---
ticker: "SHOP"
company: "Shopify Inc."
date: "2026-10-23"
view: "Avoid"
fair_value: 83
price: 168.5129
price_source: "market data quote; current snapshot"
price_time: "2026-10-23T14:43:32Z"
price_vs_fair_value: "+103.0%"
bear: 36
bull: 132
exit_price: 105
p_price_justified: "under 20%"
page: "https://alpha.dming.si/research/2026-10-23-shop"
---

# SHOP · Shopify Inc. · Avoid

Research report as of 2026-10-23. Page: https://alpha.dming.si/research/2026-10-23-shop

We value SHOP at $83 a share. The report's $168.51 price (Oct 23, 2026) was 103% above fair value: Avoid.

_Not restated: model.py's only reinvestment is the merchant loan book, not fixed or intangible capital, so the shared method's fade has no return on new capital to start from._

| Lens | Range | Marks |
| --- | --- | --- |
| Our report | $36 to $132 | fair $83 |
| Our simulation | $34.46 to $197 | median $82.98 |
| Street targets (20) | $130 to $210 | average $176, median $180 |
| Price | | $168.51 |

Bands run from our bear to bull, our simulation's 10th to 90th percentile and the Street's low to high. Street targets forecast future prices, not value today. Market data, captured Oct 23, 2026.

| Figure | Value |
| --- | --- |
| Fair value | $83 |
| Price | $168.51 (market data, Oct 23, 7:43 AM PT) |
| Five-year return | -5.8% (a year, at the price) |
| Bear / bull | $36 / $132 |
| Exit price | $105 (on Dec 31, 2031) |
| Confidence | 30% |
| Price against our scenarios | The price is justified in under 20% of our scenarios (5,000 simulated outcomes, each drawing the uncertain drivers from ranges we chose: our judgement, not a measured chance) |

**What the price needs.** Any one of these, on its own, would justify the price:

- 2026-31 revenue growth: needs 40.2%, we have 18.1%
- 2031 operating margin: needs 43.2%, we have 19.8%

_As of 2026-10-23 · model revalued at this date; unblinded_

- **Fair value today:** $83.18 per diluted share at a 9.74% cost of equity; the central operating forecast is carried forward from the 2026-10-15 underwrite and reviewed below
- **Price:** $168.5129, market data quote captured 2026-10-23 14:43:32 UTC → fair value is 50.6% below the price
- **Five-year exit price:** $105 on 2031-12-31: median of the judged exit multiple, long fade and outside view
- **Five-year return:** -5.8% a year from the captured quote through 2031, including cash paid out along the way
- **Purchase price:** $76: the highest price at which a buyer earns the 12% yearly return our call rule requires
- **Bear / bull today:** $36 / $132 from coherent operating cases
- **Price needs:** Alone, revenue growth of 40% a year 2026-31 (ours 18%) or a 43% 2031 margin (ours 19.8%); jointly, 25.6% a year with a 25% margin, 17% growth in 2032 and 28x (drivers.json, added 2026-10-23 after publication)
- **P(price justified):** 15% in 5,000 subjective draws at $168.51, a 10% market world included (reconciliation.md); not calibrated
- **Confidence:** 30% that the 2031 value lands within ±20% of $105
- **Key driver:** 2031 margin: resetting it to 12.7% lowers value to $58.96; the current Street revenue path raises value to $88.99
- **Central margin:** 19.8% after stock pay in 2031; an uncertain forecast, not a historical norm or company guide
- **Market target:** Street average $176.16 (median $180; range $130–210), captured 2026-10-23; an analyst price forecast, not validation of our operating model

**We retain Avoid at $168.51.** The current-date model replay gives fair value of
$83.18 per share, 50.6% below the quote. The operating forecast and valuation
methods are unchanged from the October 15 underwrite; there has been no new
Shopify filing since the Q2 2026 10-Q and release. We re-dated the model to
October 23, refreshed the market snapshot, reran the Street path through the
model and completed an independent opposite-side review.

The frozen October 15 report records $83.03 in its review table; replaying its
saved inputs through the current same-code model gives $83.050157. Both are
$83 at the report's whole-dollar headline precision. The comparable October 23
replay is $83.181799, a $0.131642 change from the shorter discount period
alone; the operating forecast, cash flows, and exit value are unchanged.

This revision is unblinded and cannot restore the original blind status. The
October 15 report was already partly unblinded: before its blind commit, the
Q2 2026 10-Q equity-investment note let us infer historical share prices. The
blind value and that exposure history are retained from the original report.

**Terms used below.** NOPAT is operating profit after tax and stock pay.
GMV is the value of merchants' sales. “Bear,” “central” and “bull” are
coherent downside, base and upside operating cases. “Street” means analyst
consensus, which is a comparison rather than company guidance.

## October 23 revision findings

This revision keeps the original central forecast because the review found no
arithmetic error and no dated evidence that supports replacing the forecast
with a different central margin. It also does not treat the forecast as
proven. The key unresolved judgment is whether operating leverage can raise
after-stock-pay operating margin to 19.8% in 2031. The available evidence
supports the direction of expense leverage; it does not establish that exact
2031 level. We disclose the uncertainty and show the model impact without
forcing a historical margin reset.

**Street comparison (flag 17).** The current market data snapshot is pinned in
`audit/current-market-snapshot.json`.
At 9.74%, the $176.16 average target discounts one year to $160.52; $83.18 is
0.518 times that amount. A target is not a company forecast, so we ran the vendor's
2027–2030 revenue and EPS estimates through the report's own operating model.
The cases below keep 2032 NOPAT growth at 13% and the exit multiple at 21.5x.
Only the EPS-derived cases change the operating-profit path; 2031 is an
explicit extension because market data supplies no estimate that far out.

| Case | Fair value | Change vs. $83.18 | 2031 margin after stock pay | Main assumption |
| --- | ---: | ---: | ---: | --- |
| Central forecast | $83.18 | — | 19.8% | Carried-forward model |
| Current Street revenue | $88.99 | +$5.81 | 18.9% | market data 2027–30 revenue; 2031 repeats 2029–30 revenue growth |
| Street revenue + EPS, pre-stock-pay crosswalk | $93.33 | +$10.15 | 19.8% | 2031 EPS grows 13%; subtract historical stock pay / revenue |
| Same, faster EPS extension | $100.24 | +$17.06 | 21.6% | 2031 EPS repeats 2029–30 growth of 21.3% |
| Street EPS left after stock pay | $110.43 / $117.34 | +$27.25 / +$34.16 | 23.7% / 25.5% | Same two EPS extensions; no stock-pay subtraction |

Both main EPS extensions remain below the current quote and discounted target.
The last row is an accounting-crosswalk sensitivity, not the preferred path.

Market data says its analyst EPS and net-income estimates are non-GAAP, but does not
itemize analysts' adjustments. We therefore do not state that every estimate
specifically excludes stock pay. The main crosswalk treats consensus EPS as
pre-stock-pay and subtracts stock pay at the historical 2025 ratio of 3.885%
of revenue; the alternative leaves stock pay in the EPS estimate. Under that
alternative, the same two extensions value the business at $110.43 and
$117.34. The crosswalk and all four scenario outputs are reproducible in
`audit/street-replay.py`. No 2031 consensus is
available. The 2031 growth extensions and stock-pay crosswalk are assumptions,
not observations. The independent advocate's saved-October-15 replay also
supported Avoid; it reached $98.34 using a separate revenue-to-GMV and cash
flow crosswalk. The difference from our saved-date actual-model replay is
small and comes from its simpler proportional-revenue treatment of segment
mix and loan funding; the current-date scenario above uses the report's
segment margins and `model.py` directly.

**Terminal margin (C6).** The central case moves from a 12.7% reported 2025
operating margin to 19.8% in 2031 by reducing operating expenses before
transaction and loan losses from 26.0% of revenue in 2027 to 21.5% in 2031;
losses stay at 3.8%. The 2025 margin is the highest reported since 2016, but
the data do not establish a precise cyclical boundary: 2021 margin was 5.8%,
and 2022–2023 were loss years. We do not label 2021 or 2025 as the required
“pre-upcycle” peak without evidence for that cycle date.

The Q1 2026 CFO said Shopify had kept headcount slightly down for three years,
did not expect that to change, and could continue lowering sales and
marketing, research and development, and general and administrative expense
as percentages of revenue. Q2 showed continued near-term leverage, and the
Q3 guide was 33–34% total operating expense as a share of revenue. These are
dated support for direction and near-term discipline, not a 2031 margin guide.
Holding every other input constant and changing only the 2031 margin gives
$58.96 if it is reset to the 2025 actual of 12.70%, and $35.58 at the 2021
actual of 5.83%. These are diagnostic sensitivities, not central estimates.
C6 remains an open forecast risk; neither sensitivity is adopted as the
central case. The existing Q3 and 2027 falsifiers remain the tests for whether
the operating-leverage path is succeeding.

**Call.** The revised evidence supports the same Avoid conclusion. Current
Street revenue alone adds about $6 per share to fair value; the EPS-based path
adds more, but depends on a non-GAAP-to-operating-profit crosswalk and an
unobserved 2031 extension. The current quote still exceeds the modeled bear,
central and bull values. This remains a valuation judgment, not a claim that
the company or its recent operating trend is weak.

## Carried-forward October 15 underwriting

The detailed company, segment, balance-sheet, forecast, case and original
review analyses below come from the published October 15 report. They are
retained because no new company filing changed those inputs. Any $178.08 price,
Street target or Street estimate below is the historical October 15 basis;
the current market snapshot and review findings above govern this revision.
The original report remains unchanged at its frozen published path.

## Business economics: Shopify keeps about 3% of merchants' sales and needs almost no fixed capital

- **Who pays.** Merchants, from one-person shops to General Motors and
  Balmain (Q2 2026 call). Merchant solutions earn 2.33% of GMV (2025), up
  from 2.21% (2023): payment processing and currency fees (Shopify Payments
  penetration 61% in 2024, 65% in 2025, 68% in Q2 2026, in 40 countries),
  lending interest and fees ($175M in 1H26), partner revenue shares,
  shipping labels, POS hardware and ads. Subscription revenue grows with
  merchants and plan mix (monthly recurring revenue, MRR, $221M, +19%; Plus is 34% of MRR).
- **Margins.** Gross margin 48.1% (2025) because payments carry card-network
  costs: merchant solutions 37.7%, subscription 81.1%. Operating expenses
  before losses fell from 35.7% of revenue (2024) to 31.8% (2025) and 31.5%
  (1H26) on flat headcount. Transaction and loan losses rose from 2.6% to
  3.6% and 3.8% of revenue as lending grew. Stock pay is inside GAAP
  operating expense (3.9% of revenue in 2025).
- **Capital.** The business needs almost no fixed capital (capex $26M in
  2025). The capital it uses is the loan and cash-advance book: $2,184M net
  at 2026-06-30, up from $1,784M in six months, 5.7% of gross balances more
  than 180 days late (10-Q note 6). Shopify buys these loans from a bank
  partner; that book is what grows with GMV.
- **Balance sheet.** No debt since the $920M convertible matured in 2025.
  Cash and marketable securities $5,472M; equity and other investments
  $4,854M (Affirm $1,655M, Global-E $768M, Klaviyo $280M, private stakes
  $1,555M, convertible notes $587M) and the Flexport stake $559M (10-Q
  note 4). These swing GAAP net income by hundreds of millions a quarter;
  we exclude them from earnings and add them once in the bridge.
- **What limits it.** Fees are set against Stripe, Adyen and card networks,
  and large merchants negotiate; agentic shopping (OpenAI, Google, Meta's
  Muse) may move discovery and checkout off merchants' own sites, which
  management says carries the same economics today (Q2 2026 call); AI costs
  sit in subscription gross margin; the Canada Revenue Agency (CRA) has proposed a
  transfer-pricing assessment for 2020, disputed (10-Q note 12).

## Shopify has beaten its own guide in every quarter since Q3 2024

Shopify guides growth in words ("low-thirties"). Taking word midpoints
(low 2, mid 5, high 8 in the decade named), it beat every quarter from Q3
2024 to Q2 2026 (printed by `model.py` from the releases):

| Q3 2024 to Q2 2026 | Beat | Median | Range |
| --- | --- | --- | --- |
| Revenue growth vs the guided word midpoint | 8 of 8 | +4.5 points | +2.0 to +6.0 |
| Operating expenses, % of revenue, vs guided midpoint | 8 of 8 lower | 1.5 points lower | 0.6 to 2.9 |
| Stock pay vs guided dollars | 8 of 8 lower | $10.5M lower | $6M to $23M |

Q2 2026: revenue +34% against "high-twenties"; gross profit +31% against
"mid-twenties"; opex 34.0% against 35-36%. Q3 2026 guide: revenue growth
"low-thirties", gross profit "mid-to-high twenties", opex 33-34%, stock pay
$150M, no significant FX. Our Q3 forecast is built bottom-up (GMV +31.5%,
take rate +10 bp, subscription +21%): revenue $3,787M (+33.2%), about a
quarter of the usual beat because the Q3 2025 comparison grew 32%.

## The forecast and valuation: $83 central, from a 21.5x exit multiple

`model.py`, output in `output.txt`; assumptions and
sources in `assumptions.md`.

| Central | 2026 | 2027 | 2028 | 2029 | 2030 | 2031 |
| --- | --- | --- | --- | --- | --- | --- |
| GMV ($B) | 496 | 615 | 738 | 863 | 993 | 1,122 |
| Merchant-solutions take rate | 2.42% | 2.46% | 2.49% | 2.52% | 2.54% | 2.56% |
| Revenue ($M) | 15,345 | 19,073 | 22,952 | 26,975 | 31,064 | 35,210 |
| Operating margin after stock pay | 15.3% | 16.3% | 17.5% | 18.5% | 19.2% | 19.8% |
| NOPAT ($M) | 1,183 (2H) | 2,483 | 3,206 | 3,984 | 4,782 | 5,586 |
| Owner free cash flow ($M) | 1,135 (2H) | 1,947 | 2,653 | 3,420 | 4,200 | 5,006 |
| Paid out, after operating cash ($M) | 1,073 (2H) | 1,835 | 2,536 | 3,299 | 4,077 | 4,881 |

- **Segments on their own paths.** GMV growth fades from 31% (2026) by 3-7
  points a year to 13% in 2031, from a company still accelerating (2023
  +20%, 2024 +24%, 2025 +29%) on a base management puts above 14% of US
  e-commerce. The take rate gains 2-4 bp a year as payments penetration
  rises from 68%. Subscription revenue grows 18% falling to 11%.
- **Margins.** Merchant gross margin 37.5% and subscription 79% from 2027;
  operating expenses before losses fall 1.25 points a year to 21.5% of
  revenue (they fell 4 points a year in 2024-26); losses stay at 3.8%.
- **Stock pay is a cost** (inside GAAP opex); the share count is held at
  1,294.5M (outstanding plus the 5.3M vested options; unvested options and
  RSUs are left out because their cost is in future stock pay), and
  buybacks absorb issuance. Counted once.
- **Reinvestment.** The loan book grows at 0.45% of GMV, charged against
  owner cash flow; capex equals depreciation.
- **Bridge.** Today: cash and securities $5,472M plus vested-option
  exercise cash $403M, less operating cash of 3% of the last twelve
  months' revenue ($398M), plus equity investments and Flexport ($5,413M)
  at June 30 carrying value: $8.4 a share at face. Each year's owner cash
  flow, less the growth in operating cash, is paid out at the year end
  (the $5B buyback authorization, 8-K 2026-06-02): $10.0 a share at
  9.74%. The exit value is the operating business alone.
- **Three methods, central.** Judged exit multiple 21.5x 2032 NOPAT
  ($105), set at the case's own long fade (21.1x, rounded up); long fade,
  growth 13% fading to 3% over 15 years and return on new capital 60%
  fading to 40% ($103); outside view, 31.9x the peers' latest trailing GAAP
  P/E on 2031 NOPAT ($138). The value-driver formula at 6% perpetual
  growth, 40% return and 9.74% gives 22.7x, printed beside the judged
  multiple; it prices one mature return for ever. The blind 28x implied
  6.8% growth for ever at a 40% return, above what the case's own fade
  pays for, and the reviews cut it (answer 4). The outside view sits 31%
  above the median because the peers' 32x is a market multiple of
  businesses whose prices we do not test; we believe the case's own
  multiple and fade.
- **Cost of capital.** Risk-free 5.079% (US 10-year, market data,
  2026-10-15), Damodaran's implied premium 4.20% (2026-10-01), peer median
  Blume beta 1.11 (V, MA, PYPL, MELI, AMZN, BKNG, INTU, EBAY): cost of equity
  9.74%. At Shopify's own Blume beta (2.02; raw 2.53) the cost of equity is
  13.57%, above the 12% hurdle (review flag 11, below). At that rate the central fade
  is worth 12.5x 2032 NOPAT, so the exit multiple falls with it: exit $61,
  fair value $49, purchase price $52 (review flag 7, below). Bear $25, bull $74.
- **Cases.** Bear ($36): GMV growth 17% falling to 8% ($0.85T in 2031), take
  rate slipping to 2.29% under fee pressure, merchant gross margin 35.5%,
  losses 4.3%, opex stuck near 26%, 15.5x at exit (its fade 15.3x). Bull
  ($132): GMV 28% to 16% ($1.32T), take rate 2.68%, margin to 22%, 26.5x
  (its fade 26.2x). Weighted 25/50/25 (our weights): $84. The bear's
  outside view ($69 at exit) is 92% above its median because the peers'
  32x does not fit a business growing 8%; we believe the bear's own
  multiple and fade ($36, $36).

Sensitivity of central fair value: NOPAT -20%/-10%/+10%/+20% gives $68,
$75, $91, $98. Exit multiple 20x gives $82 (the fade then sets the median),
24x $91; from 28x up the median moves to the outside view and fair value
stays near $103. Mature return on new capital from 25% to 100% moves the
fade between $99 and $108 and fair value between $83 and $85.

## Original October 15 price: 2.47 times our 2032 profit, or 53x it

From `unblinded.py` (output in
`unblinded-output.txt`):

- After today's cash and stakes ($8) and the payouts to 2031 ($10), the
  2031 exit must be $259 to earn the 9.74% cost of equity (our central
  $105); to earn 12%, $289. That needs 2032 NOPAT of $15.6B at our 21.5x
  (2.47 times our $6.3B), or 53x our 2032 NOPAT. Scaling every year's NOPAT
  by 2.24 (same multiples) reaches the price.
- The Street's revenue is close to ours: Street consensus (2026-10-15) has
  2026 at $15.06B (38 analysts; ours +1.9%), 2027 $18.98B (+0.5%), 2028
  $23.9B (-3.9%, 24 analysts), 2029 $28.6B (-5.8%, 22), 2030 $33.8B (-8.1%,
  14). The Street's EPS excludes stock pay (its 2025 figure, $1.46, is
  about operating income plus stock pay plus interest after tax, not GAAP
  $0.94); on that basis, with 3.5% earned on today's surplus cash, our
  earnings are +2% against the Street in 2027 and -14% in 2030 (7-11
  analysts). So the market is paying for a multiple far
  above any mature peer's, not for a faster business than the Street's.
- **Alpha's view** (`alpha rank SHOP --raw --json`, 2026-10-15): rank 294 of
  713, score 58.9, Neutral; quality A (rank 43), growth B, momentum C,
  expectations C, valuation D (rank 636). Alpha agrees the business is
  excellent and the price is high; our Avoid rests on value, where Alpha is
  near the bottom.
- **How the price got here** (daily closes). Q2 results on
  2026-08-05: $123.30 (08-04 close) to $144.24, +17.0% on the day against
  QQQ -0.9% ($723.85 to $717.30); $152.61 on day 5 (2026-08-11) against QQQ
  $718.45. The stock then fell to $139.82 (2026-09-01) and rose 27% to
  $178.08 by today. Headlines in that stretch (market data, read 2026-10-15)
  include Meta's Muse agentic checkout using Shopify (2026-09-21) and a
  Wedbush Outperform reiteration at $176 (2026-09-23); no company release.

## Original October 15 saved estimates: we agreed with the Street on revenue and disagreed on the multiple

The market's thesis: Shopify becomes the commerce layer under AI agents as
well as stores, keeps GMV growing near 30% for years, and deserves a
multiple no mature payments or commerce company now carries. Ours: the
operating story is strong and mostly in our numbers (they track the Street
to 2027), but a business that keeps 3% of GMV, pays card networks most of
its payments revenue and lends to its merchants is worth a network-like
multiple at most, and $178.08 needs either two and a half times our profit
or two and a half times that multiple. What settles it: whether GMV growth holds above 25% into 2028 and
operating expenses keep falling 3-4 points a year, as in 2024-26 (the bull's
path and beyond).

We have no edge on the near-term forecast; the company beats its own guide
every quarter and our Q3 numbers sit above the guide. Our edge, if any, is
discipline on the exit multiple and on charging the loan book and stock pay
as costs.

## Why the other side could be right

- Growth is accelerating, not fading: GMV +20%, +24%, +29%, +33% (2023 to
  1H26). If 2027-2031 growth stays near 25% instead of fading to 13%, 2031
  GMV is about $1.5T and, with the bull's 22% margin, 2031 profit is about
  1.45 times our central (our arithmetic: $1.5T at a 2.6% take plus $7B of
  subscriptions); with the margins in the next point it nears twice. At
  our 21.5x the price needs two and a half times; at the blind 28x, about
  twice.
- Opex fell 4 points of revenue a year in 2024-26 on flat headcount. If AI
  keeps headcount flat while revenue doubles, operating margin could reach
  25-30% by 2031, not our 19.8%.
- Agentic commerce could raise Shopify's share of the economy's GMV, and the
  Shop Pay identity network could earn more per transaction than we allow.
- But the bull case already carries 16% growth in 2031, a 22% margin and
  26.5x, and is worth $132. The price sits above our whole bear-to-bull
  range: the market is paying for growth beyond our bull and a multiple
  above every mature peer's, together. A multiple alone does not close the
  gap: from 28x up the median moves to the outside view and fair value
  stays near $103.
- The advocate's strongest set of changes it could defend (merchant
  margin held at 38.2%, faster take-rate gains, GMV halfway to the bull,
  the bridge fixes) gives $107 and a $97 purchase price on the blind 28x;
  with everything generous, $123. It could not make the case for Buy.

## Catalysts and falsifiers

- **2026-11-03 (company events page):** Q3 2026 results. Our forecasts:
  revenue $3,787M (+33.2%; guide low-thirties), GMV $121.0B (+31.5%), gross
  profit $1,805M (+29.8%; guide mid-to-high twenties), opex 32.7% of revenue
  (guide 33-34%), operating income $566M.
- **February 2027 (not confirmed):** Q4 2026 results and the first 2027
  outlook.
- **Falsifier for Avoid:** 2027 GMV growth of 28% or more with opex before
  losses at or below 25% of revenue, or a 2027 operating margin of 20% or
  more; re-underwrite then.
- **Toward the bear:** take rate below 2.35% of GMV, losses above 4.5% of
  revenue for two quarters, or 2027 GMV growth below 18%.

## Original October 15 review flags: six raised, each answered

The flags are numbered patterns in our review checklist (review.md).

- Flag 1 (medium; net move from the blind value): the value moved -$14.5 (-14.9%) from the blind $98 after
  the price was read; the largest change is the exit multiple (-$19.5),
  resting on the blind model's own fade, not on a management claim or a
  Street figure.
- Flag 3 (medium; value close to or far from the price): fair value 53% below the price (40% or more away).
- Flag 4 (low; outside view far from the median): the central outside view is 31% above its median and the
  bear's 92% (answered above and in answer 10).
- Flag 6 (medium; a lenient input chosen over a harsher sourced one; raised by
  the auditor, we had said it did not apply):
  the 28x was lenient against the case's own fade and is cut (answer 4);
  the beta is the rule's peer median, with the harsher own beta printed
  beside it (answer 9).
- Flag 7 (low; fair value moving a lot between own and peer beta): the purchase price is $76 at the peer beta and $52 at the
  own beta (-32%), because the own-beta cost of equity also lowers the
  exit multiple through the fade (answer 11).
- Flag 11 (low; cost of equity above the 12% hurdle): at the own beta only (13.57%).
- Other flags checked and not raised: 2 (view unchanged), 5 (central within 2-8% of
  the Street on revenue through 2030 and within 14% on earnings before
  stock pay), 8 (outstanding shares
  and vested options in the count; unvested awards are costed in stock pay;
  no convertible, no at-the-market program), 9 (10-K, two 10-Qs, eight
  releases, two calls read), 10 (today's batch of 15 records has 10 below
  and 5 above the price), 12 (the reviews moved the central value 14.9%),
  13 (equity investments and Flexport carry filing lines in 10-Q note 4).

## Original October 15 adversarial review: neither reviewer moved the view off Avoid

Two reviewers (model auditor, opposite-side advocate), one round. Neither
moved the view off Avoid. Rows 1-11 are the auditor's findings and rows A1-A8
the advocate's. Effects are measured on the final model
(`.tmp` scratch runs that patch one input of `model.py`), except where a
row says "blind model". Running order of the accepted changes: blind
$97.54 → bridge (1, 2, A2/A3) $102.15 → share count (A1) $102.56 →
multiples (4) $83.03.

| # | Finding | Flag | Effect | Answer |
| --- | --- | --- | --- | --- |
| 1 | The $5.4B of equity investments earn nothing for 5.2 years and are discounted at 9.74%, so they count at about 62% of face | 13 | With 2: blind model $97.54 → $102.15 (+$4.6) | Fixed. Carrying value at 2026-06-30 (10-Q note 4), added today at face. The stakes are marked to market, so holding them flat and discounting them charges a return they are expected to earn |
| 2 | Kept cash earns 3.5% pre-tax and is discounted at 9.74%; the outside view prices that interest at 31.9x | none | In 1 | Fixed. Today's surplus cash counts at face; each year's owner cash flow, less the growth in operating cash, is paid out at the year end. Evidence: the $5B repurchase authorization (8-K 2026-06-02) and $1,933M used in 1H26 (10-Q note 10). Method 3 now prices 2031 NOPAT, with no interest in it |
| 3 | Cash rate 3.5% against about 4% implied by 1H26 interest income | none | $0 now: no cash is held in the model after 2 | Moot after 2. The 3.5% remains only in the Street comparison in `unblinded.py` |
| 4 | The 28x exit is 25% above the case's own fade (21-22.5x for a mature return of 40-140%); the 31.9x peer P/E comes from slower-growing peers. What dated, price-independent evidence supports 28x? | 6 | Central 28x → 21.5x: $102.56 → $83.03 (-$19.5); bear 18x → 15.5x: $39.84 → $36.27; bull 34x → 26.5x: $135.44 → $132.18 | Fixed. We have none. The blind reason for 28x (`assumptions.md`) was Visa and Mastercard at 33-35x trailing GAAP earnings and Amazon and Booking at 32x, set a little below them. Those are today's market multiples of other companies, the evidence method 3 already uses; they say what the market pays now, not what our forecast of Shopify is worth at exit. Against 28x, all in the blind commit: our own fade on the same forecast ($126 at exit against the multiple's $159, which is 21.1x the same 2032 NOPAT), and the blind output's line that 28x needs 6.8% growth for ever at a 40% return while the case fades to 3%. Each case's multiple now sits at its own fade, rounded up to the half: bear 15.5x (fade 15.3x), central 21.5x (21.1x), bull 26.5x (26.2x). Same error and same fix as in our ServiceNow and Salesforce reports (lessons.md, 2026-10-15) |
| 5 | Tax: 1H26 cash tax $241M, about 24% of operating income plus interest, against 20% assumed; but 2025 cash tax $194M and effective rate 18.4% | none | 24%: -$3.8 ($83.03 → $79.20) | Disclosed, not changed. The two dated figures point both ways (10-K FY2025 note 21: 18.4% effective, $194M cash; 10-Q Q2 2026: $241M cash paid in 1H26 against $88M in 1H25, a half year whose timing we cannot separate from the rate). 20% sits between them; the CRA dispute stays a named, unmodelled risk |
| 6 | GMV slows from 31% to 24% in one year while the Street's revenue slows less (about +26% for 2027) | 5 | +2 points a year 2027-31: +$2.9 ($85.89) | Rejected. The Street is not evidence that does not depend on the price, and the advocate's own check (North American GMV +28% against US e-commerce +12.2%, Census Q2 2026 via secondary coverage) puts Shopify's share of new market dollars at 28-32%, which supports the fade. Our 2027 revenue is already 0.5% above the Street's |
| 7 | Shopify beat its revenue guide by a median 4.5 points in each of the last eight quarters; the model takes about a quarter of that | none | Full beat in H2 2026 (GMV growth +3.3 points): Q3 revenue $3,861M, fair value +$0.5 | Disclosed, not changed. The record was in the blind commit and weighed there: the Q3 2025 comparison grew 32%. The Q3 forecasts in the record stay as committed blind; Q3 on 2026-11-03 tests them |
| 8 | Take rate +1 bp a year | none | +$0.7 | Disclosed, not changed: see A5 |
| 9 | Own beta 2.02 | 6, 11 | $83 → $49 (-$34) with the exit multiple at the fade for that rate (answer 11) | Disclosed, as before. Our valuation method's rule is the Blume peer median (a beta pulled toward 1); the own beta is printed beside it |
| 10 | Flag 4: bear's outside view far above its median | 4 | $0 | Explained: the peers' 32x does not fit a business growing 8%. After 4 the central outside view is also 31% above its median, for the same reason |
| 11 | Auditor's rerun: the own-beta line kept the 21.5x exit, but the fade at the 13.57% own-beta cost of equity is 12.5x; the report understated the own-beta downside by about $21 | 7 | Own-beta central: exit $105 → $61, fair value $71 → $49, purchase $76 → $52; bear $25, bull $74 at own beta. Headline (peer beta) $0 | Fixed. `model.py` now sets each case's exit multiple at its own fade at the own-beta rate (bear 9.5x, central 12.5x, bull 15.5x), the same rule as answer 4. Flag 7 is now raised |
| A1 | Drop unvested RSUs and options from the share count (our own share-count check) | 8 | Shares 1,306.6M → 1,294.5M with vested-option exercise cash $403M (5.31M at $75.93, 10-Q note 10): +$0.4 after 1-2 | Fixed. Their future cost is in the stock pay we expense; counting the shares too charged it twice. Vested options stay in, with their exercise cash |
| A2/A3 | Cash and stakes at face, future cash paid out; against it, buybacks at $178 against a value near $107 destroy value | none | In 1 | Fixed with 1-2. A buyback above value moves wealth from holders who stay to holders who sell; the per-share value of the cash paid out is unchanged for the business as a whole, so it is not charged |
| A4 | Merchant-solutions gross margin: keep 38.2% instead of 37.5% (38.9% 2023, 39.1% 2024, 37.7% 2025, 38.7% 1H26; CFO 2026-08-05 "up slightly") | none | +$2.2 ($85.23) | Rejected. Every figure was in the blind commit (10-K, 10-Q, Q2 call transcript) and weighed there: a first half runs above the full year because Q4, the largest quarter, is the lowest (36.8% in Q4 2025), and 37.5% is near 2025's full-year 37.7%. Our inference, not sourced: payments' rising share keeps pressure on the segment margin. No new evidence |
| A5 | Take rate +5,+5,+4,+4,+3 bp a year against our +4,+3,+3,+2,+2; penetration 61/65/68%, take rate up about 10 bp a year in 2025-26 | none | +$1.0 ($84.00) | Rejected, same reason: blind evidence already weighed. Our inference: penetration gains slow as it nears its ceiling, and large merchants negotiate fees |
| A6 | GMV halfway to the bull (26% falling to 14.5%) and 30.25x | none | At the case's own fade (23.0x): +$8.6 ($91.59); at 30.25x: +$23.6 | Rejected. Mostly management claims; the advocate's own Census check supports our fade (see 6) |
| A7 | Opex to 20% of revenue by 2031 | none | +$5.6 ($88.63) | Rejected; the advocate also called it unsupported. Our opex grows 0.62-0.72 times revenue, close to 1H26's 0.66 |
| A8 | The outside view caps the median whenever the forecast rises; dropping PayPal and eBay gives 32.8x | none | $0 now (outside view $141, not the median) | Rejected. The peer set was fixed before the blind commit; dropping peers after the price needs a reason other than the result |

Flag 6 in full: we had written that flag 6 did not apply because the 28x
was below the peers' 32x and the networks' 33-35x. The auditor was right that the harsher sourced
alternative was our own fade, not the peers, and the multiple is cut. The
beta is not a lenient choice under the rule: the method fixes the Blume peer
median and prints the own beta beside it, here $49 against $83.

## Sources and date basis

- Company financials: SEC EDGAR, CIK 1594805; FY2025 10-K filed 2026-02-11; Q1 and Q2 2026 10-Qs filed 2026-05-05 and 2026-08-05; Q2 2026 results and the Q3 guide filed 2026-08-05. The Q2 10-Q is available at [SEC EDGAR](https://www.sec.gov/Archives/edgar/data/1594805/000159480526000047/shop-20260630.htm), and the [company's Q2 2026 results](https://www.shopify.com/investors/press-releases/shopify-delivers-big-30-growth-across-gmv-revenue-gross-profit) include the Q3 guide.
- Call evidence: Shopify Q1 2026 and Q2 2026 transcripts dated 2026-05-05 and 2026-08-05, retained under `data/`; the CFO statements are cited above with their dates.
- Original underwriting inputs: market data filings, statements, peer ratios, estimate capture and returns, plus company filings and calls, pinned by `data-manifest.json`; most vendor and company captures are dated 2026-10-15.
- Current market snapshot: market data quote, price-target-consensus, and annual analyst-estimates endpoints captured together at 2026-10-23T14:43:32.628612Z; the exact inputs are in `audit/current-market-snapshot.json`.
- Market data says Analyst Estimates EPS and net-income figures are non-GAAP; it does not itemize individual analysts' adjustments. See the Street's analyst-estimates FAQ.
- The current package was checked separately from this revision: package `20261023T130617793764Z-e02ccc2c8451`, build `54e4a1eea4cdcd425aa94df0fb227db1732d7286`; it remained published and verified and selected the October 15 report. This dated revision is not published.

## Validation and review status

- `uv run python model.py` and `uv run python unblinded.py` were rerun in this folder; saved output is in `output.txt` and `unblinded-output.txt`.
- `uv run python audit/street-replay.py` and `uv run python audit/margin-sensitivity.py` reproduce the Street cases and margin diagnostics from the pinned inputs.
- `uv run python scripts/research-events.py docs/research/2026-10-15-shop` found no 8-K or 6-K after the 2026-08-05 Q2 filing; no company filing arrived between the source capture and this revision.
- `uv run python scripts/research-check.py docs/research/2026-10-23-shop` passed the record, source pins, model and unblinded replays, decision block, and required next-results preview. Current checker flags 1 (medium), 3 (medium), and 17 (high); flag 17 is answered above and in the audit. Required review depth was met by the independent model audit and opposite-side advocate. The auditor's final review after this revision is recorded in `audit/review.md`.
- The report's earlier review table remains the 2026-10-15 baseline record. Its accepted repairs still replay correctly; this revision does not reopen them. C6 remains an open forecast judgment as stated above, not a passed check.

## Revision history

| Report | View | Central fair value | Price basis | Change |
| --- | --- | ---: | --- | --- |
| [`2026-10-15-shop`](https://alpha.dming.si/research/2026-10-15-shop) | Avoid | $83.05 model replay ($83 reported) | $178.08, market data, Oct. 15 | Original report |
| `2026-10-23-shop` | Avoid | $83.18 ($83 reported) | $168.5129, market data, Oct. 23 | +$0.13 from re-dating alone; forecast unchanged |

The report's source and figure basis is explicit: all current market values are from the pinned October 23 snapshot; the operating model remains grounded in the latest company filing available at the October 15 source cut. No trade, portfolio, model-setting or publishing action was taken.