Fashion & Apparel × Loop Returns
In fashion and apparel DTC, returns are not a customer-service function — they are the P&L. A brand shipping 30% return rates at $14.50 round-trip cost per return is burning $4.35 per order on returns alone, before any other cost. For a $75 AOV brand doing 120,000 orders a year, that is $522,000 annually in pure returns logistics, and it does not count the restocking, inspection, re-tagging, steaming, or write-off cost on items that come back damaged. Loop Returns has the best dataset in the industry on return rate per SKU, exchange-vs-refund ratio, and return reason — but that data sits in a Loop dashboard, disconnected from the checkout that is generating the next round of high-return orders. This is the integration that matters most for fashion: taking Loop's SKU-level return history and making it a real-time input to the profit floor at Shopify Plus checkout, so a dress with a known 42% return rate gets priced (or discounted, or excluded from free shipping) fundamentally differently than a T-shirt with a 6% return rate.
The headline return rate on a fashion DTC brand is typically 25–35%, but that single number hides a bimodal distribution that is the actual story. T-shirts, sweatshirts, accessories, and bottoms in stable sizing tend to return at 5–12%. Dresses, structured blazers, denim, shoes, and anything size-sensitive return at 35–55%. When you apply the same 'free returns, free shipping over $75' policy across both categories, you are subsidizing the high-return SKUs with margin from the low-return SKUs, and because customers are not idiots, they load their carts disproportionately with the high-return items. The economics are brutal: on a $128 dress with a 42% return rate, you need to reserve 42% × (shipping cost + return shipping + restocking + 22% damage write-off) = 42% × ($9.80 + $11.40 + $4.50 + $28.16) = $22.64 per order in expected return cost. If your gross margin on the dress is 58% ($74.24), your return-adjusted margin is $51.60, or 40.3% — still fine. But apply a 20% winback discount and free returns: $128 × 0.8 = $102.40 revenue, minus $22.64 return reserve, minus $53.76 COGS, minus $2.97 Stripe, minus $9.80 outbound shipping, minus $3 in Klaviyo touches = $10.23, or 10%, against a 25% floor. This is a $5M–$30M brand's entire margin problem in one SKU category, and every major DTC fashion brand has this exact leak running right now.
Three specific leaks live at the fashion × Loop Returns intersection. First, the 'bracketing' customer — someone who orders the same dress in sizes 4, 6, and 8 intending to return two. Loop sees this pattern; Shopify does not. The 'order' looks like $384 AOV, but the realized revenue is $128 and you have eaten three round-trip shipping costs. Without Loop data flowing into Agentis, your checkout happily approves bracketing carts with stacked discounts because the AOV looks great. Second, the 'repeat returner' — Loop tracks that roughly 8% of fashion customers account for 35% of return volume, and those customers have predictable return rates on every order. That insight should affect whether they get free shipping, free returns, or even whether a discount stack is approved on their next order — but it never does, because the Loop signal is siloed. Third, the SKU-level return cliff — Loop data shows that after 90 days in catalog, return rates on fit-sensitive items jump 6–12 points as the reviews surface fit complaints. Your checkout has no idea the same dress is now generating materially different return economics than it did at launch. Agentis bridges all three by consuming Loop's rolling 90-day return rate per SKU, per customer, per category, and piping it into the profit-floor evaluation at checkout as a first-class cost.
How Agentis Closes The Gap
Agentis subscribes to Loop's return.created and return.completed webhooks and polls the /returns endpoint nightly to maintain a SKU-level return-rate table, a customer-level return-rate table, and a category-level return-rate table. At checkout, Agentis applies the deeper of the three rates to the projected margin: if the customer has a 38% personal return rate and the SKU has a 42% return rate, Agentis uses 42%. It then multiplies that rate by the full per-return cost (outbound shipping that will be re-eaten, return shipping, restocking labor, and a category-specific damage reserve from Loop's return-reason data) and treats the result as a line-item cost on the order. If the return-adjusted contribution margin falls below the fashion floor, Agentis can: strip the discount, block the bracketing pattern, disable free shipping on the order, or flag for manual review. On the back end, Agentis also builds a 'return tax' report that shows which SKUs, which customer segments, and which promotions are driving unprofitable return volume — so merchandising and lifecycle teams can adjust pricing and promo targeting. In practice, fashion brands using this integration see a 200–450 bps lift in return-adjusted contribution margin within 60 days, almost entirely from catching the long tail of bracketing and repeat-returner orders.
Return rates are updated nightly via the /returns endpoint poll, and real-time webhooks fire on individual return events. Because return-rate math is inherently a trailing 90-day calculation, nightly is sufficient — the rate for a SKU does not move meaningfully hour-to-hour. Customer-level rates update within 15 minutes of a return being initiated in Loop.
New SKUs inherit the category-level return rate from Loop for the first 30 days in catalog, weighted toward the higher end of the range because launch reviews have not yet surfaced fit complaints. After 30 days, Agentis transitions to the SKU-specific rate. You can also manually override the rate at launch if you expect a particular style to return heavily (e.g., a new denim fit).
Agentis does not actually strip free returns — it adjusts the profit floor so the discount stack gets resized or the GWP gets removed instead. Customers almost never see a 'no free returns' message. In the fashion cohort we have measured, conversion rate moved within ±0.6% of baseline while return-adjusted contribution margin lifted 200–450 bps. The discount the customer gets is smaller, not the return policy.
Exchanges are weighted at ~30% the cost of refunds in the Agentis model, because the revenue is retained and only the logistics cost is lost. If your Loop data shows a SKU has a 35% return rate but a 70% exchange ratio, the effective refund rate is 10.5% and the return-adjusted margin barely moves. This is exactly the kind of nuance that single-number return rates hide.
Playbook
Mid-market beauty brands on Klaviyo see 35%+ of revenue come from flows — but most flows are margin-negative after GWP, samples, and discount stacking. How Agentis makes flow ROI profit-true.
Playbook
Home goods brands lose 6–15% of contribution margin on under-priced heavy-freight and oversize-parcel orders. How Agentis + ShipStation make real shipping cost a checkout-time input.
Free Audit — No Commitment
See exactly how much margin Agentis can recover for your store in 7 days — no commitment required.