS was not an assumption call. It was entirely produced by the share-count error. The name still passes on my own (differently argued) terminal assumptions — see §7 — but it does so with a +4.3pp margin, not +18.9pp, and the gap between those two numbers is the whole value of this memo.

1. What the business is

ServiceTitan sells vertical SaaS to residential and commercial trades businesses — HVAC, plumbing, electrical, roofing, landscaping, pool, and adjacent field-service verticals. It began with plumbing and has extended vertical by vertical. Substantially all revenue (>95%) is US and Canada.

The product is deliberately end-to-end rather than a point tool. Management describes covering five "centers of gravity" inside a trades business — CRM, FSM (field service management), ERP, HCM and FinTech — under three commercial offerings:

• Core — the entry SKU: call tracking, scheduling, dispatching, end-customer comms, marketing automation, estimating, job costing, sales, inventory, payroll integration.
• Pro products — add-on modules (Marketing Pro, Dispatch Pro, Sales Pro, Contact Center, Pricebook Pro etc.), some subscription, some consumption-priced.
• FinTech — payment processing and third-party consumer financing.

Named competitors, from the 10-K itself: Salesforce, SAP, FieldEdge, Workwave, ServiceTrade, AccuLynx, BuildOps, HouseCall Pro, JobNimbus, Jobber. The only close public comparator is EverCommerce [EVCM] — SMB service software with a payments attach — which is currently shrinking (−1.7% demonstrated CAGR). That divergence is itself informative and is the shape of the bear case.

Scale and customer concentration (from the 10-K / 10-Q, verbatim metrics)

Note what is not disclosed: seat count, technician count, ARPU, price increases, or the actual (rather than floored) value of NRR or GDR. §4 returns to this.

2. The mechanism — named, and it is mostly not ServiceTitan's

The mechanism is GTV: ServiceTitan's revenue is a levered claim on the dollars its customers invoice.

This is not a narrative reading; it is the disclosed pricing architecture. From the 10-K:

"Pricing for these subscriptions are driven by the features included in the package and are linked to the size of the customer's business, generally based on the number of field technicians at the customer but in some cases directly tied to the number of end customers or the customer's revenue."

And on the usage line:

"Usage-based services primarily consist of payment processing … We receive a fee from the third-party processors, depending on the size and type of the transaction, which we recognize net of interchange and other direct expenses."

Management is explicit that GTV, not customer count, is the operating objective:

"Because of these dynamics, we focus on increasing the GTV on our platform, rather than new customer count."

Growth decomposition — as far as the disclosure permits, and no further

Read this carefully. Roughly 80–90% of ServiceTitan's revenue growth is its customers' business growth. The company-controlled residual — product attach, pricing, net new logos — contributed +4.6pp in FY26 and +2.0pp in the most recent quarter. The take rate on GTV is rising, but slowly (+0.6bp/yr).

Is growth seat expansion or pricing? It is neither, in the usual sense. Subscription pricing is a function of the customer's technician headcount and business size, so subscription revenue expands when the customer hires — which management measures via NRR, not via a disclosed seat count. There is no disclosed list-price increase in any filing since the IPO, and no disclosed ARPU. Expansion is therefore best characterised as volume-linked seat/size expansion inside existing accounts, not price.

Why this matters more than it looks. The bull framing of vertical SaaS is that the vendor controls its own destiny through attach and pricing. TTAN's disclosure says the opposite: its dominant growth input is US residential trades activity — HVAC replacement cycles, plumbing and electrical service demand, home services spend. The 10-K's own risk factors name "changes in spending on home and commercial services" and "consumer reliance on online marketplaces" as direct revenue risks. This is a cyclical exposure dressed as a software compounder, and it is the named cause of the bear case in §8.

The product-cycle overlay — Max and Atlas

Two named, dated product initiatives sit on top of the GTV mechanism:

• Atlas — "an agentic AI layer that represents the next evolution of our Titan Intelligence AI engine," introduced in fiscal 2026, woven into Core and Pro and integrated into FinTech. Contractors interact in plain English. Designed "over time" to handle tasks and decisions autonomously.
• Max — a bundle that "packages our most advanced functionality and AI features with expert guidance to unlock AI automation." Piloted in fiscal 2026. In the Q1 FY27 release (2026-06-04) President Vahe Kuzoyan said: "During Q1, we more than doubled the number of locations on Max; we're optimizing our internal processes, automating customer onboarding and expect to again double the number of locations on Max during Q2."

Max is the only company-controlled lever with a disclosed, quantified adoption curve — and it is disclosed only as a growth rate off an undisclosed base. "Doubling locations" from an unstated starting point is not a sizeable number; it is a direction. It is the single most important disclosure to watch at the Q2 FY27 print (§ Catalyst Calendar): if management gives an absolute location count or an ARR contribution, the company-controlled residual becomes underwritable. If it stays a growth rate off an unstated base for a third quarter, treat it as immaterial to the model.

3. Subscription vs payments mix — and why the terminal-margin fear is misplaced

This was the brief's central substantive question. The answer inverts the usual worry.

The mix is remarkably stable

Growth by line: subscription +28.1% / +25.9% / +24.2% (FY25 / FY26 / Q1 FY27) — steadily decelerating. Usage +23.9% / +22.6% / +29.3% — re-accelerating in the most recent quarter, the only line that did.

So the payments attach is not diluting the mix and is not a small experiment: it is a stable ~22% of revenue that just posted its fastest growth since the IPO.

The critical accounting fact: usage revenue is booked NET of interchange

"We receive a fee from the third-party processors … which we recognize net of interchange and other direct expenses which are passed onto the customer." (10-Q, Revenue Recognition)

This is the whole answer to the terminal-margin question. A payments business booked on a gross basis (the Toast/Shopify pattern) carries interchange through cost of revenue and structurally caps blended gross margin in the 20–40% range on that line. TTAN books net, so its payments revenue arrives at near-software gross margin. The reported margin structure confirms it directly:

Platform gross margin is 78.7% and rising ~3pp/yr, with the payments mix flat at ~22%. The payments attach is not the drag. The drag is entirely professional services, a 3.1%-of-revenue line running at −136% gross margin — i.e. a ~$11.3m/quarter onboarding subsidy sitting inside cost of revenue, which management explicitly says it runs at a loss on purpose ("Professional services and other cost of revenue historically has exceeded professional services and other revenue as we invest in providing customers with implementation and onboarding services"). That line has shrunk from 5.3% to 3.1% of revenue in two years, which is by itself mechanically worth roughly +2.5pp of blended gross margin over the period.

Conclusion for the terminal margin: TTAN should be modelled as a high-gross-margin software company, not as a payments hybrid. The screen's terminal-margin assumption of 11.2% ("industry median (pre-profit)") is a generic placeholder that ignores a 78.7% platform gross margin. §7 rebuilds it bottom-up at 20%, and shows the answer is entirely governed by that parameter.

4. Accounting quality — is the reported growth real?

Summary: yes, the revenue is real, organic and recurring. The reported operating-margin improvement is mostly real too, but ~5pp of the reported loss is a market-condition equity award that is very unlikely to deliver shares. Two genuine disclosure defects exist, both of the "floor, not a number" type.

4.1 Growth is organic — acquisitions are immaterial

PaymentsToAcquireBusinessesNetOfCashAcquired: $1.184m (FY2025) and $19.781m (FY2026), against $961m of FY26 revenue — 2.1% of revenue. The only FY26 acquisition is Conduit Tech (an HVAC design and sales-proposal tool, closed in Q3 FY26): total identifiable intangibles $6.97m, goodwill +$14.4m. Goodwill is otherwise static at $845.836m from the FY2023 Aspire acquisition ($589.7m paid in FY2023, pre-IPO).

Nothing resembling the 10x Genomics or Applied Optoelectronics pattern is present. There is no settlement income, no milestone revenue, no undisclosed distributor concentration, and no customer concentration —

2,000 customers account for >60% of billings, and ~10,800 Active Customers for >97%. Revenue is subscription-and-transaction, recognised ratably or at transaction, from a diffuse SMB base.

4.2 Receivables quality — clean, with a small Q1 timing build

TTAN carries a contract asset ($64.6m at 2026-04-30) that is now slightly larger than accounts receivable ($63.4m) — unbilled revenue where the right to consideration is conditional. Both must be assessed together.

On an annual basis DSO is flat at ~43 days — a good number for SaaS. The Q1-on-Q1 comparison shows +3.8 days, which is a quarter-end timing build rather than a trend, since the annual measure did not move. The allowance for doubtful accounts fell from $11.963m to $9.585m over the quarter while AR rose, which is worth noting as a mild source of quality drag but is immaterial at this size (allowance is 13.1% of gross AR, down from 17.6% — still conservative in absolute terms). No receivables red flag.

Deferred revenue is trivially small ($19.0m) because the business bills monthly-to-annually and much of it is transactional — so deferred revenue carries no signal here and should not be used as a bookings proxy.

4.3 The operating-margin improvement is mostly real — with one large caveat

The FY25→FY26 GAAP improvement is +12.2pp, while SBC/revenue improved only 0.7pp. So the improvement is ~11.5pp of genuine operating leverage, not an SBC artifact. The TTM measure confirms it independently: TTM to 2026-04-30 operating margin is −14.3% versus −27.7% for the prior TTM — +13.3pp, so the improvement is continuing past the FY2025 IPO-cost comp base rather than being an artifact of it.

Cash-opex leverage is real and checkable at the line level. Q1 FY27 vs Q1 FY26, against revenue +24.6%:

S&M and G&A are the source of the leverage; R&D is being reinvested and is growing faster than revenue on a GAAP basis. On a cash basis R&D still grows below revenue (+18.1% vs +24.6%), so leverage is intact — but the direction is worth monitoring: if the agentic-AI build (Atlas, Max) requires R&D to grow at or above revenue, the entire margin path in §7 slows. Half the R&D increase is SBC, which is a compensation choice, not an efficiency.

4.4 The one genuine anomaly: $53.6m/yr of expense for awards struck 79% above spot

In October 2024, pre-IPO, the company granted each Co-Founder 3,241,544 performance RSUs — 6,483,088 in total, 6.8% of current shares outstanding. From the 10-Q:

"The performance-based RSUs vest based on achieving volume weighted-average closing trading prices of the Company's common stock over a six-month or 90-day period, as applicable, ranging from $140.00 per share to $440.00 per share … Any RSUs for which the applicable stock price hurdle has not been achieved on or before October 21, 2034 will automatically [be forfeited]."

The stock is $78.40. The lowest hurdle is $140 — 78.6% above spot. The highest is $440, 461% above spot.

The GAAP charge is being recognised on a derived service period regardless of whether any hurdle is met, and under ASC 718 a market-condition award's expense is not reversed if the condition is never satisfied:

This is a real distortion in both directions and it must be stated as such. It inflates the reported GAAP loss by ~5.6pp of revenue for awards that, at current prices, will probably deliver nothing — so ex this charge FY26 operating margin is −12.0%, not −17.6%. But if the thesis works and the stock reaches $140+, 6.48m shares (+6.8%) are issued, and any target above $140 must carry that dilution. Neither the GAAP number nor the non-GAAP number handles this correctly on its own; the correct treatment is contingent, and §7 / Trade Construction apply it that way.

For completeness, the other non-GAAP add-backs in FY26 are: amortisation of acquired intangibles $45.2m (4.7% of revenue, from the FY2023 Aspire deal — a real past investment, non-cash, will run off), and "loss on operating lease assets" $11.0m (down from $39.1m in FY25 — abandoned office space; genuinely non-recurring). Both add-backs are conventional; neither is aggressive.

4.5 Two disclosure defects — the "floor, not a number" pattern

These are the closest thing here to the Twist retired-disclosure precedent, and they are real.

1. NRR has been disclosed as ">110%" for four consecutive periods (FY2024, FY2025, FY2026, Q1 FY27) — never as a number. A decline from, say, 118% to 111% would produce identical disclosure. The company's single most important growth metric is therefore structurally undetectable in decline until it breaks 110%, at which point the deterioration is already several quarters old.
2. GDR has been disclosed as ">95%" for three consecutive fiscal years — same structure, same blind spot.

Two secondary items in the same family: - Active Customers (~10,800) is disclosed annually, at 2026-01-31 only — there is no quarterly series, so logo growth cannot be tracked intra-year. - NRR excludes acquired customers until a full quarter after the one-year anniversary, which is correct practice but means the metric is definitionally insulated from acquisition effects — fine here, since acquisitions are immaterial.

None of this is misconduct. It is standard newly-public SaaS practice. But it means the two metrics an investor would use to detect the thesis breaking are censored from above, and the memo should be honest that the early-warning system this name offers is weaker than the metric list implies. The observable substitutes are GTV growth (quarterly, disclosed) and the usage take rate (derivable quarterly), and those are what the monitoring plan in the Catalyst Calendar keys on.

5. Transcript mention-frequency — LOW CONFIDENCE, and here is why

Source: 8-K Exhibit 99.1 earnings releases from EDGAR, narrative section only (everything before "About ServiceTitan"), single source, no mixing. Alpha Vantage's EARNINGS_CALL_TRANSCRIPT was attempted first and returned the daily-quota exhaustion message; no transcript could be retrieved, so no call transcript is in this corpus.

Window: all seven quarterly releases that exist post-IPO (Q3 FY25 → Q1 FY27). This is not eight quarters; it is the company's entire public history. Narrative word counts: 603, 785, 538, 527, 506, 624, 544 — near constant, so raw counts are comparable without normalisation, and per-10k-word figures are reported anyway.

Zero across all seven quarters: AI (as a standalone token — it appears only inside "Agentic Operating System" / "agentic"), Titan Intelligence, Atlas, FinTech, payments, pricing, seat, technician, churn, attach, cross-sell, HVAC, plumbing, electrical, macro, housing, Aspire, Conduit.

The observations, stated as open questions rather than conclusions (per method):

1. Why did the product vocabulary collapse? The IPO-quarter release (Q3 FY25) named specific Pro SKUs — "Sales Pro", "Contact Center", "Pro products". Not one has appeared since. The framing moved from product attach to GTV plus agentic AI.
2. Why does GTV keep climbing while the trade verticals disappear? GTV goes 1 → 0 → 3 → 3 → 3 → 4 → 4 while roofing, residential, commercial each appear once and vanish. Management is standardising on a single volume metric.
3. What is "Max" actually worth? It enters the vocabulary in Q4 FY26 and doubles to 2 mentions in Q1 FY27, alongside the only quantified adoption claim in the release. Directionally the strongest emerging term.

The honesty constraint, stated plainly. These are ~550-word press releases containing two CEO/President quotes. A count of "1" versus "0" is a single sentence. The trend cannot be measured quantitatively on this corpus, and I will not pretend otherwise. What the corpus does support is a weaker but still real claim: in the most space-constrained document a company produces, four terms that had never appeared (Max, agentic, Agentic Operating System, velocity) all entered simultaneously in Q4 FY26 and all persisted into Q1 FY27, while the entire Pro-product vocabulary from the IPO quarter went to zero and stayed there. That is a deliberate repositioning of the equity story from product-attach to AI-plus-volume, and it happened in March 2026. Whether it is substance or packaging is exactly what the Q2 FY27 Max disclosure will answer.

Confidence: LOW. This metric should be re-run against a proper call-transcript corpus when Alpha Vantage quota is available. The finding above is a hypothesis for that re-run, not a result.

6. Quality Criteria — INFLECTION archetype

Archetype: INFLECTION. High and rising gross margin, negative but rapidly improving operating margin, growth above 18%. Value sits almost entirely in the terminal period. F-score and gross-profitability levels are demoted to context, per criteria.md.

Quality Criteria: PASS, with the acceleration test explicitly failed. Growth is above the threshold on level but the second derivative is negative and management has guided it to +18.1% for FY27. That is not a disqualification under the stated rule, but it is the single most important thing a reader should carry into §7 — the required path and the actual path are converging.

7. The valuation summary

Full working in TTAN_Valuation.md. Headline:

• EV $7,058m · EV/Sales 6.96x TTM (corrected). Own-history EV/S percentile: 31st.
• Implied path (reverse DCF, primary output): the price requires a 19.8% 5-year revenue CAGR, holding terminal EBIT margin at 20% and exit multiple at 22.7x EV/EBIT, WACC 10%.
• Margin = demonstrated − required = 24.1% − 19.8% = +4.3pp → PASS, thin.
• Implied multiple compression: required exit EV/Sales of 3.81x against 6.96x today — the price embeds a 45% de-rate over five years. On the non-GAAP EBIT the company guides to for FY27 ($144.5m), the stock is at 48.8x, compressing to 22.7x — a 53% de-rate.
• 12-month target multiple anchor: UNIDENTIFIED. 1.31 years of usable own-multiple history, spanning a monotonic post-IPO de-rate from 14.67x to 5.03x. No peer median substituted.

8. Downside Criteria — the permanent-loss case, with a named cause

Type: MEASURED. Logged and scored; blocks nothing.

Named cause: a US residential-trades volume contraction, transmitted directly into TTAN revenue through GTV.

This is not a volatility statement and it is not generic macro. It is the mechanism from §2 run in reverse. ~80–90% of TTAN's revenue growth is its customers' invoiced volume. GTV growth is already the binding input, and the company's own risk factors name "changes in spending on home and commercial services" and "consolidation of trades businesses" as direct revenue risks. A residential-services downturn — deferred HVAC replacement, deferred discretionary home repair — hits TTAN in three places simultaneously:

1. Usage revenue falls with GTV, immediately and roughly one-for-one (21.8% of revenue, no lag, no contract protection).
2. Subscription revenue falls with a lag, because pricing is tied to technician headcount — customers who shed technicians downgrade at renewal. NRR compresses from >110% toward 100% or below.
3. The >95% GDR floor breaks, because SMB trades businesses fail in downturns, and TTAN's base is SMB.

The comparator that makes this concrete: EverCommerce [EVCM], the closest public analogue (SMB service software with a payments attach), currently at −1.7% demonstrated revenue CAGR on a 4.22x EV/Sales. That is what the shape looks like when SMB service-software volume growth stops. TTAN is a much better asset with a far higher gross margin, but the failure mode is demonstrated in the same end-market.

Quantifying it. From the §7 grid: at a 16% terminal margin and an 18.0x exit multiple — a recession-consistent combination in which margin expansion stalls and the multiple compresses — the price requires a 31.2% CAGR, a −7.1pp margin. If revenue growth also halves to ~12%, the EV that supports is roughly $1,014m × 1.12⁵ × 0.16 × 18.0 / 1.10⁵ ≈ $3.2bn, versus $7.06bn today: an equity value of ~$3.6bn including net cash, or roughly $38/share — a 51% permanent impairment.

Probability: 25%. Basis: US residential trades activity has not contracted in the company's disclosed history (GTV +19.9% FY26 on top of a large FY25 base), so this is a forecast, not an observation; but the exposure is undiluted and undisclosed as a risk metric, the customer base is SMB, and the company has never operated as a public company through a downturn. The 25% reflects a genuine cycle risk in a name that has never been tested, not a view that a downturn is imminent.

Going-concern: NO. $421.5m cash, zero debt, $250m undrawn revolver to January 2031, positive operating cash flow ($110.1m FY26), non-GAAP operating income positive and guided higher. There is no financing risk here. The bear case is a de-rating and a growth stall, not a solvency event.

Secondary bear cases, unquantified but named: - Agentic AI compresses the seat model. TTAN prices on technician headcount. If AI genuinely automates dispatch, call-handling and back-office work — which is exactly what Atlas and Max promise — TTAN's own product reduces the metric it bills on. Management has not addressed this tension in any filing. It is the most interesting unhedged risk in the name. - Co-Founder RSU dilution on success. 6,483,088 shares (+6.8%) issue if the stock clears $140 VWAP. Any bull case above $140 must be struck on ~102m shares, not 95.4m. - Voting control. Co-Founders hold ~73% of voting power via 10-vote Class B. Public holders have no governance recourse; the name will not be an activist or takeover candidate at any price.

9. MEASURED Criteria — scored, blocking nothing

10. What is unsupported, and what I could not establish

Stated explicitly so it is not mistaken for evidence.

Unsupported / not established: 1. Consensus estimates. None retrieved. Every forward number in this memo is company guidance or my own build off the disclosed guidance-beat pattern. There is no Street comparison anywhere in this memo. 2. The 12-month price target has no identified multiple anchor and is reported as a conditional band, not a point estimate. See TTAN_Valuation.md §4. 3. Transcript mention-frequency runs on 550-word press releases, not call transcripts. LOW confidence. The Pro-vocabulary decay and the agentic-AI emergence are hypotheses for a proper re-run. 4. The peer EV/EBIT set used for the exit multiple comes from the same Tier-1 scanner whose share count on TTAN was wrong by 2.7x. I verified TTAN's own inputs against primary filings but did not verify the 7 comparators' share counts. Their market caps are individually plausible against known scale (VEEV $32.7bn, PAYC $7.7bn, MANH $9.9bn, SSNC $18.8bn, TYL $14.1bn), and the TTAN defect was specific to a recent IPO with a stale pre-IPO tag — but the exit multiple is MEDIUM confidence, not verified. 5. Max adoption is not sizeable. "More than doubled the number of locations" off an undisclosed base is a direction, not a number. No revenue, ARR or location count has been disclosed for Max or Atlas. 6. Seat count, ARPU and pricing are not disclosed in any filing. The seat-vs-price decomposition in §2 is inferred from the disclosed pricing architecture and the GTV/revenue gap; it is not directly observable. 7. NRR and GDR are floors, not values (§4.5). Their trend is unobservable. 8. The Q2 FY27 earnings date is ESTIMATED from a six-observation filing-lag regression, not announced. 9. WACC of 10% is the framework default, not a CAPM build. At 12% — arguably more appropriate for a 54%-vol zero-debt pre-profit name — the required CAGR rises to 22.0% and the margin falls to +2.1pp. The valuation conclusion is sensitive to this and it is not independently derived.

What would change the analysis fastest: - An absolute Max location count or ARR contribution at the Q2 FY27 print → makes the company-controlled growth residual underwritable and could move the exit-multiple argument materially. - Any quarter in which GTV growth falls below ~15% → the mechanism in §2 says revenue follows within two quarters, and the +4.3pp margin disappears. - NRR printing at exactly "110%" rather than ">110%" → the censored metric breaking its floor.

Sources

All figures traced to primary documents. No third-party financial data was used for TTAN itself.