Skip to main content

What is Pinch

Pinch is an AI-powered platform that helps retailers control return abuse, reseller abuse, and post-purchase fraud, surface their most valuable customers, and apply different policies to different shoppers — all without changing checkout or rebuilding their existing returns flow.

The problem Pinch solves​

Return abuse costs US retailers over $100B a year, and reseller abuse — bulk-buyers reselling on secondary marketplaces — costs another $30B+. The patterns most retailers see:

Return abuse

  • Customers wearing items and returning them as new (wardrobing)
  • Empty-box returns and counterfeit-item swaps
  • Repeat abusers who return 70%+ of every order they place
  • Bracketing — buying three sizes to keep one — at scale

Reseller abuse

  • Bot-driven order placement on limited-release and high-margin SKUs
  • Bulk-buyers using multiple identities to circumvent purchase limits
  • Resellers who tie up inventory and erode brand equity by reselling outside your authorized channel

A flat returns or order-acceptance policy can't tell these customers apart from your best ones. Pinch can.

How Pinch fits in​

Pinch sits next to your existing order and returns systems, not in the way of them. You keep Shopify, Salesforce, Magento, or whatever you use for checkout. You keep Loop, Aftership, Narvar, or whichever RMS handles your returns portal. Pinch reads order and return events from those systems, scores each customer in real time, and writes a decision back so your existing tools can act on it.

The decision is one of three: Allow (process the return normally), Deny (block it), or Review (route to a manual queue). Every customer also gets a 0–100 Pinch Assessment score and is grouped into one of four segments — VIP, Others, Suspect, or Egregious — so you can apply different policies to each.

Where Pinch acts in the lifecycle​

StageWhat Pinch does
SignupDetects multi-account abuse, disposable emails, referral fraud (optional API)
LoginDetects credential stuffing and account takeover (optional API)
OrderScores the order against the customer's full history; surfaces in analytics
Return initiatedReturns a decision: Allow / Deny / Review. Writes a tag back to your order.
Return receivedReads disposition data — what condition the item came back in. Updates the model.
WarehouseThe refund decision is made at the warehouse based on Pinch's signal and the inspected item condition.
Dispute / chargebackFeeds outcomes back into customer scoring

For most merchants, the Order and Return stages are the only ones in scope at install. The others are available when you're ready.

Who Pinch is for​

  • Retailers with 100K+ annual orders seeing return rates above 8% or abuse signals (empty boxes, wardrobing complaints from customer service, suspicious customer concentrations)
  • Brands with high-value SKUs where each abusive return costs tens of dollars in margin
  • Apparel, footwear, jewelry, electronics, beauty — anywhere returns are a meaningful share of operations

Pinch is generally not the right fit for: marketplaces with sub-$10 AOV, B2B distributors with single-digit return rates, or merchants who already enforce a no-returns policy.

What you don't have to do​

This is the part that surprises new customers most: you don't author rules. Pinch's AI creates the first set of logic for you and is verified by our team during white-glove onboarding. The product is configurable end-to-end if you want it — but the default experience is that you install, connect a couple of integrations, send us your disposition data, and your CS manager walks you through what's been set up.

What you do have to do​

Three things at install:

  1. Click the install link for your platform (or contact us for SFCC).
  2. Connect your returns platform (RMS) — Loop, Aftership, Narvar, etc.
  3. Send disposition data — either auto-sync from Aftership, or weekly CSV upload at first.

After that you're using Pinch from the dashboard: the Review queue, customer profiles, and analytics. About 30 minutes of operator time per day for a mid-sized merchant.

Next​