When the Recommended Size Is Out of Stock: Keep Fit Advice Honest

A shopper completes a foot scan, receives a size recommendation, and discovers that the recommended size is sold out. The next screen matters as much as the recommendation itself.

It is tempting to promote the nearest available size. But availability does not make a shoe a better fit. A useful journey keeps two questions separate: which option suits this person, and which suitable options can the retailer actually fulfill?

For ecommerce and product teams, this is a concrete application of AI Fit Intelligence: connecting personal fit context with a buying decision without allowing inventory pressure to rewrite the advice.

Separate the fit decision from the stock decision

Perfitt’s approach connects foot characteristics, inner-shoe information and fit data. Its published product overview describes personalized recommendations that can reflect snug, regular or roomy preferences.

Those inputs answer a different question from a stock feed. A recommendation is about the relationship between a person and a particular product; inventory tells the storefront whether a specific variant is available.

The workflow below is an implementation proposal for retail teams, not an announcement of a new Perfitt feature. Live stock checks, cross-product alternatives, replenishment alerts and analytics joins require scoping and validation with the retailer’s systems. Do not assume they come enabled with a sizing integration.

Design the unavailable-size moment explicitly

Consider a hypothetical shopper whose recommended size is unavailable. The interface should preserve that recommendation and state the stock limitation plainly. “Your recommended size is currently unavailable” is more useful than silently moving the recommendation to a size that happens to be on the shelf.

Offer another size only when fit evidence supports it

If the recommendation system identifies another acceptable option, explain the expected difference in fit and let the shopper decide. A roomier option is not identical to the first choice. If there is no supported alternative, do not manufacture one to keep an add-to-cart button active.

Evaluate another product on its own fit

A similar-looking shoe can be a useful discovery option, but copying the same size label across models is not a fit assessment. Re-evaluate the candidate product using the available product-level evidence. Then check that variant’s stock, intended use, price and delivery options. When product coverage is insufficient, present the item as something to explore—not as a verified fit match.

Make waiting a legitimate outcome

A restock alert or store-availability check can be more helpful than an unsuitable substitute. Show a replenishment date only when the retailer has a dependable date to show. Keep any alert opt-in explicit, and distinguish the requested availability notification from broader marketing enrollment.

Define the data handoff before designing the screen

A team reviewing this journey should agree on a small set of responsibilities before connecting systems:

• Product identity: map the recommended model and size to the exact sellable variant, including the relevant sizing system and market.

• Inventory freshness: define how recently availability must have been checked, and recheck it at the appropriate step before checkout.

• Recommendation context: preserve the original recommendation and distinguish supported alternatives from general product suggestions.

• Fallback behavior: decide what the interface shows when stock information is stale or a product has no fit coverage.

Assign ownership across the fit provider, commerce platform and retail operations team. A storefront should not display an apparently precise recommendation alongside availability information whose meaning nobody has agreed on.

Treat unavailable recommendations as signals—not guaranteed sales

Completed orders show fulfilled demand. They do not, by themselves, explain the shoppers who could not buy their recommended size.

Where appropriate data permissions and integrations exist, teams can record that an unavailable-size result occurred, what alternative was shown, and whether the shopper subsequently bought, requested an alert or left. Use an agreed session or customer key and collect only the information needed for that analysis.

Interpret those records carefully. Repeated checks by one person are not multiple buyers. An alert request is not a purchase commitment. Price, delivery timing and product preference may also explain why a visitor leaves. Segment by model, size, market and stock period before using the signal in a replenishment discussion.

This is where the broader Fit Intelligence opportunity becomes useful: a customer-facing interaction can inform a merchandising question. It does not automatically become a demand forecast, a production order or evidence of reduced waste.

Test the fallback, not just the happy path

Before release, test a supported second choice, no acceptable alternative, stale stock data, different regional size labels and a product without sufficient fit coverage. The correct outcome in some cases is a clear explanation and no purchase recommendation.

After launch, review alternative purchases alongside subsequent fit-related returns and customer feedback. A higher click-through rate on substitutes would be a weak result if those orders come back. Keep the original recommendation visible in the analysis so teams can tell whether the fallback helped the shopper or simply redirected a transaction.

The principle is straightforward: help the person find a suitable product that is available, rather than making an available product appear suitable. That is a practical way to connect personalization, commerce and merchandising while protecting trust.

See what better fit could do for your footwear business. Book a 15-minute Fit Growth Review.