As of November 8, 2026, Reformation's third-quarter call has given ecommerce teams a rare sentence to work with: the apparel brand said a new AI-based planning engine improved its ability to evaluate product performance and forecast demand for recut orders, which in turn supported in-stock rates and full-price selling. Net revenue rose 19.1 percent to 173.6 million dollars, direct-to-consumer revenue rose 16.7 percent to 152.8 million, and active customers reached 1.25 million, up 23.4 percent. For anyone who runs a product page, the interesting claim is not the growth figure. It is that availability is now being reported as an outcome of planning software rather than as a warehouse statistic.
What the earnings call actually claimed
The mechanism described on the call is narrower than the headline suggests, and that is what makes it useful. The engine is applied to recut orders, meaning the brand identifies a style that is selling, then decides how much more to produce and how fast. The example given was a cropped pant identified in May as a top performer, recut in 21 days, then extended into seven additional silhouettes. Availability improves here not because the warehouse got bigger but because the decision to replenish a winning style got shorter.
That is a different lever from the one most availability programmes pull. Traditional in-stock work tries to forecast demand further out and hold more buffer. A recut engine shortens the loop instead: sell some, read the signal, produce more within weeks. The buffer sits in the supply chain's reaction time rather than in inventory, which matters for a brand selling at a 60.4 percent gross margin where markdowns are the expensive failure mode.
Availability is a merchandising output, not an inventory metric
Once availability is produced by a planning decision, the metric stops belonging to the warehouse team alone. A product page that shows a full size run is the visible end of a decision made weeks earlier about how much to cut. When the page shows gaps, the cause may sit in the buy, the recut trigger, or the replenishment cadence rather than in the stock file. Treating an out-of-stock page as a logistics incident sends the fix to the wrong team.
What the quarter shows beyond the tool
The rest of the numbers describe a business growing by reach rather than by price. Direct-to-consumer revenue rose 16.7 percent while wholesale and other revenue rose 40.8 percent to 20.8 million dollars, and the store count moved from 70 to 77 with 80 expected by year end. Active customers rose 23.4 percent, but revenue per customer declined, which means the growth came from bringing more people in rather than from charging existing shoppers more. Adjusted EBITDA rose 64.2 percent to 23.1 million dollars, and the full-year revenue guide was raised to a range of 610 to 612 million dollars.
| Signal | What it says about availability | Where to look on the product page |
|---|---|---|
| Active customers up 23.4 percent | More first-time visitors meeting the catalogue | Size availability on entry styles |
| Revenue per customer down | Growth is reach-led, so a missing size costs a new shopper | Sold-out flags on best sellers |
| Wholesale up 40.8 percent | Partner orders depend on the same recut decisions | Case-pack and lead-time notes |
| Gross margin 60.4 percent | Markdown risk is the cost of overstock, not of stockouts | Discount depth on slow lines |
Read together, the signals say something simple: when growth comes from new customers, the first product page is the whole relationship. A new shopper who lands on a style with two sizes missing does not wait for a recut; she leaves. That is why in-stock rate behaves like an acquisition metric in this model, even though it is reported as an operations number.
Growth came from reach, not from price
The distinction matters for how the availability claim should be read. A brand raising prices can improve full-price selling without changing availability at all. A brand growing through reach has to convert first-time visitors, and first-time visitors are the least tolerant of a gap. The call's pairing of in-stock rates with full-price selling is consistent with that reading, but it does not prove it.
What the numbers do not prove
The reporting on the tool does not establish how much of the 19.1 percent revenue growth came from better availability rather than from pricing, new stores, wholesale expansion or category demand. The tool and the result were disclosed together, which is not the same as a causal link. Category growth in apparel has been uneven this year, and a wholesale channel rising 40.8 percent can move total revenue without any change to a single product page.
What the evidence does support is a management conclusion: availability is being planned, not just measured, and the planning loop is short enough to be judged inside a quarter. That is testable, and it is the part other ecommerce teams can copy without buying the same software.
How to audit an availability claim
Five fields, captured weekly, are enough to tell a planning improvement apart from a stocking decision. The distinction is worth the effort because the two have opposite costs: planning errors show up as empty pages during demand, stocking errors show up as markdowns after it.
FAQ
Did the AI planning engine cause the 19.1 percent revenue growth?
The reporting does not establish that. The tool and the result were disclosed together, but pricing, new stores and wholesale expansion also moved.
What does a recut order mean?
Producing more of a style that is already selling, usually in weeks rather than seasons. It shortens the replenishment loop instead of adding buffer stock.
Why track in-stock rate with discount depth?
Because availability can be bought with inventory. Without the discount figure, a higher in-stock rate may only mean more stock was held and marked down later.
Is this relevant outside apparel?
Yes, wherever a short production or replenishment loop exists. Categories with long lead times get less from recut logic and more from forecast accuracy.
