Testing an ecommerce benchmark for population, geography and definition before you use it

What to take away

Testing an ecommerce benchmark for population, geography and definition before you use it means checking the source against a fixed reference population. No England-wide platform conversion, revenue or return benchmark was found in the official sources reviewed on 6 September 2026.

No England-wide platform benchmark passed all passport fields: the ONS Retail Sales Index is Great Britain only, platform documentation is global, and supplier dashboards omit England eligibility.
A broad retail series, supplier dashboard or global survey cannot be relabelled as an England ecommerce-platform norm.
A benchmark can describe a reference distribution but cannot show that a platform or campaign caused an outcome.
Keep rejected records in a dated exclusion log so a broad figure cannot return without its original limitation.
Platform dashboards define conversion differently: Shopify uses sessions, WooCommerce uses visitors, BigCommerce and Adobe Commerce use sessions, and Google Analytics 4 uses events.

Define the platform, population and event first

Write the decision and reference population before searching. For an England ecommerce platform benchmark, specify the platform (Shopify, WooCommerce, BigCommerce or Adobe Commerce), merchant eligibility (plan, country, channel), the event (order placed, session, checkout reached) and the reporting period.

An example: an England-established seller, its own physical consumer goods, direct online store, adult UK buyers, a fixed order-acceptance event and a specified reporting period.

State whether the measure concerns operational reliability, observed orders, settled cash, accounting revenue, net contribution or causal impact.

Shopify's UK plans as of 2026 are Basic at £25 per month, Shopify at £79 and Advanced at £299. A benchmark that mixes Basic and Plus merchants has a different population than one that separates them.

WooCommerce is free software, but hosting typically costs £10 to £50 per month, and that cost affects who uses it.

Evidence passport fields

Publisher and extraction date
Geography and merchant population
Channel, goods, buyer eligibility
Event, numerator, denominator, unit
Period, timezone, revision status
Currency, VAT, accounting basis
Collection method, sample, uncertainty

Create an evidence passport for every candidate source:

publisher and field or extraction date;
geography and merchant population;
channel, goods and buyer eligibility;
exact event, numerator, denominator and unit;
period, timezone and revision status;
currency, VAT and accounting basis where money appears;
collection method, sample coverage and uncertainty;
exclusions, missingness and definition version;
platform plan, installed apps and theme;
payment gateway and checkout configuration.

The ONS quality definition treats relevance, accuracy, timeliness and comparability as distinct dimensions. Its approach helps expose gaps, but applying the questions does not make a private benchmark an official statistic.

Reject attractive mismatches

ONS publishes Retail Sales Index internet sales for Great Britain by store type, month and year. It is useful economic context within that scope. It does not identify England merchants, software supplier, session denominator, accepted orders, platform fees or contribution. It therefore cannot supply a platform conversion or commercial target.

Reject a supplier benchmark when account eligibility, plan, country, channel mix, event definition or sample construction is missing. Shopify's analytics fields reference, for example, shows that related rates use sessions, product views or checkout-reaching sessions as different denominators. That same reference fails the passport field 'geography and merchant population' because it covers global platform metrics, not England merchants.

Public documentation for WooCommerce's analytics reports conversion rate using visitors, while BigCommerce and Adobe Commerce use sessions. Google Analytics 4 measures conversions as events, not orders. A percentage without the original definition is unusable.

Do not average incompatible observations. A larger sample does not repair the wrong population, and a newer date does not fix a different event.

Build an internal reference carefully

When no external comparison passes, freeze a merchant-specific definition and establish an internal reference period. For Shopify, use the analytics fields reference to freeze the denominator. For WooCommerce, use its reports. Record catalogue range, stock policy, geography, device coverage, consent design, channel activity, price, fulfilment conditions, platform plan and app stack. Mark outages, releases, campaign changes and definition breaks.

Compare like periods only when the decision owner accepts the remaining differences. Show the number of eligible cases, missing records, revisions and uncertainty. Call the result an internal baseline, not an England benchmark.

For money, reconcile orders, captures, refunds, disputes, settlements and the ledger first. The Shopify Payments activity report warns that its balance activity is not an accounting revenue statement. The accountant and tax adviser must approve currency, revenue and VAT treatment.

Keep causal claims outside descriptive benchmarking

HM Treasury's QPIE guidance explains the need for a credible counterfactual when attributing causal impact. Its subject is policy evaluation, not merchant performance, but the evidence boundary still applies.

A Shopify dashboard benchmark describes Shopify merchants, not the impact of switching to Shopify. The research lead should record accepted, rejected and unresolved sources with reasons. Leave the decision threshold blank. Hold the comparison if geography, population, denominator, money basis or uncertainty cannot be aligned, and repeat the search when a source definition or merchant operation changes.

Keep the rejected records. A dated exclusion log should preserve the URL, page title, access date, failed passport field and reviewer. That evidence prevents a broad UK or global figure from returning later without its original limitation and makes the next research update faster.

Before you act

Define the decision and reference population before searching.
Build an evidence passport for every candidate source.
Reject sources missing account eligibility, plan, country or event definition.
Do not average incompatible observations or repair the wrong population.
Reconcile orders, captures, refunds, disputes, settlements and the ledger first.
Record accepted, rejected and unresolved sources with reasons.
Check the platform plan, app stack and payment gateway for each source.

Common questions

Why can the ONS Retail Sales Index internet sales not serve as an England platform benchmark?

It covers Great Britain by store type, month and year, so it is useful economic context within that scope. It does not identify England merchants, software supplier, session denominator, accepted orders, platform fees or contribution, so it cannot supply a platform conversion or commercial target.

What should a merchant do when no external comparison passes?

Freeze a merchant-specific definition and establish an internal reference period. Record catalogue range, stock policy, geography, device coverage, consent design, channel activity, price and fulfilment conditions. Mark outages, releases, campaign changes and definition breaks, and call the result an internal baseline rather than an England benchmark.

Why does a benchmark not prove that a platform or campaign caused an outcome?

A benchmark can describe a reference distribution, but causal attribution needs a credible counterfactual. HM Treasury's QPIE guidance explains that requirement, and although its subject is policy evaluation rather than merchant performance, the evidence boundary still applies to causal claims.

Why do Shopify and WooCommerce conversion rates differ?

Shopify's analytics fields reference uses sessions as a denominator. WooCommerce's analytics reports use visitors. BigCommerce and Adobe Commerce use sessions, while Google Analytics 4 uses events.