Pattern Group (PTRN): The 47% Growth Is Real. The "Loss" Is a Footnote. The Multiple Is Not Cheap.

On September 10, Pattern Group CEO David Wright walked into the Goldman Sachs Communacopia tech conference and described his company as a "fast-growing e-commerce infrastructure company". The stock had already done what the market does with a good story: it was up roughly 72% for the year, trading near $19.86 against a 52-week range of $8.92 to $29.80. But the last five trading sessions told a quieter story — down about 12% over the past 20 days, sitting below its 50-day average with the momentum indicator (RSI) at 39. A stock that ran this hard and then stalls is a stock that needs to be re-read, not re-liked.

The most useful thing to do right now is to separate three things the market is muddling together: a business that is growing 47% a year, an accounting artifact that makes it look like a money-loser, and a price that has already priced in a lot of good news. Strip the distortion out and you get a cleaner version of the question than the transcript's cheerleading — or a screener's scary red "negative P/E" — either lets you see.

The machine, not the metaphor

Start with what Pattern actually sells, because "e-commerce infrastructure" is doing more work here than the business does. Pattern is an online reseller. It buys inventory from consumer brands — Pandora, TUMI, P&G, Nestlé brands, Bosch — and sells it on Amazon, Walmart, and a growing list of international marketplaces. It is, in its own S-1 words, "an online reseller of consumer products" that partners with premium brands and sells through marketplaces.

What separates it from a boring reseller is the cost side. Management frames the whole company around one equation — traffic × conversion × price × availability — and around a "cost-to-serve" number: how much it costs to push one dollar of revenue through the machine. The concrete version of that claim is logistics. Pattern says 95% of its inbound volume moves in full truckloads at $0.11 per unit, versus $0.44 per unit for the parcel and partial loads a brand running it alone would typically use. That is a four-fold cost gap on the single biggest input in the business. You can take the $0.11 as management's own number, but the structural point stands: at scale, it is genuinely hard for an individual brand to replicate that truckload efficiency across the 14–18 fulfillment centers it says must ship every week.

The stickiness metric is the one to watch, and it is the strongest number in the whole report. Net revenue retention — how much the revenue from existing customers grew — hit a record 129%, up from 118% a year earlier. In plain terms: Pattern's existing brands bought about 29% more from it than the year before, before a single new brand was added. That is not a weak business. A 129% NRR is what makes a company worth a premium, not what makes it a value stock. And the mix is slowly moving off Amazon — from 94% of revenue at the IPO to about 90.6% in the latest quarter, with Walmart up threefold in a year and international (Coupang in Korea, MercadoLibre in Mexico) crossing a record $110 million.

A "loss" that is a footnote

Now the distortion, and this is the part that matters most for how you read the price. If you look up Pattern on a screener you will see a negative trailing P/E, roughly negative 27x. That reads like a company that loses money. It does not lose money. The negative number exists because of a one-time $92 million stock-based compensation charge tied to the IPO that landed in the third quarter of 2025. That single charge knocked that quarter to a reported $2.19 loss per share, and it is what makes the trailing P/E, the trailing EBITDA multiple, and a negative return-on-equity all look broken.

Set that one-time item aside and the earnings picture is the opposite of a loss. Each of the last three reported quarters produced positive GAAP net income — $27 million in the June quarter, $29 million in December — and the full year is expected to be profitable on a GAAP basis. Trailing free cash flow is about $106 million, up 92% a year. Operating cash flow is $136 million. There is $346 million of cash, a $150 million undrawn revolver, and no interest-bearing debt. Management has a $100 million buyback program on the books from March. A company with that balance sheet is not the company a red "negative P/E" describes.

So what are you actually paying? Against a market cap near $3.5 billion and an enterprise value near $3.2 billion: a forward consensus of about $0.57 per share in 2026 works out to roughly a 35x multiple on next year's profit; the 52x trailing EV/EBITDA is the inflated artifact of that one-time charge; and free cash flow of ~$106 million against a $3.2 billion enterprise value is a FCF yield of about 3.3%. That last one is the honest anchor for a retail investor — you are paying about thirty times cash flow, and you get the growth back as an option, not as a guarantee.

The factor stack, and where the risk actually lives

Here is the full read, factor by factor, against the numbers rather than the narrative.

Growth is the obvious A: revenue up 47% in the quarter, roughly 44% for the trailing year, and four straight quarters above 40%. Profitability is the weak line — not broken, but thin. Adjusted EBITDA is up 54% and the margin crept to 6.1% from 5.9%, but 6.1% is still a thin margin for a company that buys and ships physical goods. Cash flow is the quiet A: free cash flow up 92%, net debt zero. Safety is the A you rarely get on a growth stock: no debt, $346 million cash, a current ratio above 2.3.

Momentum is the factor that changed, and it changed against the stock. Up 72% for the year, but down about 12% in the last 20 days, below the 50-day line at ~$23.5 with an RSI at 39. A stock doing that is not being rejected because the business broke — it is being repriced because the easy part of the run is done and the multiple needs the story to keep improving.

The real risk is not the multiple. It is concentration. About 90% of revenue still flows through Amazon, and Amazon is both the customer and the venue that can change its rules at any time. The diversification is real — Walmart tripling, international up 87% — but it is 9% of the mix, not 90%. That is the difference between "managing the risk" and "the risk is managed." Add a business that depends on a single marketplace to a 6.1% EBITDA margin, and you have a company whose premium is only as safe as the assumption that Amazon keeps being profitable to sell through.

For a comparison, and with the honest caveat that Pattern has no clean peer set — it is a reseller with a tech platform, and none of these names is a true match — the e-commerce market shows what the premium buys. Shopify (SHOP) trades at a ~12x sales multiple; Mercadolibre (MELI) at ~2.7x; The Trade Desk (TTD) at ~2.2x sales with a 6.8x EBITDA multiple and a 16x P/E. Pattern's ~1.17x sales is actually below those on revenue, but its 52x EBITDA — even before you correct the one-time charge — is the point: the market is not paying for today's margin, it is paying for the margin it believes is coming. That is a growth bet with a reseller's cost structure underneath it, and the whole case lives or dies on whether 6.1% of EBITDA margin has a real path to 8, 10, or more.

What the stack says to do

The factor stack points one way: this is a name for a growth sleeve, not a value sleeve or a dividend sleeve. The improving report card — NRR at 129%, FCF up 92%, zero debt — is the part that justifies holding or adding on weakness, and "hold is not sell" cuts both ways: a pullback from a 52-week high is not a sell signal, and a run-up is not a buy signal either. The honest posture for a retail investor is the barbell instinct — if you want Pattern's growth, it belongs next to a boring cash-flow name that hedges the day Amazon changes the rules, not as a stand-alone conviction.

The specific trigger that would change this call is the margin, not the price. EBITDA margin that clears ~8% and keeps going is the confirmation that the multiple is buying a real improvement, not a hope. The trigger that would crack it is NRR slipping below ~120%, or non-Amazon revenue stalling while Amazon share creeps back up. Until one of those happens, the 47% growth is real, the "loss" is a footnote, and the multiple is what it is — about 35x forward and 52x on the distorted trailing number. You can respect the growth and still not chase a cooling price at a premium multiple. That is the whole job.