The Invisible Margin Killer
Returns are part of e-commerce like checkout is part of retail. In fashion, 30 to 50% of all orders are returned. Most retailers know this. What many don't know: what each individual return actually costs — and more importantly, where those costs originate.
According to the EHI Retail Institute, 53.3% of retailers estimate up to EUR 10 in costs per return. 13.9% calculate up to EUR 20. These sound like manageable amounts. But if you don't break down return costs by cause — by product, channel, and customer segment — you're only seeing the average. And averages hide the real problem areas.
A product with an 8% return rate via Google Shopping and 42% via Instagram Ads is not the same product. At least not from the perspective of your contribution margin analysis.
Why the Overall Rate Tells You Nothing
Most e-commerce dashboards show you one number: the total return rate. Maybe broken down by product category. That's not enough.
Consider a fashion shop with three products:
Product A — Sneaker "Classic"
- Overall return rate: 18%
- Via Google Shopping: 12%
- Via Instagram Ads: 35%
- New customers: 38%
- Existing customers: 9%
Product B — Dress "Bella"
- Overall return rate: 44%
- Via Google Shopping: 40%
- Via Instagram Ads: 52%
- New customers: 55%
- Existing customers: 31%
Product C — T-Shirt "Basic"
- Overall return rate: 11%
- Via Google Shopping: 10%
- Via Instagram Ads: 14%
- New customers: 15%
- Existing customers: 8%
At first glance, Product B is the problem. But look more closely: the sneakers have a 38% return rate among new customers from Instagram Ads. If you're paying EUR 80 in customer acquisition cost (CAC) for a new customer who orders EUR 45 worth of goods and then returns them — you haven't just earned nothing, you've lost EUR 80 plus return processing costs plus shipping.
This isn't an edge case. Marketing costs per order have risen by 30 to 70% over the past five years. If you don't know which channel-product-customer segment combination is profitable, you're burning money — and won't notice until the quarter ends.
Let's run the numbers concretely: You run Instagram Ads for the sneakers, pay EUR 80 CAC, the average order value is EUR 89. The margin before marketing and returns is EUR 35. Minus EUR 80 acquisition costs, you're already at negative EUR 45. When 38% of these new customers then return their order — at EUR 12 return processing cost per item — a seemingly successful channel becomes a systematic loss-maker. And you only see this when you look at all three dimensions together.
The Data Problem: Everything Lives in Different Systems
The information you need for this analysis already exists. It just lives in different systems:
- Return data in your shop system or WMS (warehouse management)
- Channel attribution in Google Analytics or your attribution model
- Customer history in your CRM or customer database
- Product data in your PIM or ERP
- Marketing costs in Google Ads, Meta Ads, affiliate platforms
To answer the question "Which product has the highest return rate among new customers from Google Ads?", you'd need to combine data from at least three systems. Manually, that means: exports, Excel, VLOOKUP, pivot tables, hours of tinkering. And by the time the analysis is done, the data is already outdated.
This is exactly the problem we describe in our article about breaking down data silos — separate systems prevent the analyses you actually need.
What a Cause-Based Returns Analysis Includes
A meaningful returns analysis goes beyond the overall rate. It breaks down return costs by cause — like a proper contribution margin analysis:
By product:
- Which products have the highest return rate?
- Which products are still profitable after deducting return costs?
- Are there size or color variants with disproportionately high returns?
By channel:
- Which marketing channel brings customers with the highest return rate?
- Does the product presentation in each channel match reality?
- Is a channel still profitable after deducting returns?
By customer segment:
- Do new customers return more often than existing customers?
- Are there "serial returners" — and how do you handle them?
- How does the return rate differ by region, age group, or order value?
By time period:
- Does the return rate increase seasonally?
- Are returns higher after discount campaigns compared to regular business?
- How has the rate developed over the last quarters?
This analysis is the foundation for informed decisions about assortment, pricing, and channel strategy. Without it, you're flying blind. More on how to truly calculate product profitability.
Why So Few Retailers Do This Today
According to the EHI Retail Institute, only 7.3% of retailers use AI in return management. 45.5% consider AI "relevant for the future." This means: the vast majority still analyzes returns manually — or not systematically at all.
The reasons are understandable:
- Data is scattered. Without a central data foundation, you can't do cross-channel analysis.
- Technical expertise is missing. SQL queries across multiple databases require data engineers or BI specialists.
- Time. Manual preparation takes hours or days — and needs to be repeated regularly.
- The overall rate feels sufficient. As long as the total return rate is "within range," nobody questions the details.
But the money is in the details. A retailer that reduces their return rate from 35% to 30% doesn't just save 5 percentage points — they save thousands of euros in logistics, packaging, and lost revenue. With 10,000 orders per month and EUR 12 return cost per item, 5 percentage points fewer returns means savings of EUR 6,000 — every month. Over a year, that's EUR 72,000 going straight to the bottom line.
And that only covers direct costs. Add to that: capital tied up in returned goods, value loss on B-stock items, lost revenue from unavailable warehouse space, and staff time spent on return processing.
How oneAgent Makes This Analysis Possible
oneAgent connects directly to your data sources — shop system, ERP, CRM, Google Ads, Meta Ads, and over 550 other systems. No data exports, no CSV files to wrangle.
You ask questions in natural language. For example:
"Which products have the highest return rate among new customers from Google Ads?"
oneAgent automatically combines the data and delivers the answer — with concrete numbers, broken down by the dimensions you need.
More example queries:
"Show me contribution margins by product and channel — including return costs."
"Which customer segments have the highest return rate on orders over EUR 100?"
"How has the return rate for dresses developed over the last 6 months?"
No SQL skills required. No weeks-long BI project planning. The analysis you need, delivered in seconds.
The critical point: oneAgent doesn't guess. Every answer is verified against your actual data — an automatic verification layer ensures the numbers are correct before you see them. That's the difference from generic AI tools that can hallucinate business figures. We explain why this matters in detail in our article about shadow AI in the enterprise.
From Gut Feeling to Data-Driven Decisions
With a cause-based returns analysis, you can derive concrete actions:
Optimize your assortment: Remove products with persistently high return rates and negative contribution margins from your catalog, or rework the product descriptions.
Adjust your channel strategy: If Instagram Ads drives traffic but returns eat up the contribution margin — the channel isn't profitable. Regardless of how good the conversion rate looks.
Differentiate customer segments: Target new customers with high return rates specifically — for example, with better size guides or restricted payment options.
Improve product presentation: If a product is returned disproportionately often through a specific channel, the expectation often doesn't match reality. Better images, more honest descriptions, and size charts reduce the rate.
Factor return costs into acquisition calculations: If you're paying EUR 80 CAC, you need to know the return probability of the acquired customer. Otherwise, no channel pays off.
Question discount campaigns: Many retailers observe that returns increase significantly after discount events like Black Friday or flash sales. Customers order multiple variants "to try out" because the perceived risk is lower at reduced prices. If you break down the return rate by campaign period, you can calculate whether a discount promotion was actually profitable after deducting return costs.
The Competitive Advantage Is Speed
The 7.3% of retailers already using AI for returns analysis have a systematic advantage: they identify problems earlier, react faster, and optimize continuously. While other retailers do a quarterly Excel evaluation, data-driven companies have the answers in real time.
This doesn't mean you need to overhaul everything overnight. But the first step — understanding which product-channel-customer segment combinations are eating your margin — can make a measurable difference immediately.
Conclusion: Returns Aren't Fate — They're a Data Problem
The return rate isn't simply "an industry problem." It's a metric that can be systematically reduced — if you know where the causes lie. For that, you don't need an overall rate, but a breakdown by product, channel, and customer segment.
You already have the data. You just need to bring it together and be able to ask the right questions.
Try oneAgent free for 14 days and ask the questions your current dashboard can't answer. No credit card required, no installation — start right away and see where your margin is really being lost.