The return policy has, over the previous decade, evolved from an operational concession to a primary trust signal in retail commerce. The evolution was, in some sense, driven by a single retailer — Asos — whose generous returns policy in the period from 2010 onwards demonstrated that a returns approach calibrated to the customer’s confidence rather than to the supplier’s operational comfort produced commercial outcomes that exceeded the additional cost of the returns themselves. The demonstration has, in the period since, been widely cited as a model for retail commerce more broadly, and a substantial fraction of the retailers I have worked with have, at some point in their planning, considered whether to implement an Asos-equivalent returns policy on their own catalogues.
The consideration has, in essentially every case I have observed, ended in the same place. The retailer has examined the operational implications, has projected the increased return rate that the policy would produce, and has concluded that the cost is, on the available data, not recoverable for the retailer in question. The conclusion is, in the great majority of cases, correct; the Asos policy works for Asos and does not, for reasons that are worth articulating in some detail, work for the median retailer attempting to clone it. This post is an account of those reasons, and of the conditions under which the policy nevertheless can be made to work.
Why the Asos policy works for Asos
The Asos returns policy permits, in essentially its current form, the customer to return any item within twenty-eight days of receipt at no cost to the customer, with the refund processed within several business days of the return’s arrival. The policy is, on the operational data Asos has occasionally disclosed, sustained by a return rate that is, for the apparel category, in the range of thirty per cent of orders by item count. The return rate is, in absolute terms, substantially higher than the return rate of retailers operating less generous policies in the same category; the higher return rate is operationally accommodated by the retailer’s logistics infrastructure, which has been designed for the volume the policy produces.
The mechanism by which the policy produces commercial returns despite the elevated return rate is, on the available analysis, approximately the following. The policy lowers the perceived risk of the purchase to the customer; the lowered risk increases the customer’s willingness to commit to purchases the customer would otherwise have considered insufficiently certain to proceed with. The increased commitment produces, on aggregate across the customer base, more orders than would otherwise have been placed; a substantial fraction of the additional orders are returned, but a non-trivial fraction are retained, and the retained fraction is sufficient to recover the operational cost of processing the returned fraction.
The mechanism’s commercial viability depends on three properties of the retailer that are, in combination, not common across the broader retail landscape. The first is the unit economics of the category — the gross margin per retained unit must be sufficient to absorb the cost of processing several returned units alongside it. The second is the logistics infrastructure — the cost per returned unit must be sufficiently low that the absorbed cost remains tolerable. The third is the LTV horizon — the customer who is acquired through the lower-perceived-risk policy must remain a customer over a long enough horizon that the lifetime value of the acquired customer exceeds the cumulative cost of their return behaviour.
Asos satisfies all three properties: the apparel category has favourable gross margins in the price points Asos operates at, the retailer’s investment in returns logistics has produced per-unit costs that are substantially below the industry norm, and the apparel category’s customer-relationship horizon is long enough for the LTV calculation to support the policy. The retailer, in other words, has constructed a defensible commercial moat around the policy; competitors who attempt to match the policy without satisfying the three underlying properties produce policies that are, in operational terms, indistinguishable from charity.
Why the policy does not work for most retailers
The retailers who have, in my observation, attempted to clone the Asos policy without satisfying the underlying properties have, with very few exceptions, produced commercial outcomes that did not justify the policy’s continuation. The pattern is approximately the following.
The retailer announces the generous returns policy. The customer base, responding to the lowered perceived risk, increases its purchasing volume. The return rate increases proportionally — in many cases more than proportionally, since the marginal customers acquired by the policy are, by hypothesis, less certain about their purchases than the inframarginal customers were. The retailer’s logistics costs increase substantially; the retailer’s gross margin per retained unit is insufficient to absorb the cost increase; the retailer either rolls back the policy after several quarters of operational losses or sustains the policy at the expense of operating losses that the retailer cannot indefinitely fund.
The cycle is well-documented across retail-trade-publication coverage of returns-policy experiments; the pattern is sufficiently consistent that it should, in principle, deter retailers from attempting the experiment. The pattern nevertheless repeats, principally because the procurement conversations that lead to the policy decisions treat the Asos example as a precedent for the policy’s commercial viability rather than as a special case whose conditions the retailer’s own operation may not satisfy.
The conditions under which the policy can be made to work
The policy can, in my observation, be made to work for retailers whose operations satisfy the three properties Asos satisfies — favourable gross margins, low per-unit returns logistics costs, and long customer-relationship horizons. The fraction of UK retailers that satisfy all three properties is small; the retailers who do satisfy them are, in most cases, the larger operators whose scale supports the logistics infrastructure investment.
For the smaller and mid-market retailers who do not, on inspection, satisfy the three properties, the policy in its full Asos-equivalent form is not commercially viable. The retailers’ available alternatives are, in approximate order of decreasing ambition, the following.
The first is the policy that approximates the Asos generosity but applies only to a subset of customers — typically members of a loyalty programme, or customers above a certain prior-spend threshold. The pattern accommodates the LTV requirement explicitly: the customers whose LTV is high enough to support the policy receive it, while the customers whose LTV does not yet justify it do not. The pattern is operationally manageable and produces, on the data I have collected from clients who have implemented it, a measurable conversion uplift on the qualifying customer segment without the unsustainable cost of the universal application.
The second is the policy that approximates the Asos generosity but is funded by a small price premium across the catalogue. The premium is, in operational terms, the customer paying for the option to exercise the generous return; the policy continues to function as a trust signal because the premium is, on the customer’s perception, modest and the option’s perceived value at the moment of decision exceeds the premium’s perceived cost. The pattern requires careful price calibration but is, in well-implemented cases, a viable middle path.
The third is the explicit communication of a less generous returns policy in the most positive operational framing the policy honestly supports. The pattern does not, in any sense, produce the commercial outcomes the Asos policy produces; it does, however, avoid the operational losses of the unviable clone and recovers some of the trust-signal benefit through the explicit and confident communication of the policy that the retailer can actually operate.
An advisory close
The Asos returns policy is, considered as a commercial mechanism, a remarkable success and one that has, over the previous decade, raised the bar of the broader retail-commerce trust contract; the policy is also, considered as a procurement reference, a highly specific case whose underlying conditions the great majority of retailers do not satisfy. The retailers who have, in my observation, made the policy work are those who have either inherited the underlying conditions or have, over a sustained period, invested in constructing them; the retailers who have attempted to leap to the policy without first constructing the conditions have, almost without exception, produced operational losses that the policy did not recover.
It is recommended that retailers considering returns-policy reform examine the underlying unit economics of the operation before committing to the policy itself; the unit economics determine whether the policy is operationally viable, and the policy is, in commercial terms, a downstream consequence of the economics rather than an independent variable that can be set without reference to them. (For the related view on how the trust landscape has, over the same period, shifted in ways that have made the returns policy a more important trust signal, see Why your product reviews feel fake (even when they are real).)