The Returns Tab Became Our Best Media Buyer
TLDR: Our ad dashboard kept telling us which campaigns were cheap. Our returns and sizing data told us which campaigns were actually profitable, and the two answers were not the same. Reading the post-purchase numbers reshaped how we split paid budget across our basics, and it cut money we were quietly burning on customers who sent product back.
I run marketing for a direct-to-consumer menswear basics brand. Underwear, socks, sleepwear, the unglamorous categories people reorder without thinking. For a long time I treated the ad platform as the scoreboard. Cost per acquisition was low on a campaign, so I fed it more. Cost per acquisition climbed somewhere else, so I starved it. That is how most of us are trained to read paid media, and it works right up until the customer does something the dashboard never shows you. They open the box, try the product, and decide whether to keep it.
The thing nobody warns a new ecommerce marketer about is that the ad platform stops watching at checkout. It counts the sale and moves on. Everything that happens after, the fit complaint, the wrong size, the refund, the support ticket, lives in a different system entirely. For us that system was a returns spreadsheet the operations team kept, and I had barely looked at it.
The Day the Two Reports Disagreed
It started with a campaign I was proud of. One of our core briefs and slips had a campaign pulling in sales at a cost per acquisition well under our target. On the ad side it looked like the best thing we were running. I was about to push more budget into it when our ops lead mentioned, almost in passing, that one style in that line was coming back more than the others.
So I pulled the returns export and lined it up against the ad data for the first time. Returns on that one style ran about 18 percent. Most of the catalog sat closer to 6 or 7 percent. Nearly every return note said the same thing in slightly different words: ran small, sizing off, expected a different fit. The campaign was not cheap. It was expensive in a place my dashboard could not see. Once I subtracted the refunds, the reship costs, and the support time, the real cost per kept customer was higher than campaigns I had been treating as mediocre.
That was the uncomfortable part. The "winning" campaign was acquiring people who did not stay. The platform rewarded me for the click and the checkout. It had no idea the product was going back.
What the Sizing Notes Were Actually Saying
I stopped reading returns as a logistics problem and started reading them as market research, which is what they are. Every return is a customer telling you, with their own money on the line, exactly where the promise and the product split apart. When 18 percent of buyers on one style write that it ran small, that is not a warehouse issue. That is a positioning and expectation issue, and it almost always starts in the ad and on the product page.
The fix was not a clever new campaign. We rewrote the product description with the fit reality stated plainly, added a sizing line that told people to size up on that specific cut, and changed the creative so the model and the copy matched what arrived in the box. None of that is the work the ad platform tells you to do. It does not appear in any optimization tip. It came straight out of reading customer complaints in volume.
Returns at the catalog level are not a small line item either. The National Retail Federation put the 2023 industry return rate at 14.5 percent of merchandise, which is real margin leaving through a door most marketers never look at. Baymard Institute has also documented for years that unclear product information and sizing uncertainty are among the reasons people abandon carts and, when they do buy, return, which you can see in their work on cart abandonment. The signal was always there. I just had not been treating my own returns tab as the same kind of evidence.
How We Re-Split the Paid Budget
The reallocation was less dramatic than it sounds, because the data made it obvious. I built a simple second column next to cost per acquisition: cost per kept customer, after returns. Suddenly the ranking of our campaigns reshuffled. A few "cheap" winners dropped because their products came back. A couple of campaigns I had underfunded looked far better once I saw their products almost never returned and those buyers reordered.
We shifted roughly a fifth of paid spend away from the high-return styles and toward the categories that held. Our sock range was a quiet example. Lower headline excitement, almost no fit complaints, steady reorders. On the ad dashboard alone I never would have leaned into it. On the kept-customer view it was one of the most profitable things we sold. We did not invent a fancy attribution model to figure this out. We put two existing spreadsheets next to each other and read them honestly.
The Shopify team makes a similar point in their guide to handling ecommerce returns: returns are not only a cost to minimize, they are feedback about where the buying experience is breaking. Once you accept that, the returns tab stops being a chore for the ops team and becomes one of the cleanest market-analysis tools you have, because the customer is paying to tell you the truth.
What I Would Tell Another DTC Marketer
Your ad platform is a great salesperson and a terrible analyst. It optimizes for the event it can measure, the conversion, and it is blind to everything after. If you let it pick your winners alone, you will keep paying to acquire customers who quietly leave through the returns door, and you will keep calling it a win because the cost per acquisition looked good.
You do not need new software to fix this. Pull your returns export. Line it up against your campaigns by product. Add one column for what an acquisition really costs you after the product comes back. Read the return reasons in bulk, because they are the most honest customer research you will ever get for free. The campaigns that survive that view are the ones worth funding. The fit notes you find will improve your product pages, which lowers returns, which makes the next round of ads cheaper for real, not just on the surface.
When you start buying media against customers who keep what they bought instead of customers who merely checked out, you spend less on returns you never saw coming and more on the people who actually wanted the product. That is a calmer, more confident way to grow, and it started for us with a tab I had been ignoring. You can see the kind of basics we build this way at Mariner.
