ADP › analyze
ADP · Analyze from before
Date: 2026-08-27 · Baseline: analyze-2026-03-30.md (2026-03-30, verdict 8.5/10 Buy)
1. What this updates
The baseline was written 2026-03-30 at a price of ~$201.25, inside its own $195-210 buy
zone, on a Fundamentals + Moat + Sentiment + Valuation convergence read as "one of the cleanest
setups evaluated." Price today is $284.68 — +41.5% in five months, above the watchlist's
legacy Trim $280+ line. That is the live question this pass exists to answer: did the business
re-rate on new facts, or did the price simply run ahead of an unchanged story?
Event list since 2026-03-30:
- Q3 FY2026 earnings (~end April 2026) — occurred, not independently re-read this pass (superseded by the full-year print below).
- FY2026 closed (fiscal year end June 30) — full-year results and FY2027 guidance reported ~2026-07-29/30.
- Multiple sell-side price-target hikes in the 48 hours after the print (Citi, Morgan Stanley, Stifel, Wells Fargo, Guggenheim, UBS).
- One open-market insider purchase — a director, 2026-05-07, near the then-price low.
- ADP's own National Employment Report has kept printing monthly; the July 2026 read (+44K, released 8/5) missed consensus (+70K) and decelerated from June's +95K.
- Next dividend increase (the 52nd consecutive) still pending, expected around the Nov 2026 board cycle — not yet announced.
- No second prior report exists — this is the first /analyze-from-before pass on ADP, so there is no second-generation persistence data to draw on.
2. The delta ledger
Lead items — SUPERSEDED (all five Price rows, re-derived from scratch per rule) and DRIFTED:
| # | Claim (baseline) | Type | Status | Old → New |
|---|---|---|---|---|
| P1 | Fwd P/E 16.8x, 35-40% discount to 5yr avg ~27x | Price | 🔄 SUPERSEDED | Vendor headline fwd P/E is now 21.25x — but that is FY2028 EPS, one fiscal year too far out ([[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] fires again). Reconstructed true FY2027 fwd P/E ≈ 23.2-23.5x off guide-consistent EPS (~$12.10-12.26). Discount to the 5yr avg narrows to ~13-15%, not 35-40%. |
| P2 | Yield 3.38% vs 5yr avg 2.0-2.2%, 50-70% above | Price | 🔄 SUPERSEDED | Dividend unchanged at $6.80/yr. True yield (dividendRate/currentPrice, not Yahoo's previousClose-based field) = 2.39% vs 5yr avg 2.04% — now only ~17% above average, not 50-70%. Most of the March signal was price-driven, and it has now unwound. |
| P3 | FV $225-310 (conservative/base/optimistic) | Price | 🔄 SUPERSEDED | Re-derived: $225-250 / $260-290 / $300-325 (see §6). DDM is nearly unchanged in dollar terms (dividend didn't move) — spot has moved from the bottom of the March range to the top of the new one. |
| P4 | Entry $195-210 | Price | 🔄 SUPERSEDED | Zone is 27-30% below spot and unreachable without a real reset. New entry $225-250. |
| P5 | Watchlist Trim $280+ (legacy fixed dollar) |
Price | 🔄 SUPERSEDED | Spot ($284.68) had already breached this stale dollar line. Re-derived as 26x fwd on the corrected FY2027 EPS ≈ $315 — above spot, so the corrected trim is not yet triggered. The fixed-dollar line was the false alarm, not the stock. |
| St2 | Q2 FY26 beat, guide raised to 9-10% EPS growth | State | 🔄 SUPERSEDED | Full FY2026 landed: revenue +7% ($21.95B), adjusted EPS +11% (beat the raised guide), Q4 adj EPS $2.64 vs $2.59 est (+17% YoY), 9th-plus consecutive beat streak intact. FY2027 guide issued: revenue +5-6%, adj EPS +9-11%, EBIT margin +70-90bps, client-funds interest revenue $1.54-1.56B (up from $1.35B). |
| St4 | Analyst consensus bearish, median PT $272, mostly Hold/Sell | State | 🔄 SUPERSEDED | Six firms raised targets sharply in the 48h post-print (Citi $230→$287, MS $240→$286, Stifel $260→$285, WFC $248→$283, Guggenheim $270→$300, UBS $260→$270). Mean target now $287, median $286 — essentially at spot (+0.8%). The contrarian "Street underweights ADP" edge the baseline leaned on is gone; Street has caught up. Rating mix is still Hold-leaning (2.72/5) but no longer priced below the stock. |
| St8 | "Monitor Q3 FY2026 (Apr 29)" open item | State | 🔄 SUPERSEDED | Resolved — Q3 and the full FY2026 print both landed. Next catalyst is now Q1 FY2027, ~Oct 28, 2026. |
| St3 | PEO soft spot — worksite employee growth ~2%, margin −70bps | State | 📉 DRIFTED | Still ~2% worksite-employee growth, still contracting. New: FY2027 guidance explicitly states margin contraction continues into next year on pass-through growth and elevated healthcare costs — no longer a one-quarter mix effect, now a guided multi-quarter drag. Breaks if PEO revenue ex-pass-throughs goes negative or worksite-employee growth turns negative. |
| St7 | Employment deceleration (Feb 2026 +63K) | State | 📉 DRIFTED | July 2026 ADP National Employment Report: +44K, missed the 70K consensus, down from June's +95K — choppier and softer, not stabilizing. Not yet showing up in ADP's own numbers (bookings hit the top of the guided range — see S4 below), so this remains a headwind that hasn't broken the thesis, only pressed on it harder. |
CARRIED (re-tested, holds):
- S1 — Moat: WIDE. Retention still 92.1% ("third consecutive year at or above expectations" per the FY2026 print — a primary-source figure, not a screener read). Float income pillar strengthened: FY2027 guide of $1.54-1.56B vs FY2026's $1.35B is a 14-16% step-up, consistent with rates staying elevated. Regulatory-complexity and brand/trust arguments untouched by any event this period.
- S2 — Revenue model (Employer Services + PEO, recurring HCM/payroll). Unchanged.
- S3 — Evergreen assessment ("stronger in 10 years"). No disconfirming event.
- S4 — Competitive landscape. This is the moat's actual stress-test result, not just a re-assertion: the baseline's own adversarial case (Rippling capturing enough share to slow bookings growth toward ~6%) did not materialize — ES new-business bookings came in +6% YoY, at the top of the guided 4-7% range, and retention held at 92.1%. A predicted risk failed to appear; that is recorded as a force for the moat, not dropped silently.
- T9 — FCF payout ratio. FY2026: $2.63B dividends / $4.78B FCF ≈ 55.0% (fin.py FCF basis) vs baseline's 54.6% — materially unchanged, still comfortably covered.
- St9 — Dividend-increase timing. Ex-dividend 2026-09-11, payment 2026-10-01, still at the existing $1.70/quarter rate — the 52nd consecutive raise has not yet been announced, on the same November-board-cycle timeline the baseline expected. Genuinely open, not silently dropped.
REFRESHED (same claim, new number):
| # | Claim | Old | New |
|---|---|---|---|
| T1 | Revenue 3yr CAGR | 7.6% (FY22-25) | 6.8% (FY23-26) — mild deceleration, tracking toward the FY2027 guide of 5-6%. |
| T2 | FCF 3yr CAGR | 19.9% (FY22-25) | 9.5% (FY23-26). FY2026 FCF $4.78B, +8.9% YoY — still outgrowing revenue (operating leverage intact) but the CAGR more than halved. Capex rose 21.6% YoY ($547M→$665M), which explains part of the deceleration; see §3 for why this isn't read as a break. |
| T3 | Net income 3yr CAGR | 11.4% | 9.0% (FY23-26); FY2026 NI +8.1% YoY. Tracks revenue deceleration, unremarkable. |
| T4 | Margin expansion (4yr streak) | +340bps gross / +320bps op / +190bps net (3yr) | FY2026 delivered +80bps adjusted EBIT margin for the full year — same pace as the +80bps the baseline itself captured in the Q2 print, so the rate of expansion held even as revenue growth slowed. FY2027 guides another +70-90bps. |
| T5 | Share count declining | -0.8%/yr (FY25) | -1.3% FY2026 (403.3M diluted vs 408.7M) — buyback pace accelerated, not decelerated. |
| T7 | ROIC recovery | 27.3% (FY25, acquisition-diluted), "recover toward 35-40% within 2-3 years" | 40.1% (FY2026: $4.41B NI / $11.0B invested capital) — the baseline's own recovery call hit the top of its projected range in one year, not two to three. Strongest single confirmation in this pass. |
| T8 | Dividend CAGR | ~14% (3yr) / ~12% (9yr, estimated) | +10.9% YoY FY2026 (div/share $5.89→$6.53) — still double-digit, modest deceleration from the 3yr figure, consistent with a maturing but still-strong compounder. |
| St1 | Debt/deleveraging flag | Debt tripled to $9.07B FY2025, Net Debt/EBITDA 0.85x, "will monitor" | Debt fell to $5.27B FY2026 (-$3.8B), net debt ≈$1.0-1.3B, Net Debt/EBITDA ≈0.16-0.20x. The baseline's flagged concern is substantially resolved — one of the strongest positive deltas in this pass. |
| St5 | Insider activity | "100% selling, zero buying, 6mo" | One open-market purchase found: Director Robert Holmes Swan bought 3,619 shares (~$745K) at $206.05 on 2026-05-07 — near the price low, sized meaningfully, not a 10b5-1 grant/vesting/conversion. Everything else since (and before) remains routine officer sales, RSU vesting, and option conversions. Not "100% selling" anymore, though still overwhelmingly sell-weighted in dollar terms. |
| J1 | Verdict / conviction | 8.5/10, Buy | HOLD, conviction 7.0 — see §4-6. |
UNTESTED this pass — say so plainly:
- T6 — SBC discipline (~1.3% revenue). Not re-pulled this pass; no reason to expect it moved, but not verified against FY2026 actuals. Carry forward as an open check.
- St6 — Institutional accumulation. The baseline's Q3-2025 13F reads (UBS +58.5%, Wellington +47.8%, etc.) are now ~a year stale and were not refreshed this pass. Needs a fresh 13F pull before being cited again as a bullish force.
No claim was RETRACTED — nothing in the baseline was wrong when written; every SUPERSEDED row reflects a real fact change (a fiscal year closing, a guide issuing, a price moving), not an error caught after the fact.
3. How the close calls were decided
T2 (FCF deceleration) vs T4 (margin expansion holding) — apparent tension, resolved. A halved FCF CAGR alongside an intact margin-expansion pace looks contradictory until the capex line is checked: capex grew 21.6% YoY (WorkForce Software integration + platform investment), which is a reinvestment story, not a margin story — operating and EBIT margins expanded on schedule; free cash flow simply carries more capex through it this year. Revenue growth itself also decelerated in line with the FY2027 guide, which explains most of the rest. Verdict: not a break, but real — flagged as the trend row it is (REFRESHED), not waved away.
St3 (PEO drift) vs S4 (moat holding) — not the same claim. PEO margin compression is a segment-level mix effect (pass-throughs + healthcare cost growth) that management has now guided to continue; it does not touch Employer Services, which is ~70% of revenue and where the actual competitive stress-test (Rippling) sits. Reading PEO softness as evidence against the wide-moat call would conflate two different mechanisms — kept as separate rows, one DRIFTED (PEO) and one CARRIED/strengthened (S1/S4).
St7 (employment deceleration) vs S4/St2 (bookings at top of range) — the macro headwind hasn't reached the fundamentals yet. July's ADP print (+44K, missing consensus) is a genuine, worsening macro signal. But ADP's own bookings (+6% YoY, top of guided range) and retention (92.1%) show no sign of it yet — consistent with the baseline's own resolution ("payroll is mission-critical regardless of macro"). Both facts are recorded; the thesis isn't broken, but the cushion for a miss has thinned now that the multiple has re-rated up (see §6).
P1/P2 (the valuation collapse) — one mechanism, not two independent confirmations. The forward-P/E discount narrowing and the yield-premium narrowing are the same event read twice: the price rose 41.5% while the dividend didn't move and EPS grew at a normal double-digit pace. That's one re-rating, not two separate bullish-thesis-losing signals — noted so it isn't double-counted against conviction.
4. Thesis persistence and conviction delta
Thesis persistence (Structural + Trend claims only): 12 of 13 rows survived as CARRIED or REFRESHED (T6 SBC is UNTESTED, not failed) → 92%.
This is the textbook high-persistence pattern the protocol calls out — but read it correctly here. High persistence against a price move usually means "the business held and the multiple moved, so there's still room" (a re-rating still in progress). Here the multiple has already moved the same 40%+ the business earned, largely because the market caught up to a forward-year vendor-EPS artifact that (once corrected) never made ADP as cheap as the March headline implied. The business held. The opportunity that the March report was pricing did not repeat itself — it was captured.
Conviction delta: 8.5 → 7.0. Driven by, in order of weight: 1. P1-P4 SUPERSEDED — the discount to history compressed from 35-40% to ~15%, and the entry zone is 27-30% below spot and unreachable without a real pullback. The margin of safety that anchored the 8.5 is mostly gone. 2. St4 SUPERSEDED — the contrarian "Street is wrong" argument is gone; targets caught up to price. 3. St3 + St7 DRIFTED — two headwinds (PEO margin, employment) both moved further against the thesis, neither yet broken. 4. Offsetting upward: T7 (ROIC recovery ahead of schedule) and St1 (deleveraging resolved) are genuine positive deltas that keep this from falling further — the balance-sheet and capital-efficiency concerns the baseline flagged for monitoring are both now resolved favorably.
Net: quality unchanged-to-better, price no longer offers the cushion that made this a high-conviction buy. That is a HOLD, not a sell — nothing here argues for reducing the existing 1-share position — and not a fresh BUY either.
5. What is genuinely new
- The forward-EPS-wrong-fiscal-year pitfall fires on ADP. Yahoo's headline
forwardPE(21.25x) is built on FY2028 EPS ($13.40), not FY2027.epsCurrentYear/priceEpsCurrentYear($12.26 / 23.23x) carries the correct year and reconciles to the FY2027 guide (9-11% growth off the FY2026 base). This is the same mechanism already on file in [[pitfall-vendor-forward-eps-is-the-wrong-fiscal-year]] (TSM, QCOM, BSX, MDT, LEN, AMD, ANET) — ADP is simply another instance, not a new mechanism, so no new note is written this pass; worth adding to that note's ticker list on a future pass. - The stale fixed-dollar
Trim $280+was a false alarm. Spot breached it, but the corrected multiple-based trim (26x fwd on the true FY2027 EPS) sits at ≈$315 — comfortably above spot. This is the live "is the trim decision real" question the task flagged, and the answer is no, not yet — but only because the old dollar trim was miscalibrated, not because the stock hasn't moved. - A genuine insider buy (St5) that the baseline's "100% selling" framing missed simply because it hadn't happened yet.
6. Updated verdict
Rating: HOLD — conviction 7.0/10 (was 8.5)
ADP delivered on almost everything the March baseline underwrote: revenue and earnings grew in line with (FY2026) or ahead of (ROIC, deleveraging) plan, the moat's own stress-test (Rippling vs. bookings) came back in ADP's favor, and the balance-sheet concern the baseline flagged for monitoring is resolved. This is not a story that broke. But the price did the compounding the March report was underwriting — +41.5% against FY2026 fundamental growth of high-single-digits — and a chunk of the apparent March discount was a vendor forward-EPS artifact that, corrected, was never as extreme as it looked. The stock has moved from "cheapest in 15 years, buy zone" to "modestly below its own historical multiple band, fairly priced by a Street that has caught up."
Valuation
| Model | Read | Note |
|---|---|---|
| Graham IV | $61.12 | 5% weight — floor reference only, structurally undervalues asset-light names. |
| Bogle (5yr) | 9.2-15.2%/yr | Yield 2.39% + 9-10% EPS growth ± P/E reversion; no scenario here matches March's 13-22%. |
| DYT | $272 (2.5% band) – $309 (2.2% band) | Much narrower than March's $227-340 — most of the yield signal has already resolved. |
| DDM (2-stage) | $211-286 | Numerically close to unchanged — dividend didn't move. Spot has gone from the bottom of this range to the top. |
| P/FCF | Actual 23.0-24.0x vs 20-26x scenario band | Sits mid-band — fairly valued on cash flow, not a bargain, not stretched. |
Fair value range: Conservative $225-250 · Base $260-290 · Optimistic $300-325. Spot $284.68 sits at the top of the base case — a very different picture from March's spot-at-the-bottom-of-conservative.
Entry: $225-250 (top of conservative band; a real pullback, not the old unreachable $195-210). Trim: 26x fwd — at the true (guide-reconciled) FY2027 EPS ≈$12.10-12.26, this implies ≈$315-320, near the stock's own 5yr-average forward multiple. Re-set this judgment at the next pass; do not let it drift back to a fixed dollar.
Break triggers (next pass should re-test these first)
- PEO revenue ex-pass-throughs turns negative, or worksite-employee growth goes negative (St3 escalates from drift to break).
- Private-sector ADP payroll prints stay below ~50K for three-plus consecutive months and ES bookings guidance gets cut as a result (St7 finally reaching the fundamentals).
- The Nov 2026 dividend increase is skipped, delayed past the usual cycle, or comes in materially below the ~10-11% recent pace (St9 resolving badly).
- FY2027 EPS growth guided or delivered below the 9% floor at any quarterly update.
Upgrade conditions
- A pullback into the $225-250 entry band with the thesis otherwise intact — the fastest path back to BUY/ACCUMULATE.
- Confirmed FY2027 delivery at or above the high end of guide (11% EPS growth) with PEO margin stabilizing, which would support pushing the trim multiple up toward 27-28x on a second consecutive beat-and-raise.
7. What this pass did NOT test
- T6 (SBC discipline) — not re-pulled; carried forward untested.
- St6 (institutional 13F accumulation) — the cited Q3-2025 filings are now stale; needs a fresh pull before being cited as a bullish force again.
- Q3 FY2026 earnings (April 2026) were not independently re-read — superseded in substance by the full FY2026 print, but if a specific Q3-only data point is ever needed (e.g., a mid-year PEO trajectory check), it wasn't separately verified this pass.
- Peer relative valuation (PAYX, Workday) was not refreshed this pass — the baseline's PAYX/peer comparables were not re-pulled; the fair-value range above rests on ADP's own historical bands and DDM/DYT/FCF models, not a fresh peer table.
- This is ADP's first
/analyze-from-beforepass — there is no second-generation persistence data yet; the 92% figure above is a one-pass result, not a two-pass-confirmed one.
Baseline superseded: Output/Stocks/Technology/ADP/analyze-2026-03-30.md (2026-03-30).
Primary sources this pass: .mcp/fin.py ADP --news, yahoo-finance get_stock_info/get_recommendations/get_holder_info (insider_transactions), roic.ai per-share and yield-analysis endpoints, ADP Q4/FY2026 earnings coverage (BigGo Finance, Investing.com transcript), ADP National Employment Report (July 2026, released 2026-08-05).