Verification-First Return Prediction for Shopify Plus
Know which carts come back — before you change anything.
Install the ReturnML Shopify app — no dev work required. For 60 days we score every cart silently, completely invisible to your customers. Then we prove our predictions against your own order history before you change a single thing.
Only 5 founding pilot slots available
Fashion e-com loses €850B/year to returns
Most brands fight it with two bad tools.
Built for Nordic, UK & US fashion brands on Shopify Plus — dresses, denim, and activewear, where size & fit drive 45%+ of returns.
Blanket free returns
Lose €8+ on every returnable order. You train customers to buy five, keep one.
Charge everyone
Kill conversion. Good customers leave. Return rate barely budges.
ReturnML
Score every cart silently for 60 days. See exactly what returns cost you — then decide if and when to act.
How it works
From Shopify install to verified proof — in your first week, not months.
- 01
Connect
Install the ReturnML app from the Shopify App Store — no dev work required. Prefer to integrate yourself? A single API call works too.
- 02
Rewind
In your first week, we backfill 12–24 months of your order and returns history, train on the older portion, and show verified accuracy on the recent months our model never saw.
- 03
Watch
For 60 days ReturnML scores every live cart in the background. No customer-facing changes, no pricing changes, zero conversion risk.
- 04
Decide
Get the Money Report: verified accuracy and euro impact on your own carts. Turn on actions only once the numbers convince you.
Live demo
See the silent scoring in action
Add items to the cart below. When you hit checkout, ReturnML scores the cart in real time — exactly as it would in shadow mode. The customer sees nothing; the score and simulated action are just logged for the Money Report.
Modanova
Shopify PlusNew Arrivals
Silk Wrap Dress — Red, M
DRESS-M-RED
High-Rise Jeans — W32
JEANS-32
Oversized Blazer — Black, L
BLAZER-L-BLK
Linen Shorts — Sand, S
SHORTS-S-SND
Ribbed Knit Top — Cream, XS
KNIT-XS-CRM
Your Cart (0 items)
Add items from the store to get started
After 60 days
The Money Report
Verified prediction accuracy on your own carts, euros lost to flagged carts, and the simulated profit of each intervention — before you turn anything on.
48,211
Carts scored
6,904
Flagged high-risk
61%
Verified return rate among flagged
€54,300
Return costs identified
Simulated net profit — opt-in coupon action, by operating point
Conservative
+€6,800
Balanced
+€11,200
Aggressive
+€14,900
Each threshold trades flagged-cart volume for precision. This is what the verified outcome data says you would have kept had you offered the coupon at each operating point — you pick the threshold, not us.
Merchandising insights
- Worst-performing SKU: Silk Wrap Dress (Red, M) — 62% return rate
- Bracketing hotspot: 3+ sizes of the same style ordered together, concentrated in denim
- Return rate by payment method: Klarna carts return 1.4× more often than card
Numbers above are illustrative. Your report is built from your own scored carts joined against your actual Shopify return outcomes. Cost per return, coupon %, and adoption rate are assumptions you can edit — nothing here is estimated from other merchants.
One call. Full picture.
This is what runs under the hood of the Shopify app — call it directly if you'd rather integrate yourself. Risk score + simulated profit per action, logged for the Money Report; nothing here is executed against the customer in shadow mode.
// POST /v1/score
{
"session_id": "usr_8f2a",
"mode": "shadow",
"cart": [
{ "sku": "DRESS-M-RED", "qty": 2 },
{ "sku": "JEANS-32", "qty": 1 }
]
}
// Response (47 ms) — logged, not executed
{
"return_risk": 0.83,
"simulated_actions": [
{ "type": "coupon_5pct", "delta_profit": 4.10 },
{ "type": "charge_delivery", "delta_profit": 2.40 },
{ "type": "do_nothing", "delta_profit": 0.00 }
],
"explanation": [
"3 similar items in cart",
"62% historical return rate",
"COD payment method"
]
}Not just a guess. Verified.
Most tools say “this order is risky”. ReturnML says “offering the non-returnable coupon would have made you €4.10 more than doing nothing — and here's the verified data that proves it.” That's causal ML, not correlation.
Verified before you act
The shadow pilot IS the proof: we show you precision and recall on your own carts before a single prediction changes anything for a customer.
<50 ms p99 latency
Built for checkout speed. ReturnML sits in your critical path and still returns before the page finishes loading.
Guardrails included
VIPs and first-time buyers are automatically protected. Never charge a top customer by accident.
Explainable by default
Every score comes with SHAP-based reasons your ops team can actually read — and your legal team can defend.
Once the Money Report convinces you
3–6% fewer returns.
Zero conversion loss.
That's the simulated upside for merchants who turn on live actions after the pilot. For a €100M fashion brand, that's €2–3M in net margin per year. You keep more sales and keep the right customers.
Launching with Shopify Plus. Salesforce Commerce Cloud and custom stacks follow via the same API.
Common questions
- Will customers notice anything during the pilot?
- No. The 60-day shadow pilot is fully read-only — we score every cart in the background. No checkout copy, pricing, or flow changes for a single customer until you explicitly opt in to actions.
- How does the retroactive report work?
- We don't ask you to trust our model. On install, we backfill 12–24 months of your order and returns history, train on the older portion, and show you its precision on your own historical orders — before a single live cart is scored.
- How is accuracy verified?
- We join every scored cart against actual return and refund outcomes from Shopify's returns API once the return window closes — typically 30–45 days. Labels only count after that window plus a buffer, so the accuracy numbers you see are verified on your own data, not a vendor benchmark.
- What do you need access to?
- Just the Shopify app with minimal read scopes — orders, returns, and the customer metadata needed to score a cart. We don't touch your theme or checkout flow, and no write access is required for the pilot.
- What data do you store?
- Hashed cart/session IDs and outcome labels only — no plaintext PII, GDPR-clean. Under the pilot terms, anonymized outcome data helps improve our cross-merchant models; this is stated transparently upfront and you can opt out at cancellation.
- What does the pilot cost?
- Nothing for founding merchants. The retroactive report, the 60-day shadow pilot, and the Money Report are free. You keep the report whether or not you continue with ReturnML afterward.
- Which actions can we configure?
- The flagship action is an opt-in, non-returnable coupon offered at checkout. A personalized delivery charge is available as an advanced option. You set the thresholds; ReturnML returns the simulated profit delta for each option so you decide — and these only go live after the pilot.
- How long does integration take?
- Under a day. Install the Shopify app, or fire a single POST to /v1/score yourself if you'd rather integrate directly. No data pipeline, no model training, no sprint.
- What if our stack isn't Shopify?
- We're launching with Shopify Plus first. ReturnML's scoring API is platform-agnostic under the hood, so Salesforce Commerce Cloud, Magento, and custom stacks can follow via the same REST API.
- How are VIP customers protected?
- Guardrails are built in for when actions go live. First-time buyers and high-LTV customers are automatically excluded from any friction actions — you never risk charging your best customers by mistake.
Ready to see what your returns are actually costing you?
Claim one of 5 founding pilot slots — read-only, free for 60 days. We'll show you the numbers on your own data before you change anything.