INTU › analyze
INTU · Analyze from before
A portfolio-specific passage was removed from the public build.
The headline of this pass. The baseline named "Investor Day Sep 17 2026" as its first trigger. It fired, the stock fell again, and the price is now −15% from the baseline date and −58% from the 52-week high. Yet not one of the baseline's seven pre-registered break triggers fired. The FY2026 10-K (filed 2026-09-09) closed the two questions the baseline could not reach, and they resolve in opposite directions: stock-based compensation is $2.056B — 95% of the entire GAAP-to-non-GAAP bridge, which fails the baseline's own upgrade condition; and the FY26 free-cash-flow surge that drove the last upgrade is ~65% a one-time deferred-tax swing, now traced to a specific statute. Set against that, the question this pass was commissioned to answer — does a free QuickBooks tier change the terminal gross margin — resolves clearly in the baseline's favour. Verdict holds; conviction returns to where it stood before the FCF number moved it.
1. What this updates
The baseline was written two days after the FY26 print and before the FY26 10-K was filed and
before the Investor Day. Three things are now available that it explicitly could not reach: the
FY2026 balance sheet, the FY2026 stock-based-compensation line (an UNTESTED row for two
consecutive passes), and the 2026-09-17 Investor Day.
Event list since 2026-08-27
| Date | Event | Ledger rows touched |
|---|---|---|
| 2026-08-27 → 09-22 | Price $348.00 → ~$295.83 (−15.0%). 52w high $703.96 ⇒ −58%; ~−52% YTD | All Price rows, J1 |
| 2026-08-26 → 08-27 | JP Morgan downgrades OW → Neutral, PT $605 → $331 (a $274 cut). Six PT cuts. The only two PT raises come from bears marking to market — Goldman (Sell) $276→$304, Piper (UW) $250→$290 | Price, J1 |
| 2026-09-09 | FY2026 10-K filed (accession 0000896878-26-000037) — balance sheet, SBC, deferred taxes, headcount now readable | St3, St4, St5, T4 |
| 2026-08-17 / ~09-01 | Two securities class actions (not one) — see §8 | 🆕 |
| 2026-09-01 / 09-08 | Form 4s: annual PSU/RSU settlements + two small sales. No open-market buying | St6/St7 |
| 2026-09-17 | Investor Day — FY27 + Q1 guidance reaffirmed, dividend +15% confirmed, ~440bp GAAP margin expansion, "Intuit Intelligence"; CEO concedes FY26 new-customer misses in both DIY tax and QuickBooks Online | S3, S4, T1, T2, J1 |
| 2026-09-17 → 09-22 | −8.3% over three sessions ($318.13 → ~$292) | Price |
| 2026-09-16/18 | Fed delivers its first rate hike in three years; 10-yr Treasury back above 5%. Sector-wide software de-rate on 09-18 (CRWD −2.8%, HUBS −4.2%, NET −4.3%, OKTA −3.4%, INTU −3.7%) | Price — not company-specific |
Two forces drive the price and only one belongs to Intuit. The 09-18 decline was a sector move on a macro catalyst. Recorded as a Price force, not a Structural one — per this command's own rule that a narrative shift is "a Price force wearing a Structural costume."
2. The commissioned question: does QuickBooks Free change the terminal gross margin?
It is a funnel — and the strongest evidence is that Intuit raised prices at the top of the stack in the same month it launched the free tier at the bottom.
The barbell, in one table
On 2026-08-01, three weeks before the print, Intuit repriced QuickBooks Online:
| SKU | Before | After 2026-08-01 | Change |
|---|---|---|---|
| Free | — | $0 | launched, unchanged |
| Lite | — | (price not disclosed) | unchanged |
| Ledger | — | — | unchanged |
| Simple Start | $38 | $38 | unchanged |
| Essentials | — | $85 | — |
| Plus | $115 | $140 | +21.7% |
| Advanced | $275 | $340 | +23.6% |
A company losing a price war does not raise its two highest tiers by 22–24% in the same month it launches a $0 tier. It is segmenting: give the bottom away to widen the funnel, charge the complex customers more. That is the opposite of margin compression — it is mix engineering.
Four independent legs, all primary-source
1. It is a new front door, not a repricing of the base. Goodarzi, Q4 FY26 call: "A key component of our new-to-the-franchise strategy is widening the front door with QuickBooks Free and QuickBooks Lite. These offerings create low-friction entry points to reach millions of businesses earlier in their journey." The target is businesses that are not QuickBooks customers today, not the 8.9M paying base.
2. The free tier is genuinely free but severely metered, and it monetises before it upsells. Per Intuit's own product pages: 2 custom invoices/month · 1 estimate · 2 receipt uploads · 1 contractor · 5 mileage trips · 1 connected bank · 3 reports · single user; no accountant access, no multi-user, no Intuit Assist/Intelligence, no bill pay, no payroll, no inventory. The hook is explicit: invoices become unlimited if you enable QuickBooks Payments — Intuit takes a payments rate before it ever sells a subscription. Goodarzi confirmed the economics: "we have over 20,000 customers… that are active, not only in QuickBooks Free and/or they have upgraded to upper SKUs, and we're monetizing payments."
3. QuickBooks pricing power is still rising, in the guidance.
| Disclosure | Source |
|---|---|
| "QuickBooks Online accounting revenue grew 20% in Q4, driven by higher effective prices, customer growth and mix shift" | CFO, Q4 FY26 |
| FY27 GBS guide: "growth driven by ARPC expansion due to higher effective prices and increased adoption of services" | CFO, FY27 guidance |
| Online ecosystem ARPC growth accelerated to +15% in FY26 | CFO, Q4 FY26 |
4. Gross margin is expanding, not compressing. 78.1% (FY23) → 78.7% → 79.6% → 80.2% (FY26). Four consecutive years. A business being forced into a price war shows it in gross margin before it shows it anywhere else, and Intuit's is at a record. A free cloud-ledger tier carries near-zero incremental COGS; the terminal gross margin is set by the paid mix, and the paid mix is moving up (QBO Advanced / Intuit Enterprise Suite revenue +38%; IES annualised revenue >$145M, 4× YoY).
⚠️ Two corrections to the framing this pass was given
(a) The price cut is in TurboTax, not in QuickBooks. These were conflated. FY27 guidance is explicit: "These actions result in lower tax ARPC in fiscal 2027." That is a real, dated, management-executed price concession — in DIY consumer tax only. QuickBooks got a free acquisition SKU on top of a 22–24% increase at the top. Two different moves, two different segments.
(b) No competitor forced it — on the record, and management pushed back on the premise. Goodarzi, asked directly why the GBS algorithm was cut: "We are creating the pressure. We are not being pressured to make the change." Intuit names no competitor anywhere in the transcript, the Investor Day release, or the coverage; the language is only "low-cost providers" (tax) and "competitive on the low end" (QuickBooks). The obvious referents are Wave / Xero / Zoho at the QuickBooks low end and FreeTaxUSA / Cash App Taxes / H&R Block Free in DIY tax — but that attribution is ours, unverified, and no competitor unit or ARPU data was disclosed by anyone.
The honest caveat — and it is not small
QuickBooks Free is not a margin event, but it is evidence of a real problem it does not yet solve. Total online paying customers grew only +3% to 8.9M, and Goodarzi said plainly: "I'm not happy in 2 areas: DIY tax and the fact that our online paying customers only grew 3%." At Investor Day he went further — "we fell short of our new customer targets" and management "underestimated the resources needed to scale" the Big Bets. 20,000 QuickBooks Free users against an 8.9M paying base is 0.22%, from one month of data. And ~3/4 of mid-market customer additions came from upgrades or desktop migrations — harvesting the installed base, which is a finite lever ([[pattern-margin-expansion-is-a-finite-growth-lever]], same shape).
The sharpest unresolved version of the question, which management itself concedes is open: do the cheap tiers attract genuinely new customers, or do they accelerate existing customers migrating down into lower-revenue SKUs? The Aug 1 price architecture is designed to prevent the second (the free tier is metered to near-uselessness for anyone with a real business), but there is no disclosed cohort data yet. That is the single best question to ask of the Q1 FY27 print.
Reconciling the 440bp margin guide with the price cuts — the arithmetic, not the adjective
| FY2026 actual | FY2027 guide (mid) | |
|---|---|---|
| Revenue | $21.448B | $23.396B (+9.1%) |
| Operating income, as reported | $5.884B | $7.449B |
| GAAP operating margin | 27.43% | 31.84% → +441bp ✓ (matches the Investor Day's "~440bp" exactly) |
| Gross profit at 80.2% GM | $17.195B | $18.756B |
| Implied total operating expense | $11.311B | $11.307B — flat to the dollar |
The entire 440bp is zero operating-expense growth on +9% revenue. Management named the drivers and they are internally consistent: "margin expansion driven by savings from workforce changes announced last quarter and continued efficiency gains… partially offset by deliberate choices… to increase investments in sales and marketing as we focus on growing our new-to-the-franchise customers." It is a reallocation, not a cost cut: back-office savings fund the acquisition push.
Three components, in descending order of certainty:
- $293M of FY26 restructuring does not recur = 137bp, or 31% of the whole expansion. Mechanical. (FY25 restructuring was $15M.)
- The May 2026 workforce action lands in FY27, not FY26. FY26 year-end headcount rose 18,200 → 18,600 — because the ~3,000 affected US employees remained on payroll through July 31, 2026, the fiscal-year-end measurement date. (This reconciliation is inference — the 10-K does not state it — but it is consistent with the press reporting and it materially increases the credibility of the flat-opex guide rather than undermining it.)
- AI cost-to-serve. >$135M of AI-powered customer-service efficiencies; CTO Alex Balazs: 70% of code pull requests AI-delivered, 40% coding-velocity improvement. Real but unaudited.
⚠️ Note the asymmetry the headline hides: non-GAAP margin expansion is only ~260bp against the 440bp GAAP figure. The 180bp wedge is restructuring roll-off plus the SBC definition change — i.e. a meaningful share of the impressive-sounding GAAP number is non-operating optics. And the restructuring-savings component is one-time in character, so it is not a durable offset to a structurally lower entry price point.
3. The buried finding — Intuit made its non-GAAP stricter, and the August sell-off priced a definitional artifact
CFO Sandeep Aujla, Q4 FY26 call, in the guidance preamble:
"Share-based compensation expenses will no longer be excluded from our non-GAAP financial measures [effective August 1, 2026]. We view share-based compensation as a recurring component of our compensation program and believe including this expense… reflects our core operating results."
And sized explicitly: "non-GAAP diluted EPS is expected to be $22.88 to $23.12 or growth of 23% to 24%. This includes a $5.81 impact from share-based compensation expenses."
What that does to the "$23.00 vs ~$27.30 consensus miss"
| Basis | FY26 | FY27 guide (mid) | Growth |
|---|---|---|---|
| Non-GAAP as guided (new, SBC-included) | $18.62 (restated) | $23.00 | +23.5% ✓ ties to the company's "+23–24%" |
| Non-GAAP on the OLD definition (add back $5.81) | $24.27 (as filed) | $28.81 | +18.7% |
| Street consensus on 2026-08-25 (~$27.30) — built on the old definition | — | — | Guide was ~5.5% ABOVE it |
On a like-for-like basis the FY27 EPS guide beat consensus. The 12% drop on 2026-08-26 priced a 16% "miss" that was substantially a change in the definition of the metric. This is the mirror image of [[pattern-non-gaap-definition-change-leaves-a-restatement-fingerprint]] — that note covers a company widening non-GAAP to flatter itself; INTU narrowed it, and the market punished the honesty.
Three corroborations: 1. The company states the per-share impact itself ($5.81 FY27; $1.48 in the Q1 guide). 2. The stated +23–24% only reconciles against a restated FY26 base of $18.62, not the as-filed $24.27: $23.00 ÷ 1.235 = $18.62 ✓. 3. FY26 SBC from the 10-K cash-flow statement, $2.056B, after tax ÷ diluted shares = $5.63/sh — within 3% of the guided $5.81 on a FY27 share count.
⚠️ But the Street has since re-based — so this is an explanation, not a live edge
Consensus FY27 EPS has been revised down 13.7% to ~$23.49, which sits above the guide midpoint of $23.00. The Street corrected the definition within three weeks. Consensus PT fell $456 (07-31) → $406 (09-18), −11%.
What survives, therefore, is not "the market is wrong about earnings." It is narrower and still worth holding: the August 12% gap-down was priced against a number that no longer existed; the earnings estimate was repaired but the multiple never recovered. What remains in the price today is a pure multiple de-rate on (a) a lowered growth algorithm and (b) a macro rate shock — not an earnings cut. That distinction is the whole investment case, and it is weaker than a live mispricing would be. Weighted accordingly in §9 (+0.3, not +0.5).
One durable benefit does survive intact. The baseline retracted "Intuit's non-GAAP flatters less than half as much as Adobe's" because the FY26 GAAP→non-GAAP gap was +47.4%. On the FY27 definition the gap is GAAP $20.24 → non-GAAP $23.00 = +13.6% — now genuinely among the cleanest in large-cap software. The company fixed exactly the thing the last report caught it on, and that is a permanent improvement in earnings quality, not a quarter's optics.
4. The delta ledger
28 claims tracked. 0 retracted · 5 superseded · 10 refreshed · 10 carried · 3 drifted · 0 untested-twice.
🔄 SUPERSEDED — facts changed (lead, per protocol)
| # | Baseline claim | Replaced by | Cause · strength |
|---|---|---|---|
| St5 | "SBC not retrieved — second consecutive pass untested." | FY26 SBC = $2.056B = 9.6% of revenue (FY25 $1.968B; FY24 $1.940B). It is 95% of the entire GAAP→non-GAAP bridge — see the reconciliation below. | FY26 10-K, primary-source, corroborated by the guided $5.81/sh. Three-pass question closed. ⚠️ This FAILS the baseline's own upgrade condition, which asked for the gap to be amortization-led. It is overwhelmingly SBC-led. |
| St5-b | (implicit) the non-GAAP basis is stable | Non-GAAP definition changed 2026-08-01: SBC now included. Commitment: SBC to 9% of revenue by FY28, 8% by FY30; non-GAAP EPS growth ≥ high-teens on the new stricter basis. | CFO prepared remarks; reaffirmed at Investor Day. Favourable — §3. |
| T1 | "Revenue growth +14%, held exactly." | FY27 guided +9–10%. And the 3-yr algorithm was cut: GBS from 15–20% → 10–15% CAGR; Consumer 4–8%. Blended ≈ +10% — "durable double-digit" is now the floor. | Primary-source. Unfavourable, and larger than first apparent — a 500bp cut to the GBS algorithm, not a one-year air pocket. Management's framing: "We are creating the pressure. We are not being pressured." |
| P-rows | FV $385–475 · entry $305–350 · trim 28x ttm | FV $350–440 · entry $280–320 · trim 29x ttm | Re-derived from scratch — §7. |
| J1 | ACCUMULATE [7.5] | ACCUMULATE [7.0] | §9. |
The FY26 GAAP → non-GAAP reconciliation, in full (press release 2026-08-25; $M and $/sh):
| Item | $M | $/sh |
|---|---|---|
| GAAP net income / diluted EPS | 4,566 | 16.46 |
| Share-based compensation | +2,056 | +7.42 |
| Amortization of acquired technology | +174 | +0.63 |
| Amortization of other acquired intangibles | +485 | +1.75 |
| Restructuring charges | +293 | +1.06 |
| Other net items | −167 | −0.61 |
| Income tax effects | −675 | −2.44 |
| Non-GAAP net income / diluted EPS | 6,732 | 24.27 |
SBC is $2,056M of $2,166M net add-backs — 95%. Total intangible amortization is $659M ($2.38/sh). The baseline hoped for amortization-led and got SBC-led. That is the correct reading and it is unfavourable — though it is prospectively cured by the definition change (§3).
📉 DRIFTED — moving against the thesis, not yet breaking it
| # | Claim | The drift | Where it breaks |
|---|---|---|---|
| T4 | "FY26 FCF $8.617B, +41.7%; FCF margin 40.2%." The single largest driver of the 7.0 → 7.5 upgrade. | The fact holds; the trend claim does not. The 10-K explains the gap the baseline flagged as unexplained, and it is not working capital and not deferred revenue: deferred income taxes swung +$1,714M (FY26 +$1,279M vs FY25 −$435M) = ~65% of the $2,631M OCF increase. Working capital was flat (−$4M vs −$206M); deferred revenue moved only +$53M. Cause, now verified: the One Big Beautiful Bill Act (enacted 2025-07-04) reinstated immediate §174 domestic R&D expensing with an election to deduct prior-year capitalised R&D. Long-term deferred tax assets fell $1,222M → $172M; the net position went from +$1.2B DTA to −$67M DTL. A cash-tax timing catch-up, non-recurring in magnitude. Clean FCF ≈ $6.9B, +13% — in line with net income. FCF margin ≈ 32%, not 40.2%. | FY27 GAAP tax rate guided ~27% (vs 24.1% FY26). If FY27 OCF fails to exceed ~$8.0B, the 40% FCF margin is confirmed a one-year artifact. |
| S4 | "Switching costs ~zero at simple, stable at complex; the sub-$50k DIY filer is already lost and structurally unrecoverable — but the damage is confined to that segment." | The containment failed; the damage propagates upward. "More than 3/4 of TurboTax Live customer additions in FY26 came from DIY upgrades into Live" — and FY27 TurboTax Live revenue growth is guided to mid-teens, down from +37%, explicitly "due to fewer DIY upgrades." A broken DIY funnel starves the premium assisted business the baseline treated as insulated. The most important new negative in the file. | TurboTax Live revenue growth below ~10%, or new-to-franchise Live customer growth (+15% in FY26) going flat. |
| St4 | "Debt/assets 18.0%, positive tangible book." | Total debt $5,973M → $7,669M ($1,249M ST + $6,420M LT). Debt/assets 18.0% → ~20.9% (Yahoo's 22.9% uses a higher $8.42B figure incl. leases). $1.75B senior notes issued 2026-06-11 ($750M 4.95% due 2031, $1,000M 5.50% due 2036) — stated use: refinancing the $750M 5.25% notes due Sept 2026 and $500M 1.35% due July 2027, i.e. a refi, not buyback funding. Equity fell $19.71B → $18.99B because $6.76B was returned against $4.57B earned (148% of net income). | Mild, and not a risk. $970M of the "debt" is a warehouse revolver against the consumer/SMB lending book — asset-backed, not corporate leverage. Cash + investments $7.2B ⇒ net debt ≈ $0.5B on a $78B cap; interest coverage $5.88B ÷ $256M = 23×. Breaks only if net debt passes ~1.0× EBITDA (~$6.7B). |
🔁 REFRESHED — same claim, new number
A portfolio-specific passage was removed from the public build.
✅ CARRIED — re-tested, still true
| # | Claim | The check |
|---|---|---|
| S1 | Tax compliance is mandated and annually recurring; Intuit's risk is price, not existence | Confirmed from outside Intuit: the IRS requirement that a certified professional sign the return and defend the audit was reaffirmed across the 2026 season. Management's response remains a price response. |
| S2 | Regulatory intangible — e-file authorization, 41 state engines, EFIN/ETIN, annual re-certification; AI reduces none of it | No regulatory change. Untouched. |
| S5 | Brand no longer supports a price premium | Strengthened to an explicit, repeated management admission. Q4 call: "Price is now the #1 reason customers leave TurboTax." Investor Day: "where we lost share was in DIY… The number one reason why customers left us was price." FY27 guidance carries "lower tax ARPC." The baseline inferred this a year early. |
| S7 | GBS is the largest segment and the best-defended | GBS $12.9B = 60% of revenue, +16% FY26, guided $13.1–13.2B, +13–14% FY27. Both halves hold. QuickBooks Free does not contradict it — §2. |
| T5 | Share count falling on the doubled buyback | Shares out 279.1M → 268.4M (−3.8%); diluted 283M → 277M (−2%). $5,412M repurchased (+95%). ⚠️ Efficiency note: $5.4B of buyback moved the diluted count only 2%, because $2.06B of SBC offsets it. |
| T6 | Gross margin stable ~79.6%; no moat-erosion signature | 80.2%, rising for a fourth straight year (78.1 → 78.7 → 79.6 → 80.2), against a 77% break trigger. The strongest single counter-evidence to the price-war narrative. |
| T7 | 16 straight dividend raises, ~15% | +15% to $1.38/qtr, payable 2026-10-16. Reaffirmed at Investor Day. |
| St1 | (the baseline's FCF retraction) FY26 FCF was $8.617B, not a decline | Retraction stands. 10-K confirms OCF $8,838M − capex. (Magnitude now qualified — see T4.) |
| J2 | Moat 7.0/10, EVERGREEN | Held. ROIC 15.3% → 16.8%; gross margin rising; regulatory intangible untouched; accountant network re-evidenced; no named AI-native consumer filer materialised — §5. |
| P-DYT | DYT void — the yield rose because the price fell | Still void, and more so. Yield 1.87% vs 5yr avg 0.68% = 2.7×, on a price −58% from the high. [[pitfall-dyt-inverts-when-price-caused-the-yield]] — applied, reported, discarded. |
5. Break-trigger status — the baseline's own falsification list
The baseline pre-registered seven break triggers. Tested individually against primary sources: zero fired. That is the backbone of the verdict and it should be read literally.
| # | Break trigger (as written 2026-08-27) | Status | Evidence |
|---|---|---|---|
| 1 | FY27 revenue guide < 8% | 🟡 NOT FIRED — thin | Guided +9–10%; reaffirmed 09-17. Only 100bp of cushion, and the new 3-yr algorithm blends to ~10%. |
| 2 | Dividend raise < 8% | ✅ NOT FIRED | +15% to $1.38/qtr, 16th consecutive; reaffirmed 09-17. |
| 3 | QBO Accounting < 15% for 2 quarters | ✅ NOT FIRED — comfortably | Q4 +20%; FY26 +23%. FY27 guided to keep growing on higher effective prices. |
| 4 | TurboTax units decline again in the 2027 season | ⏳ NOT YET TESTABLE — season runs Jan–Apr 2027 | The guide implies the opposite. TurboTax revenue +2–3% on explicitly lower ARPC and a stated assumption that total IRS filers are flat ⇒ units must grow to clear the guide. Management has staked the guide on a unit recovery. First read: April–May 2027. |
| 5 | A named AI-native filer discloses > 2M returns | ✅ NOT FIRED | No consumer AI-native filer disclosed anything of the sort for the 2026 season. The funded AI-native entrants (Accrual, $75M, Feb 2026; Basis AI, $100M at $1.15B, Feb 2026) target CPA / partnership-return workflows, not consumer DIY. The damage Intuit took came from conventional low-cost filers, not from AI — management says "low-cost providers," never AI, and Goodarzi said of the May layoffs "none of it had to do with AI." This materially weakens the Goldman $0.12-per-return bear case as an explanation of FY26/27. |
| 6 | Gross margin < 77% | ✅ NOT FIRED — moving the other way | 80.2%, fourth consecutive year of expansion. |
| 7 | GBS growth below ~12% | 🟡 NOT FIRED — but the floor moved under it | FY27 guided +13–14%. ⚠️ The new 3-yr GBS algorithm is 10–15% CAGR — its lower bound now sits BELOW the break trigger. Management has published a range that contains the failure condition. |
| — | Operating margin failing to expand | ✅ NOT FIRED | +441bp GAAP guided (~260bp non-GAAP). |
| — | Online paying customers flat or negative | 🟡 NOT FIRED — borderline | +3%, decelerating ~2pts. Management calls it out as unacceptable. |
| — | FY27 non-GAAP EPS guided below +15% | ✅ NOT FIRED | +23–24%, on a stricter basis (§3). |
⚠️ A vendor trap caught live this pass — it was about to manufacture a credibility problem
fin.py / Yahoo report OpInc $6.177B for FY26. The filed "Total Operating Income As
Reported" is $5.884B. Yahoo's field excludes the $293M restructuring charge and is therefore the
pre-special-items figure, overstating operating margin by 137bp (28.80% vs 27.43%) and ROIC by
~80bp.
This mattered. Used naively, the Investor Day's ~440bp margin-expansion claim would have reconciled to only ~300bp, and the ~140bp shortfall would have read as management hype. On the correct as-reported base it reconciles to +441bp, exactly. A vendor field nearly created a management-credibility problem that does not exist. Related: [[pitfall-earnout-revaluation-inflates-operating-income]] — same class, different line item.
(The OpMargin 18% < NetMargin 21% tripwire the baseline flagged is still firing in Yahoo's
snapshot — [[pattern-net-margin-above-operating-margin-is-a-tripwire]]. No claim was built on it.)
6. How the close calls were decided
The two dominant forces are large and opposite, and they do not cancel to zero
| Force | Direction | Size |
|---|---|---|
| The non-GAAP definition change (§3) | Favourable | Explains the August 12% gap-down as a definitional artifact; permanently repairs the earnings-quality gap the baseline retracted a claim over. But the Street re-based within three weeks — this is an explanation, not a live edge. |
| The deferred-tax explanation of the FCF surge (T4) | Unfavourable | Removes ~65% of the $2.63B OCF increase that was most of the 7.0 → 7.5 upgrade. Clean FCF +13%, not +41.7%. Now traced to a named statute — certain, not suspected. |
They differ in kind, and that decides the weighting. The FCF finding is a correction to our own prior reasoning, verified to a specific cause — it permanently lowers the quality of the evidence that justified the last upgrade. The definition change was an edge on 2026-08-26 and is no longer one at 2026-09-22, because consensus FY27 EPS has already moved to ~$23.49. A certain correction to our own analysis outranks an edge that has already closed.
Net: conviction steps back 0.5 to [7.0] — precisely the level the 2026-08-04 report held before the FCF number moved it. That symmetry is deliberate: the upgrade rested on a number now shown to be two-thirds timing, so the upgrade is returned.
Was the Investor Day a thesis break? No — but it was not neutral
Against the baseline (real): the admission of new-customer misses in both segments — QBO, not only tax, the first time the acquisition problem touched the defended side of the house. The 3-year algorithm cut (GBS 15–20% → 10–15%). And a quiet governance point: long-term guidance is now given at segment level only, an explicit reduction in disclosure granularity, which arrives in the same event as a reclassification that moves the worst asset out of the best segment.
For the baseline (also real, and larger in dollars): guidance reaffirmed three weeks after a −12% day — management did not blink; dividend +15% confirmed; the margin guide reconciles to the dollar; and QuickBooks list prices went UP 22–24% at the top in the same month the free tier launched.
The decisive test was whether the State evidence accumulates into a Structural break. It does not, on the one measure that must move first: gross margin, expanding for a fourth year. S7 holds; S4 drifts.
Why S4 drifts rather than carries
The baseline's S4 held that the DIY loss was contained — "already lost, structurally unrecoverable at a premium price," rated 3.0/10 and written off. That containment failed. Over three-quarters of TurboTax Live's growth came from upgrading DIY customers, so a broken DIY funnel mechanically slows Live — guided from +37% to mid-teens. The baseline's defence of the top end depended on a feeder it never identified as a dependency. This is the pass's only genuine deterioration of a Structural claim, and it is why the verdict did not simply carry.
Why the price fall is not itself evidence
−15% since the baseline, of which the 09-18 sector move on a Fed hike and a 5%+ 10-year is a visible share. A rate-driven de-rate of every long-duration software name is a Price force and belongs in the discount rate, not in the moat rating. It justifies a lower terminal multiple — applied in §7 — but it is not a finding about Intuit.
The analyst tape, read for information rather than direction
JPMorgan's OW → Neutral with a $605 → $331 PT (−45%) is the standout capitulation. But the shape of the tape is more informative than any single call: the only two PT raises in the window came from the two bears — Goldman (Sell, $276 → $304) and Piper (Underweight, $250 → $290). That is bears marking to market, not bulls capitulating. Consensus mean PT ~$406 against a ~$296 spot, on 31–34 analysts. Recorded as a weak contrarian positive; not weighted in the verdict.
7. Valuation — re-derived from scratch
⚠️ Yahoo's PE(fwd) 10.76 is again a phantom. It implies forward EPS of $27.08, matching no
guided year — FY27 non-GAAP is $22.88–23.12, FY27 GAAP is $20.12–20.36. Plausibly an old-definition
blend. Not used. PE(ttm) 17.70 × EPS(ttm) $16.46 = $291.4 ✓ reconciles to spot and to the
company's own FY26 GAAP EPS — the ttm field is verified good and the fwd field is not.
[[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]].
Multiples at ~$295.83
| Basis | EPS / CFPS | Multiple |
|---|---|---|
| FY27 non-GAAP as guided (SBC-INCLUDED — the strictest number Intuit has ever published) | $23.00 | 12.9× |
| FY27 consensus (re-based) | $23.49 | 12.6× |
| FY27 non-GAAP, old definition (+$5.81) | $28.81 | 10.3× |
| FY27 GAAP guide (mid) | $20.24 | 14.6× |
| TTM GAAP (= FY26 actual) | $16.46 | 18.0× |
| FCF/share, as reported | $31.11 | 9.5× |
| FCF/share, clean of the deferred-tax swing | $24.92 | 11.9× |
| "Owner FCF" (clean FCF less SBC — the honest per-share figure) | ~$17.50 | ~16.9× |
The owner-FCF line is what keeps this from being a screaming bargain. SBC is a real economic cost — the proof is that $5.4B of buyback moved the diluted count only 2%. Charging SBC against FCF, the name is ~17× owner earnings, not 9.5× headline FCF. This is why the FCF model is weighted down again this pass, as the baseline did for a different (and now-resolved) reason.
| Model | Weight | Range | Note |
|---|---|---|---|
| Forward earnings — 15–19× FY27 non-GAAP $23.00 | 50% | $345–437 | Multiple cut from the baseline's 16–20× for three named reasons: the GBS 3-yr algorithm was cut 15–20% → 10–15%; TurboTax Live and Credit Karma both decelerate; and the risk-free rate rose (10-yr > 5%, Fed hiking). |
| Owner-FCF — 20–24× on ~$17.50/sh | 20% | $350–420 | Corroborates the earnings range from a cash basis. |
| Bogle — 1.87% yield + high-teens non-GAAP EPS growth − multiple compression | 20% | ~13–16%/yr expected return | Supports the range; does not set it. |
| Graham IV $161.89 | 10% (0.5 weight) | Drags | BVPS $70.76 is ~$52/sh goodwill. Reality check only. |
| DYT · DDM | 0% | Void — reported and discarded | Price caused the yield. |
Fair value: $350 – $440 (midpoint ~$395)
Down ~8% from the baseline's $385–475 (mid $430). The cut is entirely a multiple decision, not an earnings decision — FY27 EPS is guided higher than the baseline modelled, on a stricter basis. The multiple comes down because the growth algorithm came down and the discount rate went up. (Sanity check: sell-side consensus mean PT is ~$406, inside this range and near its upper half.)
| Baseline (08-27) | Now (09-22) | |
|---|---|---|
| Price | $347.22 | ~$295.83 |
| Fair value | $385–475 | $350–440 |
| Entry | $305–350 | $280–320 |
| Strong buy | <$285 | <$262 |
| Trim | 28x ttm ≈ $461 | 29x ttm ≈ $477 |
| Conviction | [7.5] | [7.0] |
On the trim, and why the multiple went UP while fair value went DOWN. These are not in conflict.
The baseline set 28x ttm ≈ $461 ≈ 20× FY27 non-GAAP believing $23.00 excluded SBC. It includes it.
The same economic trim level therefore lands on a higher multiple of the now-stricter EPS:
29 × $16.46 = $477 ≈ 20.7× FY27 non-GAAP — the baseline's judgment, restated onto a denominator we
now understand. The basis stays ttm because Yahoo's forwardPE is broken (a 20x fwd trim would
render ~$550, ~15% too high) and its trailingPE is verified.
[[pitfall-multiple-trim-inherits-the-broken-vendor-field]]. Re-check the basis after the Q1 FY27
print, when ttm EPS steps up and the rendered dollar rises on its own — the intended behaviour.
Zone note, stated plainly: spot ~$295.83 sits inside the re-derived $280–320 entry band, at roughly the 40th percentile, with ~$16 of room beneath it. Unlike the baseline — which sat 0.8% inside its band and said so — this is a real in-zone reading. The entry band moved down with fair value, which is the correct direction and the discipline [[pitfall-stale-entry-zone-suppresses-a-name]] exists to enforce in both directions.
8. What is genuinely new
| Finding | Detail | Direction |
|---|---|---|
| Non-GAAP now includes SBC (from 2026-08-01) | $5.81/sh FY27 impact. The August "EPS miss" was a definition change. §3. | 🟢 Large — but the Street re-based |
| SBC quantified: $2.056B = 9.6% of revenue | 95% of the entire GAAP→non-GAAP bridge. Committed to 9% by FY28, 8% by FY30. | 🟡 Level high, trajectory committed |
| FCF surge is a deferred-tax swing, cause identified | +$1,714M YoY swing = 65% of the OCF increase. OBBBA §174 R&D expensing election; LT DTA $1,222M → $172M. Clean FCF +13%. | 🔴 |
| QuickBooks list prices RAISED 22–24% on 2026-08-01 | Plus $115→$140, Advanced $275→$340; Free / Lite / Ledger / Simple Start unchanged. The barbell. | 🟢 The single best counter to the price-war thesis |
| QuickBooks Free is a metered $0 tier monetised by payments | 2 invoices/mo, 1 bank, 3 reports, single user, no accountant access, no AI. Unlimited invoices if you enable QuickBooks Payments. >20,000 customers in month one. | 🟢 Small but real |
| Credit Karma Tax free filing quantified | $0 federal / $15 state. TY2025 pilot: ~80% customer incrementality; 34% of Gen Z CK Tax filers were new to Credit Karma. | 🟢 |
| 3-yr algorithm cut: GBS 15–20% → 10–15% | Consumer 4–8%. Blended ≈ 10%. Goodarzi: "We are creating the pressure. We are not being pressured." | 🔴 Bigger cut than first apparent |
| TurboTax Live decelerating 37% → mid-teens | Because >3/4 of its adds were DIY upgrades | 🔴 The key new negative |
| FY26 restructuring $293M | = 137bp, or 31% of the guided FY27 margin expansion, and mechanical. FY25 was $15M. | 🟡 |
| The May 2026 cut lands in FY27, not FY26 | FY26 headcount rose 18,200 → 18,600 because the ~3,000 affected US staff stayed on payroll through 07-31. (Inference — the 10-K does not say so.) | 🟢 Raises the credibility of the flat-opex guide |
| AI cost-to-serve quantified | >$135M customer-service efficiencies; 70% of code PRs AI-delivered; 40% coding velocity | 🟢 Unaudited |
| Non-GAAP margin expansion is only ~260bp vs 440bp GAAP | 180bp wedge is restructuring roll-off + the SBC change — non-operating optics | 🟡 |
| Mailchimp becomes its own reportable segment FY27 | $1.3B, flat to −1%. Removes the drag from GBS optics. Long-term guidance now segment-level only. | 🟡 Disclosure granularity reduced |
| Credit Karma is bigger than rated | ~1 in 9 US credit-card and personal-loan originations; +3/+4pts share in 2 years | 🟢 |
| TurboTax distribution into LLMs | "expanding the distribution… including through leading LLM experiences"; Intuit signed deals with Anthropic and OpenAI | 🟢 Partnering with the disruptor |
| TWO securities class actions, not one | Baldwin (Rosen/Pomerantz), class period 2025-08-22 → 2026-05-20, alleging overstated TurboTax competitive position. Grant & Eisenhofer (filed 08-17), class period 2025-02-25 → 2026-06-01, alleging concealed genAI competitive pressure and a misrepresented Mailchimp turnaround. Lead-plaintiff deadline 09-08 passed. | 🟡 Immaterial to cash flow at this size; monitor |
| Insider picture resolved | Prabhu bought ~$542K on 2026-05-26 — the only open-market buy in a year. Cook sold ~$375M Dec 2025 at $630–680. No buying since 08-01. | 🔴 Mild |
| Macro: Fed hiking, 10-yr > 5% | Sector-wide software de-rate on 09-18 | 🔴 Not company-specific; priced into the multiple |
9. Thesis persistence and conviction delta
Thesis persistence: 12 of 15 Structural + Trend claims survived as CARRIED or REFRESHED = 80% (down from 93%).
- Structural: 6/7 held — only S4 drifted, and on a dependency (DIY → Live) the baseline had not identified.
- Trend: 6/8 — T1 superseded by a published algorithm cut; T4 drifted on the deferred-tax finding.
80% persistence against a −15% price move is still a discount-widening setup, but a weaker one than last pass — and the weakening is honest rather than cosmetic.
Conviction: 7.5 → [7.0] ⬇️
| Row | Move | Weight |
|---|---|---|
| T4 DRIFTED | The FCF surge that supplied most of the 7.0 → 7.5 upgrade is ~65% deferred-tax timing, now traced to OBBBA §174. Clean FCF +13%, not +41.7%. | −0.5 — the whole downgrade |
| S4 DRIFTED | DIY funnel damage propagates into TurboTax Live (+37% → mid-teens). New deterioration of a Structural claim. | −0.3 |
| T1 SUPERSEDED | GBS 3-yr algorithm cut 15–20% → 10–15%; its floor is below our break trigger. | −0.3 |
| St5 SUPERSEDED | The bridge is 95% SBC — the baseline's stated upgrade condition (amortization-led) failed. | −0.2 |
| §2 — the Aug 1 price increase | Plus +21.7%, Advanced +23.6%, entry SKUs held. The commissioned question resolves for the baseline: funnel, not margin event. | +0.4 |
| §3 — non-GAAP definition change | Explains the August gap-down; permanently repairs earnings quality. Discounted because consensus re-based within three weeks. | +0.3 |
| §5 — zero break triggers fired | Seven pre-registered falsification tests, all held, all primary-sourced. | +0.2 |
| St3 / T6 / T7 | ROIC 15.3% → 16.8% on a filed balance sheet; gross margin 80.2% rising; dividend +15% reaffirmed. | +0.1 |
| §2 — headcount timing | The May cut lands in FY27; the flat-opex guide is more credible than it looked. | +0.1 |
Net −0.5. Returned to the level held on 2026-08-04, for the honest reason that the number which moved it is now known to be two-thirds timing.
10. Verdict — ACCUMULATE [7.0]
Is ~$295.83 an add, a hold, or a downgrade? An add — measured, not aggressive.
The case without hedging: this is a 12.9× forward-earnings business — on the strictest non-GAAP definition it has ever published — guiding +23–24% EPS growth, expanding gross margin for a fourth consecutive year, earning 16.8% ROIC on a filed balance sheet, carrying ~$0.5B of net debt against $7.2B of cash and investments, covering a 16th consecutive +15% dividend raise six times over, and trading 58% below its high. Seven pre-registered break triggers were tested against primary sources and not one fired. The commissioned question — whether a free QuickBooks tier breaks the QuickBooks moat — resolves against the bear case, and the decisive proof is that Intuit raised Plus and Advanced list prices 22–24% in the same month it launched the free tier. The price concession is real, dated and management-executed, but it is in consumer DIY tax, not QuickBooks.
Against that, three things genuinely got worse and none are noise: the FY26 cash-flow story is two-thirds a §174 tax-timing catch-up; the DIY funnel damage has reached TurboTax Live, which the baseline treated as insulated; and management published a 3-year GBS algorithm (10–15%) whose lower bound sits below our own break trigger. That is why fair value comes down 8%, conviction comes down half a point, and the position gets added to rather than built into.
Sizing note: spot is at the 40th percentile of the entry band with ~$16 of room to the floor, and the next binary information event is ~60 days away. Add in a tranche; keep capacity for <$262, which the strong-buy level now names.
Break triggers (re-set — these replace the baseline's list)
- Q1 FY27 operating expense growing more than ~2% YoY. The 440bp margin guide requires flat opex dollars on +9% revenue while S&M rises. This is the first and best audit of the whole FY27 story, and it is 60 days away.
- FY27 revenue guidance cut at any quarterly update, or GBS guided below 12%.
- TurboTax Live revenue growth below ~10%, or new-to-franchise Live growth flat — S4's drift completing into a break.
- The 2027 tax season delivering a second consecutive year of TurboTax unit decline despite the lower ARPC. The guide mathematically requires unit growth; if units fall anyway, the price lever has failed and the Consumer thesis breaks outright.
- Online paying customers flat or negative (8.9M, +3%) — the QuickBooks Free funnel failing at the top. Or: evidence that the free/Lite tiers are cannibalising paid SKUs rather than adding new logos — watch ARPC and paying-customer count together, never separately.
- Gross margin below 78% — tightened from 77%, because the series is at 80.2% and rising; the first decline is the signal, not an arbitrary floor.
- FY27 OCF below ~$8.0B — confirming the 40% FCF margin was a one-year deferred-tax artifact.
- Dividend raise under 8% at the FY27 print (carried unchanged).
Upgrade to [7.5]
Q1 FY27 opex flat or down with S&M up · online paying customer growth re-accelerating above 5% · QuickBooks Free disclosing a paid-conversion rate or new-logo mix · FY27 OCF tracking above $8B ex-deferred-tax · TurboTax Live holding mid-teens rather than decelerating further.
Downgrade to [6.5]
Any FY27 guidance cut · GBS below 12% · TurboTax Live below 10% · a second year of TurboTax unit decline on lower prices · Q1 FY27 opex growing mid-single-digits, which would break the margin guide and the credibility of the reallocation story together.
11. What this pass did NOT test
- The Investor Day slide deck was not read primary-source. The official PDF exceeded the fetch size limit and Seeking Alpha returned 403. Every Investor Day figure here is either (a) identical to a figure in the 2026-08-25 transcript, which was read verbatim, or (b) second-hand via MarketBeat / Truist / StockOpine. They are mutually consistent, but the deck itself is UNTESTED — including any QuickBooks Free unit economics or FY30 targets disclosed only there. Pull it at the next touch.
- No explicit FY30 revenue or operating-margin target was found. Only the FY30 SBC-at-8% target is confirmed. If the deck contains one, we do not have it.
- The QuickBooks Free cannibalisation question is open, and management concedes it is open — whether cheap tiers attract new customers or accelerate existing ones migrating down. No cohort data exists yet. This is the most important unresolved item in the file.
- No competitor was ever named, and no competitor unit/ARPU data was obtained. The Wave / Xero / Zoho and FreeTaxUSA / Cash App Taxes attribution in §2 is ours and unverified.
- QuickBooks Lite list price could not be found.
- The class-action dockets were not read. Consolidation, lead plaintiff, named individual defendants, and whether the genAI/Mailchimp complaint is a separate docket from Baldwin are all unverified.
- 10-K Note 15 (restructuring) detail — the $293M is the press-release add-back total only; the severance / real-estate / other split is unknown.
- The headcount reconciliation (why 18,600 is higher after a 17% cut) is our inference, consistent with press reporting but not stated in the 10-K. It carries weight in the flat-opex argument, so it is flagged as thin.
- TurboTax price vs mix — third consecutive pass unresolved. Intuit still discloses no ASP series ([[pattern-disclosed-asp-series-separates-price-from-volume]]). FY27 will be the first year the direction is unambiguous, because management has guided ARPC down.
- FY27 TurboTax unit count — the decisive Consumer test, not knowable until April–May 2027. Flagged so the next two passes do not pretend otherwise.
- A conflicting third-party figure is on the record and unresolved: MarketBeat's Investor Day write-up reports "nearly 8 million online paid customers, up 4%" against the company's own 8.9M, +3%. The transcript figure is treated as authoritative here.
Sources
Primary
- Intuit FY2026 10-K, filed 2026-09-09 · cash flow statement · balance sheet · debt note
- Intuit Q4/FY2026 earnings call transcript, 2026-08-25 — via roic.ai get_latest_earnings_call (all quoted management statements)
- Q4/FY26 earnings press release (8-K) (the GAAP→non-GAAP reconciliation)
- Intuit Hosts Investor Day, Reaffirms Q1 and FY2027 Guidance, 2026-09-17
- June 2026 $1.75B senior notes 424B5 · May 2026 restructuring 8-K
- QuickBooks Aug 2026 price changes · Is QuickBooks free? · Free vs paid plans · Free Credit Karma Tax
- roic.ai get_cash_flow / list_employee_counts INTU; yahoo-finance MCP income_stmt, balance_sheet, get_historical_stock_prices; python .mcp/fin.py INTU --news --holders — all retrieved 2026-09-22
Secondary - Intuit Investor Day Puts AI at Center of Growth Reset After Customer Misses — MarketBeat/Yahoo - Intuit Investor Day: rebuilding customer growth without giving up earnings growth — StockOpine - Truist reiterates Hold, $300 target after Investor Day — Investing.com - CrowdStrike, HubSpot, Intuit, Cloudflare, Okta Shares Plummet — StockStory/TradingView (the 09-18 sector move and its Fed-hike cause) - Intuit CEO says the 17% workforce cut had nothing to do with AI — CNBC - Pomerantz class-action announcement, 2026-09-01 · Grant & Eisenhofer filing, 2026-08-17 - When AI Files Your Taxes: Who Pays When It Fails (2026-season AI-filer landscape; Accrual and Basis AI funding; IRS human-signature requirement)
Chain: Baseline analyze-2026-08-27.md · original analyze-2026-08-04.md