Financebotresearch desk研究台

DOCU › analyze

DOCU · Analyze from before

HOLD Technology

Date: 2026-08-27 | Baseline: 2026-05-06 (conviction 6.5, HOLD) | Price then: $46.53 → Price now: $63.80 (+37.1%)

1. What this updates

Baseline file: Output/Stocks/Technology/DOCU/analyze-2026-05-06.md, the company's only prior verdict-bearing report. No second prior report exists, and no health/moat/value/sentiment/earnings file was written on DOCU between the two dates — this pass is the first re-test of every claim in the file.

A portfolio-specific passage was removed from the public build.

Event list since baseline (what actually happened): 1. Q1 FY2027 earnings, reported ~June 4, 2026 — the baseline's own named catalyst. Revenue $830.2M (+9% YoY, beat $824M consensus), adj. EPS $1.09 (beat $0.99-1.00 consensus). IAM reached 12.6% of ARR, up from 10.8% at the Q4 FY26 exit (11% as baseline reported it). FY2027 guidance raised to $3.490–3.502B revenue (~9% growth) from the initial ~$3.42B (~6%) guide baseline flagged as the disappointment. Non-GAAP op margin guide raised to 30.5–31.0%. Record quarterly buyback: $317.5M. Dollar net retention >102%, +1pt YoY. 2. Billings grew only 3% YoY in Q1 FY27 against 9% revenue growth — a new, unflagged-in-baseline decel signal (billings lead revenue). 3. DocuSign integrated IAM into Google Cloud's Gemini Enterprise for Legal (~Aug 2026) via a DocuSign-built connector — DOCU stock fell ~3.2% on the announcement (AI-disruption fear), even though the integration makes DocuSign a component inside the AI-agent stack rather than a target for replacement. 4. Continued, uninterrupted insider selling — CEO, CFO, COO/President, CLO and most directors sold on a near-monthly cadence from March through August 2026, all at prices well below today's spot. Zero insider buys found in the window. 5. No confirmed material progress by OpenAI's DocuGPT since the Sept-2025 announcement that baseline cited as the "primary catalyst" for the 2025 collapse — coverage as of August 2026 still frames it as a multi-year risk, not a realized one. 6. Stock re-rated up 37% with no single confirmed catalyst — an Aug-13 sector-wide software rally (+8.4% that day) accounts for part of it; the stock is up ~18% in the trailing month into the Sep-3 print. Sell-side has not kept pace: mean analyst target $57.52 is now below spot ($63.80). 7. Q2 FY2027 earnings have NOT yet printed — due ~Sep 3, 2026, about one week after this report. This is the baseline's own named checkpoint ("IAM reaches 15%+ of ARR by Q2 FY2027 = upgrade signal") and it remains open.

2. The delta ledger

No RETRACTED or SUPERSEDED rows this pass — the baseline's structural read and its own self-caught trap (the FY2025 DTA windfall) both held up. Leading with the two most consequential REFRESHED rows, then the rest.

# Claim (baseline) Type Status Old → New Force / source
1 Vendor fwd P/E 9.24x (baseline, basis unstated) Price 🔄 re-derived, not carried 9.24x → true current-FY 14.10x (vendor forwardPE now 12.52x, also wrong-year) pitfall-vendor-forward-eps-is-the-wrong-fiscal-year fired again — see §3
2 "$1.76B buybacks yielded only 4.1M net share reduction (2%)" Trend 📈 REFRESHED — materially better 4yr net reduction 2% → diluted shares −7.7% YoY (212.8M→196.5M, Q1FY26→Q1FY27); record $317.5M single-quarter buyback fin.py + yahoo quarterly income stmt, 5-quarter series
3 Billings "record, >$1B" (Q4 FY26, no growth rate given) Trend 🆕 NEW, negative — → +3% YoY in Q1 FY27 vs +9% revenue growth WebSearch, corroborated by two independent outlets (StockTitan, Fool transcript)
4 FY2025 net income $1.068B inflated ~$840M by DTA release; normalized ~$248M State ✅ CARRIED, and validated by time TTM net income now $315.2M, clean (the DTA quarter has fully rolled out of the TTM window) Sum-of-4-quarters check: $78.2+90.3+83.7+63.0=$315.2M, matches vendor TTM exactly
5 FY27 revenue guide ~$3.42B (~6% growth), "fell below 10% threshold" Trend 🔁 REFRESHED, upgraded $3.42B guide → $3.49–3.50B guide (~9%), raised at Q1 print Company press release, 3 independent outlets
6 Gross margin "extremely stable ~79%" Trend 🔁 REFRESHED, mild erosion Non-GAAP GM 82.3% (Q1 FY26) → 81.5% (Q1 FY27), GAAP GM steady 79.4% Company Q1 FY27 press release
7 SBC consumes 58.7% of FY26 FCF Trend 🔁 REFRESHED, improving 58.7% → 55.2% (TTM SBC $618.1M / TTM FCF $1,120.2M, sum-of-quarters) Yahoo quarterly cashflow, reconstructed TTM (guard-compliant)
8 ROIC ~30%, improving Trend ✅ CARRIED Op income and buyback-shrunk equity both point the same direction; no reversal fin.py + quarterly statements
9 IAM 11% of ARR (Q4 FY26), target 18% by FY27-end State 🔁 REFRESHED, on pace 11% → 12.6% (Q1 FY27), +1.8pt in one quarter Company press release
10 Net retention — not quantified in baseline State 🆕 NEW, positive — → >102% direct-customer NRR, +1pt YoY Company Q1 FY27 transcript
11 Institutional ownership ~87% State 🔁 REFRESHED 87% → 93.5% Yahoo get_stock_info
12 Insider selling, no buying — yellow flag State ✅ CARRIED, unchanged Selling continued every month Mar–Aug 2026, zero buys Yahoo insider_transactions, ~30 rows reviewed
13 Analyst consensus "predominantly neutral, show-me mode" State 🔁 REFRESHED, now a caution Mean target $57.52 vs $46.53 spot (above) → mean target $57.52 vs $63.80 spot (below) Yahoo get_stock_info; sell-side has not caught up to the rally
14 Zero financial debt, clean balance sheet Structural ✅ CARRIED Debt still $183-185M (capital leases only); D/A ~4.4% fin.py balance sheet
15 Moat: brand, switching costs, network effects, AI-data moat, compliance Structural ✅ CARRIED, one reinforcing datapoint New: DocuSign became a connector inside Google Gemini Enterprise for Legal rather than a target displaced by it — consistent with baseline's own "compliance/audit-trail layer is the real moat" defense read WebSearch, DocuSign IR press release
16 Disruption threats: OpenAI DocuGPT, Microsoft Copilot bundling, Adobe Structural ✅ CARRIED No confirmed enterprise-scale DocuGPT traction found; still framed as multi-year risk by outside coverage WebSearch, multiple outlets
17 Evergreen rating 6/10 Judgment ✅ CARRIED No structural break; deferred to Phase 5 with everything else —
18 Graham IV $17.97, not weighted (asset-light SaaS) Price ✅ CARRIED New value $18.08 (immaterial move); same conclusion — not appropriate for this business fin.py
19 "IAM reaches 15%+ of ARR by Q2 FY27 → upgrade" Judgment (baseline's own open item) ⏳ UNTESTED — the print hasn't happened Interim Q1 read (12.6%) is consistent with reaching 15%+ but does not prove it Q2 FY27 report due ~Sep 3, 2026
20 Fair value $50-60 base / $75-90 bull / $35-40 bear Price 🔄 re-derived from scratch See §6 —
21 Entry/trim price bands ($40-43 add … $65-80 trim) Price 🔄 re-derived from scratch, converted to multiple See §6 —
22 Conviction 6.5, verdict HOLD Judgment 🔁 REFRESHED to 6.0 See §4 —

3. How the close calls were decided

Row 1 — the forward P/E. This is the pass's single most consequential finding and the exact failure mode the task brief warned about. Yahoo's forwardPE (12.52x) and forwardEps ($5.098) do not reconcile against a bottom-up estimate: the company's own FY2027 guide implies non-GAAP operating income of roughly $1.07-1.08B (30.5-31% margin × $3.496B revenue midpoint); adding net interest income and taxing at DocuSign's typical ~19% non-GAAP rate on ~193M diluted shares produces ~$4.5-4.7 EPS — matching Yahoo's other field, epsCurrentYear ($4.524, priceEpsCurrentYear 14.10x), almost exactly. forwardPE/forwardEps is on the next fiscal year (FY2028), same defect documented on TSM, QCOM, AMD, ANET, MDT, LEN. fin.py's PE(fwd) field inherits it. True current-FY (FY2027) P/E is 14.10x, not 12.52x — and the baseline's own 9.24x is very likely the same defect, though it cannot be re-verified retroactively since the underlying vendor snapshot from May is not archived. Treated as re-derived, not carried, per the Price-row rule; flagged as a live instance of pitfall-vendor-forward-eps-is-the-wrong-fiscal-year (see Flags, below — this pass did not edit the note, since editing Knowledge/Playbook is outside this delegation's stated scope).

Row 2 vs Row 12 — buybacks working vs. insiders selling. These point in different directions and were checked for contradiction before either got a status. They are not actually in tension: the buyback is a board-level capital allocation decision funded by real FCF (confirmed: TTM FCF $1,120.2M, sum-of-quarters, comfortably covers the $1.0B TTM buyback pace), and it is visibly shrinking the share count now (−7.7% YoY diluted shares) in a way the FY23-26 annual data did not show. Insider selling is personal, near-monthly, and pattern-consistent with scheduled 10b5-1 plans — no lumpy, off-cycle sales that would suggest a change in view. Both are recorded as-is: buybacks REFRESHED positive, insider-selling flag CARRIED unchanged. Neither explains away the other.

Row 3 — billings deceleration vs. row 5's raised guide. This is the genuine tension in this pass. Revenue guidance went up (a Trend force in the thesis's favor) while the leading indicator that predicts revenue — billings — grew at a third of revenue's rate in the same quarter (a Trend force against it). Billings is the earlier signal in the causal chain (booked-but-not-yet-recognized business), so weight was given to it as the more forward-looking number: the raised guide reflects business already booked; the weak billings print is what determines whether FY2028 guide can be raised again. This is why the row is scored NEW-negative rather than dismissed, and why it is the single most important thing the Sep-3 print needs to resolve — a rebound in billings growth toward high-single-digits/low-double-digits would neutralize this concern; a second quarter of sub-5% billings growth would be the first real evidence the deceleration thesis (baseline's "sub-10% growth doesn't justify adding") is worsening, not stabilizing.

Row 15 — the Google Gemini integration: structural reinforcement or Price force in Structural costume? Per the command's own guardrail, "the AI narrative shifted" is not automatically a business-model change. Tested against that bar: DocuSign built the connector, retains the customer relationship and the compliance/audit-trail data, and is not disintermediated — this is a genuine (if early) confirmation of the specific moat defense the baseline's Moat Analyst named in May ("the legal infrastructure layer... hyperscalers cannot easily replicate"). It earns a CARRIED-with-reinforcement, not a new Structural claim, because it is one data point on an existing thesis, not a new mechanism. The market's -3.2% reaction on the announcement is a separate, Price-level fact and is not read as evidence against the Structural claim.

4. Thesis persistence and conviction delta

Thesis persistence: 100% (9 of 9 baseline Structural + Trend claims survived as CARRIED or REFRESHED; rows 1, 3, 10, 19, 20, 21 are Price/State/NEW/Judgment rows and are excluded from this count by the command's own formula). Every Structural claim held without qualification (rows 14-16). Every Trend claim moved in a favorable or neutral direction (rows 2, 5, 6, 7, 8) — margin softened 80bp, everything else improved.

This is the textbook case the command's own guidance names explicitly: high persistence against a big price move. The business held; the multiple moved. At baseline, DOCU traded at what the report called a genuine discount (8.5x reported P/FCF, 9.24x claimed fwd P/E) with real FCF and a clean balance sheet — "you're paying almost nothing for the IAM option." At $63.80, on a corrected 14.1x true fwd P/E and a re-derived base fair value of $55-70 (§6), that free option has been substantially priced in.

Conviction: 6.5 → 6.0. Named drivers: - Down: the entry-zone opportunity that was the core of the May bull case (buy under $46, cheap on every metric) no longer exists at $63.80 — margin of safety is gone (rows 1, 13, 20, 21). - Down: the new billings-deceleration signal (row 3) is unresolved and sits directly on top of the exact "core decel worry" this pass was told to watch for. - Down, temporarily: an imminent binary catalyst (Q2 FY27 print, ~Sep 3) with the baseline's own named checkpoint still open (row 19) adds near-term uncertainty that should not be papered over with false confidence either way. - Up: every Trend row that moved, moved favorably — the share count is genuinely shrinking now (row 2), SBC is a smaller drag on FCF (row 7), IAM is tracking toward its target (row 9), and net retention improved (row 10). - Net: these roughly offset in direction but not in magnitude — the valuation cushion loss and the new billings risk are judged to slightly outweigh the fundamental improvement, because the fundamental improvement was already partially priced in by the 37% rally, while the billings risk is not yet resolved either way. This is a close call, not a confident downgrade.

5. What is genuinely new

  • Billings growing at a third the rate of revenue (row 3) — no baseline counterpart; this is the finding most likely to move the verdict at the Sep-3 print in either direction.
  • Net retention >102%, improving (row 10) — baseline never quantified this; it is a genuine positive on the specific "core decel worry" flagged for this pass.
  • The Google Gemini Enterprise integration (row 15 discussion, §3) — a live instance of the moat-defense strategy the baseline predicted DocuSign would need, playing out in real time.
  • The vendor forward-P/E defect confirmed on DOCU specifically (row 1) — extends a pattern already documented on 9 other tickers to a 10th, and very likely explains part of why baseline's own valuation looked as cheap as it did.
  • Analyst mean target now below spot ($57.52 vs $63.80) — a genuinely new state: "Hold" from the sell side now means something different than it meant in May.

6. Updated verdict

Action: HOLD. Do not add — no margin of safety remains at $63.80. Do not trim — position is still too small to matter mechanically, and the fundamentals have not broken. The next seven trading days (through the Sep-3 print) are the highest-information window this thesis will see all year.

Valuation (re-derived from scratch, per Price-row rule)

TTM figures (sum-of-four-quarters, guard-compliant): revenue $3.286B, net income $315.2M (clean — no DTA distortion), FCF $1,120.2M, SBC $618.1M, SBC-adjusted FCF (FCF − SBC) $502.1M. Shares outstanding 190.94M. Cash + ST investments ~$814M.

Basis Multiple range Implied fair value/share
Reported FCF ($1,120.2M) 10-14x $63 - $86
SBC-adjusted FCF ($502.1M) 18-22x $52 - $62
True current-FY non-GAAP EPS ($4.52) 12x (bear) / 15-18x (base-bull) $54 - $81

Base fair value: $55-70. Bull fair value: $80-95 (IAM hits ~18% of ARR by FY27-end, growth reaccelerates, re-rates to 18-20x true fwd P/E). Bear floor: $40-48 (billings deceleration continues into revenue, IAM stalls sub-15%, DocuGPT/AI-agent disintermediation accelerates) — this floor is higher than baseline's $35-40 despite no change in business risk, purely because the FCF/share base is larger on a smaller share count.

At $63.80, the stock sits at the top of base fair value / bottom of bull fair value — fully, not cheaply, priced for the "IAM works" case, with no discount left for the "billings say slow down" case.

Entry / trim (multiple-based, per CLAUDE.md valuation rule)

Zone Level Action
Strong add <$45 Deep value re-opens only on a genuine breakdown
Add $45-52 Back inside base-fair-value floor with margin
Hold $52-77 Current price ($63.80) sits here — no action
Trim 20x fwd (current-FY basis; ≈$90 at today's $4.52 EPS, recomputed live by the site) Full trim ceiling — matches bull-case fair value top
Partial trim (25-50%) 17x fwd (≈$77) First profit-taking level if the print confirms IAM ≥15% ARR and the stock runs into the mid-$70s-$80s

Trim basis stated explicitly per the pitfall guard above: the 20x multiple is applied to epsCurrentYear ($4.524, current-FY FY2027), not the vendor's forwardPE/forwardEps field, which is one fiscal year ahead and would set the trim ~13% too high. Whoever recomputes this on the site should verify against priceEpsCurrentYear, not forwardPE.

Break triggers (updated)

  • IAM still under ~13% of ARR after the Sep-3 (Q2 FY27) print — the baseline's own downgrade trigger, now concrete.
  • Billings growth stays at or below mid-single-digits for a second consecutive quarter — new trigger, this pass's addition; the leading indicator the baseline never tracked.
  • A confirmed enterprise-scale DocuGPT or Gemini-native displacement (a named customer loss, not just a partnership announcement) — unchanged from baseline.
  • Insider selling turns lumpy/off-cycle (a sale outside the routine near-monthly cadence, or a resignation) — refinement of baseline's existing yellow flag.

Upgrade conditions

  • IAM ≥15% of ARR and billings growth reaccelerates above high-single-digits at the Sep-3 print → conviction back toward 7.0, consider whether the position is worth sizing up from token size.
  • A pullback into the $45-52 add zone on no change to the fundamental picture → the May thesis re-opens at a real discount again.

7. What this pass did NOT test

  • Q2 FY2027 results — the single most decisive open item, due ~Sep 3, 2026, after this report's date. Row 19 is explicitly UNTESTED, not assumed to be trending toward either outcome beyond the interim Q1 datapoint.
  • The exact cause of today's (Aug 27) 7.5% single-day move — corroborated only that DOCU has been on a broad software-sector rally through August (sector +8.4% on Aug 13 alone); no company-specific news item was found that explains today's move specifically. Treated as unexplained price noise, not a fundamental force, and not used to support any claim in this ledger.
  • A direct reconciliation of the baseline's original 9.24x forward P/E — flagged as very likely the same wrong-fiscal-year defect found today (row 1), but the underlying May vendor snapshot is not archived and cannot be re-pulled to confirm with certainty. Recorded as a single-source, non-corroborated inference, not a fact.
  • The 10-Q's own billings disclosure verbatim — the 3%-YoY figure is corroborated across three independent outlets (StockTitan, Fool transcript, Benzinga) citing the same press release, but this pass did not open the primary 10-Q filing itself. Given three independent secondary sources agreeing on the same specific figure, this was judged sufficient corroboration rather than single-source risk, but a primary-source read is the natural next step before the Sep-3 print if time allows.
  • Deep-dive Moat Analyst adversarial stress-test refresh — the Structural claims were re-tested against "did anything happen," per the command's standing prior that these move slowest, rather than fully re-run from scratch. No evidence surfaced that would have changed the answer, but a full stress-test was not repeated line-by-line.

Sources