VEEV

We rate Veeva Hold. Our value is $296 a share against $318.34 this morning and $323.04 at Friday's close. The shared rule calls a stock a Buy at or below our value and an Avoid when our value is below 85% of the price (above $348.55 here); the price sits between.

This report replaces a Buy no one researched. Our 2026-10-15 report said Avoid at $202. Changes to the shared valuation method since then restated it to $322 and Buy, with no new Veeva work. Two audits found that Buy unsound. We agree on one point and only one: the restated exit multiple, 48.7x operating profit, averages Veeva's own ten-year median, earned while it grew 25% a year, with its peers'. Our forecast has Veeva growing 9% in 2032, when it is valued. Software companies that slowed like that fetched a median 0.86 times this kind of anchor five years later. That gives 41.9x and $296.

Terms.

• Bear, central and bull cases: the central case is our best estimate; the bear is a severe but plausible bad outcome, the bull the matching good one. Each is one consistent business.
• Long fade: values the company in January 2032 by letting its profit growth fall in a straight line from 8.5% to 3% over 15 years, and its return on new investment from very high to 40%, discounted at 9.57%.
• Outside view: values the same 2032 profit at a multiple other companies traded at. The method's default, anchor (d), is the average of Veeva's own median enterprise value to operating profit (EV/EBIT) over its last ten fiscal years and its peers' median of their own.
• Operating profit (EBIT) is GAAP, after stock pay. NOPAT is operating profit after tax.
• Fiscal years end January 31: FY32 is the year to 2032-01-31.

1. What is new since 2026-10-15: no filing, one event ahead

• No 8-K has been filed since the Q2 FY27 10-Q of 2026-08-27 (scripts/research-events.py, 2026-10-26). Q3 FY27 results are not yet dated by the company; last year's came on 2025-11-20.
• Veeva's news page adds nothing after the IQVIA clinical-operations release of 2026-10-13, which the earlier report weighed. Vault CRM stands at 14 of the top 20 biopharmas committed, six to Salesforce (press release 2026-09-23).
• Investor Day, 2026-11-05. What could change: the $6B 2030 revenue run-rate goal (our central reaches $5.7B leaving FY31; reaching the goal is worth about $20 a share), the CRM share target and the margin path. A new goal is a target; we would move the forecast only on a plan that backs it with dated evidence, such as a higher FY28 guide.
• The price has risen 1.7% since the 2026-10-15 report ($313.11 then) and 108% from its 2026-06-22 close of $153.16 (market data closes, data/daily-closes.json).

The operating forecast is therefore the 2026-10-15 report's, imported unchanged into model.py; its sources are in that report's assumptions.md:

The Street's revenue (market data, read 2026-10-26) is $3.69B, $4.14B, $4.65B, $5.14B and $5.78B for FY27-31: within 5% of ours each year, but compounding at 11.8% from FY27 to FY31 against our 10.2%. That growth gap is where we differ from the Street on the business (section 5).

2. The exit multiple: why 41.9x and not 48.7x

Anchor (d) for Veeva is 48.7x FY32 operating profit: the mean of Veeva's own ten-year median (55.9x, fiscal years 2017-2026) and the median of six peers' own medians (41.4x: Adobe 41.2x, Intuit 41.6x, Autodesk 49.3x, Tyler 66.0x, PTC 40.1x, Roper 29.6x; Salesforce, Workday and Guidewire lack five valid years). Veeva's 55.9x came while its revenue grew a median 24.9% a year; the peers' while theirs grew 12.7%. Our central case grows 9.1% in FY32 (anchor_growth.txt). The anchor studies did not condition on growth, and left open whether to discount a company's own fast-growth median (docs/research/studies/2026-10-23-exit-anchor-survivors/results.md, "Open item").

What we use: what companies that slowed like Veeva actually fetched. In the survivor-free study's software sample (starts 1996-2021, listed and delisted companies, peers as they stood), we take every start whose own multiple was earned at 20% growth or more and which grew 12% or less five years later: 86 starts from 29 companies (Adobe, Autodesk, Cadence, Synopsys, Check Point, F5, Fair Isaac, Microsoft, Oracle, PTC, Roper and others). Their own multiple five years on was a median 0.862 times the anchor (d) they started with (exit_outcomes.txt):

48.7x × 0.862 = 41.9x, our outside view. It is the report's one change from the method's default; the method still sets the bear and bull multiples from it (29.1x and 62.1x, the software group's 25th and 75th percentiles).

Why the ratio is used as it stands, not relative to all software: across all 605 software starts the realised multiple ran 0.83x the anchor for 1996-2009 starts and 1.30x for 2010-2021, which is why the method leaves the anchor unscaled. Starts like Veeva ran 0.90x and 0.83x in the same two eras. Their discount does not depend on which decade comes next, and it is below 1 in both. These companies also slowed further than Veeva (a median 4.8% five years on, against Veeva's 9.1%), so the ratio may be harsh rather than lenient.

What our blind draft did, and why we dropped it. The draft (commit 631134c) moved each leg of anchor (d) along a slope measured across 610 software company-years in 2016-2025 (ten points of growth carried 1.35x the multiple), which gave 35.9x and $273. Our reviewer tested that slope against outcomes (review/redteam/backtest_growth_move.txt, rerun by us with the same result): applied to past starts, even with the actual growth five years on given in advance, it predicted the realised multiple worse than the unmoved anchor (mean absolute log error 0.553 against 0.460), in both eras. Regressing realised over anchor on the growth gap gives 0.12 (standard error 0.21), not 3.03. A cross-section of companies at different growth rates overstates how much one company's multiple falls as it slows: part of the gap is quality that stays. Our auditor separately found that the draft's medians lost the pairing of each year's multiple with that year's growth (+$11.30) and that the slope's choices of range and company effects moved the value by up to $26 (review/codex-audit.md). The draft's figures stay in the table below as one view.

The judgment range, and which anchor gives which rating (market.txt; operating case fixed; at $318.34 the rule says Avoid below a fair value of $270.59 and Buy at or above $318.34):

Our judgment range is $273-304: from the draft's growth slope to the earlier era's ratio. All of it is Hold. Below it, the rows rest on today's or a sub-group's multiples, which the method's ten-year window does not use; above it, only the unmoved anchor, which ignores that Veeva will grow at a third of the rate its own multiple was earned at.

3. Value: $296, exit $360

Fair value = free cash flow paid out each year plus the exit price, both discounted at 9.57%, plus net cash of $7,209M ($43.68 a share) at face today. Cost of equity 9.57% = 5.29% + 1.018 × 4.20% (the method's dated risk-free rate and Damodaran's 2026-10-01 premium; beta the peers' median, Blume-adjusted). Veeva's own beta (Blume 0.956) gives $302; Damodaran's software beta (Blume 1.19) $282.

The central exit in other terms: $360 is 28.7x FY32 operating profit, 34.3x FY33 NOPAT. At a 40% mature return that implies 7.2% growth for ever, against the fade's own 4.8%.

Sensitivities (central, output.txt): revenue growth ±1.5 points a year +$22 / −$20; Commercial subscription growth −4 points a year (a deeper CRM loss) −$17; non-GAAP margin ±2 points ±$15; stock pay 1.5 points lower +$11; mature return 25% or 60% −$2 / +$1; exit ratio 0.833 or 0.904 −$6 / +$8, unmoved anchor +$26.

4. The flags

Flag 2: the view changed from Avoid to Buy without explanation. The change came from the method, in steps (bridge.txt; the split inside the restatements is our auditor's, rerun in bridge.py):

The Avoid became a Buy through the outside view and the rule change, with no change to Veeva's forecast. We keep the method's structure and change one input, the exit multiple, for a reason specific to Veeva: its own multiple was earned at growth its forecast does not have, and companies in that position fetched less. The view is now Hold, $94 above the original report and $25 below the restatement. The original's 19.8x NOPAT exit is not restored: it was the fade alone, and the method now averages the fade with an outside view by the owner's decision of 2026-10-16 (the fade alone had put 23 of 23 reports below the Street).

Flag 4: the outside view is far from the fade. The central outside view ($527) is 172% above the fade ($194). The input behind the gap is the exit multiple against the fade's path: the fade pays 15.4x FY32 operating profit, because 8.5% growth fades to 3% over 15 years at 9.57%; the outside view pays 41.9x, what companies that slowed like Veeva fetched. Which do we believe? Neither alone. The fade is our own discount rate applied to our own path; the outside view is what buyers paid. The method takes their mean by rule. The 2026-10-18 study supports the outside view's construction (own and peers' medians together missed less than either alone); it did not test the fade-and-outside-view mean, which our draft wrongly implied.

Flag 1: the move from the blind value. $273 to $296 (+$23.48, +8.6%). The one change is the exit ratio above. It rests on historical outcomes in the study's data, rerun by us, not on a management claim or a Street figure. It moves toward the price, and the reviewer who found it was arguing the Buy side; we accepted it because the draft's slope failed the test the outside view exists to pass.

Flag 19 (judged exit more than 1.5x from anchor (d)) does not fire: 41.9x is 0.86x the anchor.

5. Market expectations and how we differ

• What the price needs: at $318.34, revenue growth about 1.5 points a year above ours FY28-33, which is the Street's path (11.8% a year FY27-31 against our 10.2%) and close to management's $6B goal (+1.3 points); or an exit at 47.6x, nearly the unmoved anchor; or margins 3 points higher.
• The Street targets $296.50 on average, below the price; discounted a year at 9.57% it is $270.60. The three targets set in the last month averaged $345 (market data summary).
• Alpha's view (Capfolio's ranking model, alpha rank VEEV, 2026-10-26, cache marked for rerating): rank 84 of 696, score 88.1, Strong; momentum A (rank 32), quality, growth, expectations and valuation C.
• Timing, beside the view: a strong uptrend, the highest close since at least January at $323.04 on 2026-10-23, above its 50-day average of $277 and up 32% since the Q2 release. Momentum is not an input to our value.

Where we differ: growth from FY29, when the six top-20 accounts that chose Salesforce leave Veeva CRM (support ends 2029-12-31). Our Commercial subscription line slows to 5-6% then; the Street's FY30-31 revenue rests on eight analysts and does not show the dip. What settles it: the FY28 guide (about 2027-03-03) and Commercial growth through FY29. If revenue compounds near the Street's 12%, our value reaches the price.

6. Why the other side could be right

• Growth may not fall to 9%. R&D and Quality grew 19% in H1 FY27; AI products (Falcon, Study Builder Agent) could add a price line; the CEO expects to win back some of the six top-20 accounts in 2027-28. Street- like growth (+1.5 points) gives $319, the price; the bull gives $468.
• Veeva's premium may outlast its growth. A regulated system of record, no debt, 45% margins before stock pay. The unmoved anchor gives $322.
• Our ratio rests on 29 companies and overlapping start years; the two eras give $291 and $304.
• Against us on the other side: the 2016-2025 window we keep was rich by longer history. On the latest year-end multiples the value is $251, an Avoid; the fade alone is $193.

7. How each review finding changed the value

Two reviewers read draft 631134c separately: a red-team advocate for the Buy side (Claude Opus, red-team skill; review/red-team.md, scripts in review/redteam/) and a model auditor (Codex, codex exec, read-only, reasoning effort xhigh; review/codex-audit.md, script review/codex_audit.py, output review/codex_audit.txt, run against 631134c). Every figure below was rerun by us.

Net: $273 to $296. The view stays Hold. The auditor did not rerun after the change (one round; the call did not change); the replacement ratio was checked by rerunning the red team's backtest, which reproduced exactly.

8. Catalysts, falsifiers and the next results

• 2026-11-05: Investor Day (2030 goal, CRM share, margins).
• Q3 FY27 results, expected late November (not dated by the company): the 2026-10-15 report's forecasts stand and are graded from that record (revenue $962.1M, range $950-972M; non-GAAP operating income $438.5M; non-GAAP EPS $2.47). The model is unchanged, so this report adds none. A Q3 beat matters for value only if it carries into the FY28 guide: 1.5 points of growth a year is $22 a share.
• About 2027-03-03: Q4 FY27 results and the FY28 guide.

Falsifiers: FY28 revenue guided at 15% or more, or R&D and Quality holding near 18% (toward the bull, $468); two or more departing top-20s back on Vault CRM by 2028; Commercial subscription growth below 5% in any FY28 quarter or an FY28 non-GAAP margin below 43% (toward the bear, $172); evidence that slowing software leaders keep their full historical multiple (the method's 48.7x and $322).

Revision history

• 2026-10-26: same forecast; the exit multiple at anchor (d) times what Veeva-like starts fetched five years on, replacing the restated Buy ($322).

Sources

• Veeva filings and calls saved with the 2026-10-15 report (../2026-10-15-veev/data/): 10-K FY26, 10-Qs Q1 and Q2 FY27, 8-K exhibits through 2026-08-26, three call transcripts.
• SEC EDGAR filing index via scripts/research-events.py, 2026-10-26.
• Veeva news page, read 2026-10-26.
• Market data fiscal-year-end enterprise values, income statements and revenue for Veeva, its nine peers (2026-10-26) and the software universe of the survivor study (fetched 2026-10-24); the study's own data for the realised multiples; data/.
• Damodaran implied equity risk premium and the method's risk-free rate (2026-10-01).
• After the blind commit 6191a78: market quote, daily bars and analyst ratings; market data quote, price targets and estimates (via scripts/research-check.py); Alpha rank.