On’s margin upgrade puts the return on expansion in focus

On’s November 10 guidance lifts 2026 gross margin excluding tariff refunds from at least 65% to around 66%. Comparable adjusted EBITDA margin remains 19.5–20%. Third-quarter release

That combination makes the next stage of the company’s story more interesting than a headline about higher margins. Our interpretation is that the important question now concerns the passage from product-level economics to earnings after the costs of expansion. Better economics on each sale can support a stronger business, but the way management spends that benefit determines how much ultimately reaches shareholders.

Keep the refund separate

A refund can return real cash and correct an earlier cost. Excluding it from a recurring earnings assessment does not make it worthless. It prevents a payment associated with past imports from being assumed to recur on every future pair of shoes. A useful model needs separate entries for the recovery and the economics of selling the next product.

There is also a reason to avoid attributing all margin strength to that recovery. On’s second-quarter gross margin was already 65.4%, without tariff refunds. August results This earlier observation gives the underlying business a relevant comparison, although a single quarter cannot establish a permanent level. Product mix, seasonality and the timing of costs can still change the outcome.

The annual guidance deserves particular care. A former minimum and a new approximate target are different kinds of statement. Subtracting them does not reveal an exact amount of incremental spending. Nor does an unchanged profit-margin range establish that operating performance has deteriorated. It tells readers to ask for the bridge between gross profit, operating costs and the final result, rather than fill that bridge with assumptions.

Direct selling moves costs as well as revenue

On’s annual report places payment-processing, selling and distribution expenses below gross profit. 2025 annual report That accounting structure matters when judging a brand’s channel choices.

Consider a simplified comparison. Selling through a retailer leaves part of the consumer’s payment with that retailer. Selling directly lets the brand capture more of the transaction, while taking responsibility for more of the shopping experience. A higher gross margin can therefore coexist with greater spending on stores, customer acquisition, delivery or service. This is an analytical example, not an estimate of On’s channel profitability.

The right comparison is contribution after those costs over a customer’s relationship with the brand. Repeat purchases can spread acquisition spending over more orders. Conversely, a customer who buys once during a costly campaign may add revenue without producing an attractive return. Neither a beautiful flagship store nor rapid online sales growth answers that question alone.

For future disclosures, useful evidence would include productivity at established stores, repeat purchasing and fulfillment costs per order. The absence of a detailed channel contribution measure leaves uncertainty. Readers should retain that uncertainty rather than substitute gross margin for an unavailable answer.

Expansion must earn its place

At September’s investor day, On targeted an adjusted EBITDA margin of at least 22% by 2029 and identified football and golf as expansion areas. Strategy and targets These are management ambitions, not secured outcomes.

Entering an adjacent sport creates two possible paths. Existing design capabilities and customer recognition might make the next category cheaper to build. Alternatively, each new audience might require its own development, sponsorship and distribution budget. The first path can spread existing costs. The second can keep resetting the investment bill even as the company gets larger.

The distinction calls for milestones rather than an immediate verdict. A launch can reasonably absorb money before meaningful revenue arrives. Over time, however, a durable category should attract repeat demand without requiring promotional support to rise at the same pace. A useful review would compare cohorts of launches, separating initial setup costs from the expenses needed to sustain each business.

There is a similar test for physical expansion. Comparing a newly opened store with a mature one can unfairly penalize investment. Combining them indefinitely can conceal weak economics. Tracking comparable groups over their maturation period would make both mistakes less likely.

Ask what the next franc produces

Three questions follow. Does the next increment of spending generate additional demand, or mainly shift purchases between channels? Does that demand persist without heavier discounting? And does the resulting profit become cash after inventory and investment needs?

These are proposed tests, not claims that On is failing them. A growing company can rationally accept a temporarily flat operating margin to build a more valuable business. The case becomes stronger when management can connect spending today with repeat customers, improving productivity and cash generation later. It becomes weaker when expansion is supported only by broader reach and launch publicity.

The cash test also needs more than a quarter-end snapshot. Building stock ahead of a launch can absorb cash before sales arrive; reducing purchases can temporarily release it. The useful question is whether inventory turns into customer receipts on a reasonable timetable without creating a later need for discounts. Looking across several periods helps distinguish preparation for growth from an accumulation of unsold goods.

There is no need to force these different questions into one score. Demand can be healthy while a store network remains expensive, and a useful investment can initially depress cash flow. Keeping separate records of customer behavior, operating costs and cash commitments makes it easier to identify what actually improved and what still needs time.

The present evidence supports examining that investment process, rather than treating every gross-margin improvement as immediately distributable profit. For On, the more useful measure of premium positioning is how consistently it finances worthwhile growth after the full cost of serving the customer. That can be tested over time without turning one earnings release into a prediction about the share price.