Phase Space AI

Valuation

Doximity [DOCS]

Doximity [DOCS] — Valuation

Spot $21.62 (2026-07-29) · shares 183.060m (filed balance sheet, 31 Mar 2026) · market cap $3,957m Net cash +$748.6m = cash $219.178m + marketable securities $529.423m − debt $0 (operating leases $10.2m excluded by convention; stated per §9.7). EV = $3,208m. TTM revenue $644.9m · TTM GAAP EBIT $214.9m → EV/Sales 4.98x · EV/EBIT 14.9x

---

1. COMPANY STATE — declared first, because it selects the instrument

STATE A — mature and structurally stable.

This contradicts the brief's working hypothesis of State C, and the contradiction is the point. State C requires "pre-profit or thin-margin." DOCS is neither, and applying a State-C instrument to it would be the error valuation.md was written to prevent.

The test, run explicitly:

Why it is not State C: State C's test is "pre-profit or thin-margin, but with usable evidence." DOCS has an 89.1% gross margin and a 33.3% operating margin. There is no margin inflection to underwrite — the margin is already earned and is currently contracting. Growth deceleration is not a State-C condition; it is a State-A company whose growth rate is falling.

Consequence of State A (per valuation.md): the instrument is "fix the underwritten margin, solve for price-implied growth, and show the neighbouring sensitivity surface." That is what §3 does. Terminal margin is built from the company's own normalised economics; industry data is a sanity band, never an override, and no percentile clamp is applied.

Evidence grade: A. Audited statements, five-year margin history, explicit forward guidance, and a consensus that matches guidance to the decimal.

---

2. Terminal margin — built, justified, and bridge-reconciled

m_EBIT,T = 34.0%

Source: the company's own demonstrated operating margin and its stated trajectory. No default, no median, no cap. DOCS's five-year mean GAAP operating margin is 34.1%; trailing FY2026 is 33.3%. 34.0% sits between them.

Opex bridge (per valuation.md, on FY2026 actuals held forward):

m_EBIT,T  =  m_gross,T − R&D − S&M − G&A
   34.0%  =    89.1%   − 20.0% − 25.0% − 10.1%

Reconciliation against FY2026 actual: 89.1% − 20.3% − 25.3% − 10.2% = 33.3% ✓ (the filed figure). The terminal bridge therefore holds the current expense structure essentially flat, with R&D easing 0.3pp as the AI build normalises — it does not assume expansion.

Hard constraint: m_EBIT,T (34.0%) ≤ m_gross,T (89.1%) — satisfied with 55.1pp of headroom. The sufficiency test also passes trivially: 55.1% of revenue remains for all R&D, selling and administrative cost, against a company currently spending 55.8%.

Is it below the company's own trailing actual? No — 34.0% > 33.3%. Per the brief, a terminal margin below trailing would require explicit justification; none is needed here.

Why not higher (e.g. FY2025's 39.9%)? Because FY2025 was the peak of a five-year range and the margin has since contracted 6.6pp on a deliberate R&D step-up that management has guided to continue (FY2027 adjusted EBITDA guided down 6–10%). Underwriting the peak would be the mirror of the defect §9.8 warns about. 34.0% is the demonstrated central tendency, which is the correct anchor for a State-A name.

Sanity band (not an override): high-margin ad-supported platforms run 25–40% GAAP operating margins. 34.0% sits inside that band. The band is reported and is not permitted to move the number.

---

3. Implied-path test — THE VALUATION CRITERIA (primary long-horizon output)

Terminal value is ~86% of EV on this build, well above the 60% threshold, so the reverse DCF is mandatory as the primary output and a forward DCF is supporting evidence only.

3.1 The exit multiple is DERIVED, not assumed

Per valuation.md rule 4, the exit multiple is not a free parameter:

EV_T / EBIT_T  =  (1−t)(1−g/ROIC) / (WACC−g)
               =  (1−0.21)(1−0.03/0.40) / (0.09−0.03)
               =  0.79 × 0.925 / 0.06
               =  12.18x

Parameters and their basis: t = 21% (US statutory; DOCS's effective rate is below this, so the multiple is conservative). g = 3.0% — terminal growth, set at nominal GDP-like and below the FY2027 guided 3.9%, because a pharma-marketing budget line cannot compound above its end market indefinitely. ROIC = 40% — a deliberate haircut from the measured >70%, since terminal ROIC should not assume the current near-zero-invested-capital structure persists. WACC = 9.0% — unlevered, no debt, β>1 on a 74%-drawdown equity.

Implied compression, stated as a number: current 14.9x EV/EBIT → exit 12.18x = −2.72x, −18.3%. This is a compression, not an expansion, and it is consistent with a business decelerating from 13% to ~4% growth. Per valuation.md rule 3, an exit multiple below today's is often correct and is so here.

3.2 The result

3.3 But the margin against guidance is the number that matters, and it is negative

The demonstrated CAGR is a trailing measurement of a growth rate the company has already told us has ended:

The price requires DOCS to compound revenue at 13.1% for five years. Management guides 3.9% for next year and consensus agrees to the decimal. Nothing in the disclosure bridges 3.9% to 13.1%.

Verdict on the Valuation Criteria: FAIL. Per criteria.md: PASS WITH ARGUMENT requires "a specific, evidenced reason (a named product cycle, mix shift, pricing action)" for the price to require more than demonstrated. DOCS has a candidate — the Clinical AI Suite — but §2 of the research establishes it has no attributable disclosed revenue, which the criteria explicitly disqualifies: "Narrative does not qualify." The required 13.1% sits above the most recent actual year, far above guidance, and above consensus in both forward years.

Stated fairly against the name: the margin versus the trailing 3-year CAGR is genuinely positive (+2.4pp), and if one believes the trailing rate is the right basis, this is a marginal PASS. The reason it is scored FAIL is that the company itself has published a different number for the forward period, and a dated management guide outranks a trailing average as evidence about the future. This is the specific judgement in this memo most likely to be wrong, and it is flagged as such.

3.4 Sensitivity — over the EXIT MULTIPLE, never over scenario probabilities

Required revenue CAGR, terminal margin held at 34.0%:

The flip point is where the real judgement lives: the price is justified on guidance (~3.9%) only at an exit multiple of ~18.5x EV/EBIT — i.e. above today's 14.9x, on a business growing 4%. That requires the multiple to expand through a deceleration, which the warranted-multiple identity does not support at any plausible (g, ROIC, WACC).

Reaching the screen's 1.3% required CAGR needs ~21.5x, which is 1.77x the identity-derived multiple. This is where the screen's +14.1pp came from.

Second sensitivity, over the terminal margin (exit held at 12.18x):

Even at a 40% terminal margin — above every year DOCS has ever printed except FY2025 — the price still requires 9.5%, which is 2.4x guidance.

---

4. 12-month target — the multiple anchor is UNIDENTIFIED

4.1 The anchor test, and why it fails

DOCS's own EV/Sales history, built daily from TTM revenue as known at each date (lagged 45 days to filing, so the series is never forward-looking), share count and net cash held at verified current values so the series measures multiple movement:

DECLARATION: the 12-month multiple anchor is UNIDENTIFIED.

Not because the history is short — 5.1 years and 1,277 observations is ample — but because it spans a regime change in both the multiple and the underlying business, and the percentile is therefore not measuring mean reversion.

Evidence that the regime changed, on both sides:

• Multiple regime: the median EV/Sales by year runs 49.00 → 17.23 → 12.23 → 9.86 → 18.26 → 5.76. There is no stationary distribution to take a percentile of. Restricting the window does not rescue it: the current percentile is 3.3rd on 2021+, 4.7th on 2023+, 10.7th on 2025+, while the median of those same windows is 13.98x, 12.16x and 16.05x. The percentile is stable and the level is not — which is the signature of a non-stationary series.
• Business regime, which is the disqualifying one: the multiples in that history were earned by a company with NRR of 157% / 117% / 114% / 119% growing revenue 20–66%. Today's company has NRR of 109% and guides +3.9%. Reverting to a 13.98x median would imply +181% on the equity. That is not mean reversion; it is pricing today's business at the multiple of a different business. This is precisely the error the brief names as "the defect that drove every high-growth rejection in this system's early record," run in reverse — and the discipline is the same in both directions.

No peer median is substituted. Per criteria.md and valuation.md rule 6, the reference class also fails independently: there is no set of n ≥ 5 mature profitable firms in physician-audience pharma advertising with dispersion in the matched dimension. UNIDENTIFIED stands.

4.2 The target, built on a stated non-percentile basis

Because the own-history anchor is unidentified, the 12-month target is built on an absolute free-cash-flow yield — and this basis is declared, not smuggled in as if it were the standard instrument.

Near-term base (per valuation.md step 1, from near-term consensus/guidance, not a house long-run forecast): FY2027 revenue $670.3m (AV consensus = guidance midpoint), FY2027 adjusted EBITDA guided $323–335m (midpoint $329m, −8.0% YoY). FY2026 FCF was $317.5m on $357.8m of adjusted EBITDA (88.7% conversion); holding conversion gives FY2027 FCF ≈ $292m.

Named product-cycle events inside 12 months (each dated in the Catalyst Calendar): Q1-FY27 print (Aug 2026) — the first test of the +3.9% guide; the Aledade and Photon partnership ramps; any first dollar disclosure for the Clinical AI Suite; the pharma budget-setting cycle (calendar Q4).

12-month target: $22.79, +5.4% to spot. Basis: 7.0% FCF yield on guided FY2027 free cash flow. Exit/anchor multiple basis: UNIDENTIFIED — own-history percentile is 3.3rd but is not admissible as an anchor (regime change); no peer median substituted.

Monotonicity check (valuation.md rule 2): V_bull $30.79 > V_base $22.79 > V_bear $17.03 ✓. Required on value, not on every parameter.

Consistency check against the implied path: the 12-month target (+5.4%) and the 5-year implied-path FAIL are not in conflict — they answer different questions over different horizons, which is why both are required. The name can be roughly fairly priced for twelve months and still require an unachievable five-year path, and that is exactly what this build says.

---

5. Comparator set used

For the exit multiple: NONE. Derived from the warranted-multiple identity using DOCS's own g, ROIC and WACC. For the 12-month anchor: NONE ADMISSIBLE. Own history declared UNIDENTIFIED (regime change); reference class fails the n≥5 / matching-operating-model / dispersion test of valuation.md rule 6. Sanity band only, not used in any calculation: high-margin ad-supported platforms at 25–40% GAAP operating margin.

This is a deliberate empty set. Recording it as empty is the finding; populating it with a healthcare-IT median would reproduce §9.9 (an anchor that is precisely identified and uninformative).

---

6. Criteria summary