One trend, many percentages
It is common to read that online shopping now accounts for a certain share of retail sales, and then find a different figure in another report. Both may be honest. The share depends on what is counted in the top and bottom of the fraction, and small choices there can change the answer substantially.
Understanding those choices matters more than memorising any single percentage, because the figures are revised, restated and redefined over time.
What goes into the numerator
The “online” part of the fraction usually means sales ordered over the internet, but where the boundary sits varies:
- Ordering channel. Some definitions count any order placed online, even if the goods are collected in a shop. Others count only home delivery.
- Devices. Orders placed through apps and mobile browsers are usually included, but orders by phone or message may not be.
- Business type. Some figures cover only retailers that sell mostly online, while others also count the web sales of traditional shops.
- Cross-border purchases. Whether a purchase from an overseas seller is included depends on the source.
What goes into the denominator
The total that online sales are compared against is just as variable. National statistics bodies often define “retail” in a specific way that excludes some categories. Common exclusions or inclusions include:
- Motor vehicles and fuel.
- Food service, such as restaurants and takeaways.
- Travel, tickets and other services.
- Financial products and utilities.
A share calculated against all consumer spending will usually look smaller than one calculated against a narrow set of goods that are easy to ship. Because the categories that move online most readily, such as books, electronics and clothing, are only part of total spending, headline shares can feel at odds with people’s everyday experience.
Category differences hide the average
A single national figure blends very different shopping behaviours. Some goods are routinely bought online, while others, such as fresh groceries, furniture or items people want to inspect first, may follow different patterns. Habits for groceries also differ from habits for gifts or electronics. Reporting one average can make it seem that everyone is shifting at the same pace when the picture is actually uneven.
Services add another layer. Ordering a ride, booking a hotel or arranging a delivery meal is a digital transaction, but whether it counts as shopping depends on the report. Booking data for travel, for example, tells only part of the story, and food delivery is often treated separately from goods retail.
Habits change in steps, not smooth lines
Shopping behaviour does not typically move along a neat curve. Events such as shop closures, new delivery options, changes in prices or shifts in work patterns can produce sudden jumps. Some of those jumps partly reverse, as people return to earlier routines, while others persist. A sharp rise during a particular period should not automatically be read as a permanent change, and a later dip should not automatically be read as a collapse.
Many people also mix channels. They may research a product online, try it in a shop and buy it through an app, or the reverse. When a purchase involves several steps, deciding which channel gets the credit is a judgement call, and different sources make different judgements.
Where the data comes from
The main sources each see a different slice:
- Government surveys of retailers. These typically ask businesses to report sales, sometimes split by channel. They depend on businesses classifying their own sales and on the sample of businesses chosen.
- Payment and transaction records. These show real spending but only for the cards, banks or processors included, and they may not distinguish online from in-person purchases cleanly. The strengths and gaps of transaction records are a topic in their own right.
- Consumer surveys. These ask people how often they buy online. Answers rely on memory and on how the question is worded, and such surveys may over- or under-represent certain groups.
- Company announcements. Firms may report their own growth, which is informative but selective and not always comparable.
Payment systems also matter. Wider use of cashless payments makes more spending traceable, so part of any apparent growth in measured online activity can come from better recording rather than a change in behaviour.
Value, volume and prices
Share-of-sales figures are normally based on money spent, not on the number of items or orders. If prices rise across the board, the value of sales can climb without any change in how many goods are bought. If online prices rise or fall at a different rate from shop prices, the share can shift for that reason alone. Returns are another factor: online orders are often returned more than in-person purchases, and whether figures are measured before or after returns can change the picture.
Share also moves when the total changes. If overall spending falls, the online share can rise even though online sales are flat. This is why the direction of the percentage is not enough on its own; the underlying amounts matter.
Reading a growth story sensibly
A few questions can keep a shopping story in perspective:
- What exactly counts as online, and what counts as retail?
- Is the figure a share, a total amount or a growth rate, and over what period?
- Does it adjust for price changes?
- Which categories are included?
- Who collected the data, and from whom?
Online shopping has become a normal part of life for many people, and few doubt the broad direction of travel. The details, however, such as how large the share is, how quickly it is changing and how it differs between groups and places, are much more sensitive to definitions than a single headline number suggests. Comparing two reports is only meaningful when their definitions match, and often they do not.