Your product returns are not a logistics problem: the advice never happened

You reduce ecommerce returns before the order, not after it. In the US, 19.3% of online sales were returned in 2025 (NRF and Happy Returns), and the top reasons shoppers give are fit, then a product that did not match what they expected. Put plainly: nobody advised the customer at the moment of choosing.

What a return weighs in 2025, and what the numbers leave out

Start with what gets measured. In October 2025 the National Retail Federation and Happy Returns published their annual estimate: $849.9 billion of merchandise returned in the US over the year, or 15.8% of total retail sales. Online, the rate climbs to 19.3%. One item in five comes back. The study rests on two surveys run in summer 2025, of 2,006 consumers and 358 retail professionals.

Digital Commerce 360 added a seasonal reading in January 2026: according to Salesforce, returns represented 14% of all purchases between November 1 and December 31, 2025, worth $181 billion, up 10% on the previous holiday season. The peak selling weeks are also the peak returning weeks. What nobody publishes is a clean return rate by country or by category from a primary source. Sector percentages circulating on logistics blogs rarely name where they come from. We do not repeat them.

What can be said fits in one sentence. Physical stores return less than online stores, and the difference has nothing to do with the carrier. In a shop the customer touched, tried, asked a question and got an answer. Online, they chose alone, between several product pages, and that solitary choice is what comes back in a box.

Why customers send your products back: the reading that changes everything

Surveys of return reasons look alike from one country to the next. Shorr Packaging polled 2,013 Americans in November 2025. The result: 44% of returns came from an item that did not fit or was the wrong size, 31% from a damaged product, 11% from an item that did not match its description, 9% from a simple change of mind. Rithum, across more than 6,000 consumers in the US, Canada, the UK, France and Germany (May 2025), found 61% citing poor fit and 33% a product that did not match its photos or description.

Now read those lists differently. Remove the damaged parcel, which belongs to shipping and packaging. Everything left, fit, size, a product unlike what was imagined, a change of mind, describes a single situation: the customer made a decision without the information they needed, and the product arrived to correct it. These are not logistics returns. They are returns of decision.

The "not as described" return is a return of expectation

Signifyd, quoted by Digital Commerce 360 in January 2026, ranks "significantly not as described" as the top reason for all 2025 returns, at nearly 48%. Read that figure with care, because it includes a share of fraud the same company says is rising. But the point lies elsewhere. A product "not as described" is almost never a product badly described. It is a product the customer built an idea of, alone, from a page written for everyone and therefore for no one. The expectation was set wrong. The parcel exposed it.

The "fit" return hides a decision made without help

Fit dominates apparel, and this article is not written for fashion retailers whose problem is body shape. It is for brands whose basket needs advice: skincare, supplements, sports equipment, jewelry, gear, technical products. For them, "fit" means dosage, range, compatibility, level, format. A customer who orders the cream for oily skin when theirs is dehydrated makes exactly the same move as the one who orders a medium instead of a large. They chose without anyone asking the question that would have settled it.

Bracketing proves it by absurdity: 51% of Gen Z shoppers order several variants to keep the right one and send the rest back, against 24% of baby boomers (NRF and Happy Returns, 2025). The customer organizes for themselves the fitting the site never offered. They turn your warehouse into a changing room, at your expense.

The diagnosis in four questions

Before renegotiating your carrier contract, pull your returns from the last three months and answer four questions. In an hour they tell you whether your problem is operational or commercial.

First question: what share of your returns carries a "no defect" reason? Strip out damaged parcels and picking errors. What remains (wrong choice, does not suit me, not what I expected, changed my mind) is your decision return rate. For most advice-driven brands it is the majority, and the surveys above suggest the same for the market as a whole.

Second question: what did the customer see between landing and add to cart? If they read a page, compared two tabs and clicked, they chose alone. If they answered questions about their need and received a reasoned recommendation, they were advised. There is no third case. And the first one produces the returns.

Third question: how many similar products did you show them at once? Three serums, five formulas, eight models. The wider the offer, the more fragile the solitary decision, a mechanism we documented in our piece on the paradox of choice in ecommerce. A rich catalog with no guide is a catalog that comes back.

Fourth question: what does your returns curve look like after an acquisition campaign? If cold traffic returns markedly more than loyal customers, you hold the proof. The loyal customer knows themselves and knows your products. The newcomer buys blind whatever the ad showed them, and they are the one filling the return boxes.

What logistics fixes, and what it cannot

Let us be fair to return portals, prepaid labels and exchange policies. They matter, because customers look at them before buying: 88% expect free returns, 41% read the return policy before ordering and 47% have stopped shopping with a retailer because of it (Rithum, 2025). Poor returns logistics loses sales. Good logistics saves some.

But it does not reduce returns. It makes them smoother, faster, sometimes cheaper per unit. The product still comes back, still has to be inspected, repackaged, sometimes marked down. And the opportunity cost is real: 60% of retailers surveyed by NRF and Happy Returns in 2025 had, at some point in the year, to choose between shipping new orders and processing returns. When the warehouse saturates in December, it is the holiday sales that wait.

The last row matters. We are not claiming everything gets solved before checkout. A crushed box or a picking error belongs to operations, and that is where logistics excels. But on the first three rows, which weigh heaviest in every survey, it arrives after the battle.

Putting advice back before the order: what a video funnel does

The in-store salesperson reduces returns by asking three questions before handing over the product. A video funnel does the same thing at the scale of a website. The visitor answers a quiz of 5 to 15 questions asked on video by the seller: which skin, which goal, which level, which use, which budget. From those answers they receive a personalized analysis video of 10 to 15 minutes, assembled automatically from segments the seller really filmed. In it, the seller explains what they understood of the need, and why they recommend this exact product rather than its neighbors in the range. Then the visitor buys.

That mechanism acts on every decision return reason. "Wrong choice" recedes because the recommendation is made for this person, not for an average customer. "Not what I imagined" recedes because a human showed the product, described it with its limits, set the expectation. Bracketing loses its purpose: when someone has told you which one to take and why, you do not order three. The salesperson happened.

Anna Velazia, a jewelry and crystal-healing brand using VideoFunnel, built its funnel around 14 questions and a 12-minute analysis video. It puts 70% of its ad budget behind that funnel, which delivers a ROAS three times higher than its other campaigns. Those are its figures, on its products, not an average. What they say is that a customer who received twelve minutes of personal explanation before buying does not buy like a customer who read a product page.

Let us be precise about what we do not promise. We publish no average drop in return rate attributable to the video funnel, because our customers do not report it in a comparable way and we do not invent statistics. What we do assert is a cause-and-effect link that every survey of return reasons makes plain: a return of decision assumes a decision made without advice, and the video funnel puts advice back where it belongs, before the order. The full setup is in our guide to the video funnel for ecommerce.

"My customers will never answer fifteen questions before buying"

It is the objection we hear most, and it deserves a numbered answer. Interact, across its own quiz base in 2026, measures a 65% completion rate and, for ecommerce stores, 37.6% of visitors who start go on to leave their details. RevenueHunt, on 45 million responses collected from more than 20,000 stores (2026), reports 69% completion and 5.5% ordering after the recommendation, with an average order value 11 to 15% higher. People answer. They answer because, for once, someone asks what they want instead of showing them everything in stock.

Turn the objection around. The customer who will not answer five questions is the one who, today, opens three tabs, hesitates ten minutes, orders two variants and sends one back. The time they refuse to give you before the order, they make you pay for after it. And a chatbot does not replace that advice: it answers the questions the customer knows to ask, not the ones they should ask, a difference we measured in our comparison of an AI chatbot versus quiz and video.

A return is a sale that was left to decide itself

Reread your return reasons with that sentence in mind. Behind "does not suit me", "not what I expected", "I ordered two sizes", there is always the same moment: the one where the customer, alone in front of your page, placed a bet. Logistics absorbs the lost bet. It does so better and better, and good for it. But the bet is placed earlier, in the minutes before add to cart, exactly where your store has no salesperson.

The video funnel puts a salesperson in that precise spot. It asks the questions, listens to the answers, recommends one product and explains why, on video, with a face. The customer buys what fits them, and knows why. The parcel that arrives looks like the one they were told about. That is not a logistics problem solved. It is a sale that was finally advised.