A return is a sale that unravels. You paid to acquire the customer, paid to ship the product, and now you'll pay to ship it back, inspect it, restock it or write it off, and process the refund. And most of the time, the customer isn't unhappy with you. They just guessed wrong about the size.
That last part is the opening. A large share of returns aren't quality problems. They're information problems, and information problems can be solved before the order is ever placed.
Most returns are decided before checkout
It's tempting to think of returns as something that happens after delivery. Really, they're set in motion at the moment of purchase, when a shopper commits to a size, a color, or a spec without being fully sure.
Someone buying a jacket online can't try it on. They squint at a size chart, guess, and hope. Someone buying a desk can't stand next to it. They eyeball the dimensions and picture their room, sometimes wrong. The uncertainty at that moment is the seed of the return, and the return is just that uncertainty coming due weeks later.
Reduce the guessing at checkout and you reduce the returns downstream. That's the whole strategy.
The three questions behind most returns
Preventable returns tend to trace back to three unanswered questions.
Will it fit me? The apparel and footwear classic. The customer wasn't sure between a medium and a large, picked one, and got it wrong.
Will it fit my space or my other stuff? Furniture, appliances, parts, accessories. Will the couch clear the doorway? Is this charger compatible with my model?
Is it actually what I think it is? Material, color accuracy, capacity, features. The rug looked grey on screen and arrived beige. The "large" bottle turned out smaller than expected.
Every one of these is answerable with information you already have. The problem isn't that the answer doesn't exist. It's that the shopper couldn't get it at the moment they needed it.
Answer fit questions before checkout, not after
A chatbot trained on your sizing charts, specs, and product details can meet the shopper at the exact moment of doubt and answer the question that would otherwise become a return.
Picture Kestrel Apparel, a small clothing brand. A shopper asks, "I'm 5'8 and usually wear a medium in other brands, what size should I get in your relaxed-fit shirt?" A human might answer that the next morning, long after the shopper bought a medium and set up a return. A bot trained on the brand's fit notes answers in the moment: "Our relaxed fit runs roomy, so a medium will be loose on you. If you like a closer fit, size down to small." That single exchange can be the difference between a kept order and a round-trip shipment.
The mechanics are simple. Feed the bot:
- Your full size charts, with real measurements, not just S/M/L labels
- Fit notes (runs small, true to size, roomy in the shoulders)
- Product dimensions and weights
- Compatibility details (works with these models, requires this adapter)
- Material and care specifics
In SpideyChat you'd upload these as documents or add them as Q&A pairs, then place the bot on your product pages so the answer is one message away from the buy button.
Prevent returns, don't just delay them
Here's the tradeoff worth being honest about. A chatbot can reduce returns two ways, and only one of them is good.
The good way: it gives accurate information that helps a shopper choose correctly, so the product that arrives is the product they wanted. That prevents a return and keeps a happy customer.
The bad way: it oversells, glosses over a poor fit, or nudges someone into a purchase that isn't right for them. That might delay a return, but it creates an unhappy customer and, often, a return anyway, plus a bad review.
Aim only for the first. A bot that tells a shopper "honestly, this runs small and might be tight on you, you may want the next size up" is doing exactly its job, even when that means talking someone out of the size they were about to buy. Sometimes the best return-prevention is an honest "this might not be the right product for you." That builds the kind of trust that brings people back.
The spec data most stores are missing
You can't answer what you haven't written down. The most common reason fit bots underperform is thin source data.
Run through this checklist for your top products:
- Real measurements for every size, not just letter labels
- Fit guidance in plain language (how it runs, where it's tight or loose)
- Dimensions for anything that has to fit a space
- Weight, capacity, and material where they matter
- Compatibility notes for anything technical
- Honest color descriptions, since screens lie
If your product pages already carry this, great, a crawl will pick it up. If the details live only in a supplier PDF or a staffer's memory, get them into the bot as documents or Q&A pairs. The bot can only be as helpful as the specifics you give it.
Measure the returns you prevented
Return prevention is a little tricky to measure, because you're counting things that didn't happen. Still, you can watch the right signals.
| Signal to watch | What a good result looks like |
|---|---|
| Fit and sizing questions in chat | High volume answered instantly, pre-purchase |
| Return rate on top products | Trends down after adding fit answers |
| "Wrong size" return reason | Shrinks as a share of returns |
| Repeat purchase rate | Holds or rises, since customers got the right item |
Tag or review your chat transcripts for sizing and fit questions so you can see how many shoppers ask before they buy. Then watch your return reasons over the following weeks. If "wrong size" or "didn't fit" starts shrinking as a share of returns, your pre-purchase answers are doing their job.
One more habit pays off. When a return does come in with a preventable reason, treat it as a content gap. The customer thought the shirt would fit and it didn't, so add or sharpen the fit note that would have set the right expectation. Over time, your returns themselves become a to-do list for making the bot smarter.
Returns will never hit zero, and they shouldn't. Some are legitimate, and a generous return policy is part of why people trust buying online. But the returns born from a guess at checkout are avoidable, and the fix is mostly about getting the right information to the right shopper at the right second. Put your sizing and spec answers where the doubt lives, keep them honest, and watch how many round trips you quietly prevent.