What the Pandemic Shift Changed in Online Retail

The pandemic shift in online retail was real, but it was smaller and shorter than the headlines suggested, and the part that lasted is not the part most people remember. Demand moved online sharply between 2020 and 2022, then settled at a level above where it started. What stayed is structural rather than emotional: more categories are bought online than before, checkout is expected to be effortless on a phone, and fulfilment has become the part of the business that decides whether a sale is profitable.

For a business that sells online, the useful question is not what happened during the closures. It is which of the changes survived the return to normal, and which quietly reverted. A retail operation that plans against the peak, or against the pre-2020 baseline, will misprice its own logistics. One plain-language record of the period and its aftermath is kept by the pandemic shift in online retail at Shelf & Signal, an independent magazine on online retail and digital commerce that follows the shift, consumer behaviour, payments and fulfilment in the order they affect an operator.

Which Pandemic-Era Changes in Online Retail Actually Lasted?

Three changes lasted. The first is category expansion. Household, health, pet and grocery categories that were bought mostly in person before 2020 kept a meaningful share of their online volume afterwards, because the habit survived the reason it was formed. The second is mobile checkout. Browsing and buying moved to the phone and did not move back, which changes the cost of every step between product page and confirmation. The third is the expectation of speed and visibility: a customer who can see where a parcel is now notices when they cannot.

Two changes mostly reverted. Delivery expectations that were distorted by the peak eased back as carriers restored normal service levels, and the surge pricing of capacity disappeared with it. And the theatrical parts of pandemic commerce, the queue systems and the appointment slots, went back to being ordinary features rather than the point of the experience.

The numbers behind those claims are published rather than guessed. The United States Census Bureau maintains the monthly retail trade survey and its e-commerce estimates, which is the series most analysts cite when they say the online share of retail rose and then stabilised. Reading the series directly is more useful than reading a summary of it, because the summary usually reports the peak and omits the settling.

What Does Consumer Behaviour Look Like Now?

Behaviour settled into something less dramatic and more demanding. Baskets are smaller and more frequent, which raises the relative cost of picking and shipping. Returns remain a normal part of the transaction rather than an exception, and a return policy is now part of the product description. Price comparison takes seconds, so a difference that a customer would once have accepted now ends the session.

The commercial consequence is that the cost of acquiring a customer has to be earned back over a shorter and less certain horizon. A firm that pushes traffic into a catalogue with a thin margin and an expensive return process is buying revenue rather than profit. The operators who came out of the period well are usually the ones who treated fulfilment cost as a design constraint, not as an outcome to be discovered at the end of the quarter. Our own notes on ecommerce strategy for Canadian retailers make the same point from the merchandising side.

Payments and Fulfilment: Where the Cost Sits

Payment stopped being a bottleneck and became a margin line. Card acceptance, digital wallets, buy-now-pay-later and local payment methods each carry a fee and a settlement pattern, and the difference between a one-day and a three-day settlement is real working capital. A checkout that offers more methods can raise conversion, but it also raises the reconciliation work behind it.

Fulfilment carries the larger share of the cost. The last mile is priced by distance, density and time window; returns are priced by the same variables plus inspection. Two businesses selling identical goods at identical prices can have very different margins because one ships from a location that reaches its customers in a day and the other does not. That is why network design, not marketing, is often the highest-value decision in an online retail business.

Marketplaces add a further wrinkle. Selling through a platform buys reach and hands over the customer relationship, the data and part of the margin. It is a legitimate strategy, but it should be chosen deliberately rather than arrived at, and the terms should be read before the first listing. The same discipline of checking a counterparty before committing is what a due diligence check before a deal is for, whether the counterparty is an acquisition target or a company that sells on your behalf. Our guide to digital transformation in retail covers where those systems meet the store.

What Retailers Should Keep Measuring

Four measures answer most of the questions raised by the last five years. The online share of total revenue, tracked quarterly rather than celebrated annually. Contribution margin per order after payment fees, fulfilment and expected returns. Repeat purchase rate by cohort, which is the only honest test of whether the habit survived. And the cost to serve by postcode or region, which is where the quiet losses live.

None of these require new systems. They require the decision to measure the same way for several quarters in a row, which is harder than it sounds when a business is busy. The pandemic period rewarded speed and tolerated sloppiness in unit economics because volume covered it. The period after it does not. An operator who knows the contribution per order, the repeat rate and the true cost to serve can make decisions about assortment, pricing and delivery that a competitor reading only the headline growth figures cannot.

That is the durable lesson. Online retail did not become a different business in 2020. It accelerated, and then it charged for the acceleration through its cost base. The firms still trading well are the ones that priced the new structure correctly and stopped planning against a peak that is not coming back.