Commerce Budgets
Part of Off-site commerce advertising
Auditing the sales population in a UK off-site report
Identify which shoppers and transactions a UK retailer's off-site sales report includes before interpreting attributed outcomes.
Before using a retailer sales number to judge an off-site campaign, identify whose purchases it can include. A report may count only customers whose retailer identity can be linked to an eligible ad exposure. That is a narrower population than everyone who saw the campaign or bought the product. This audit concerns the outcome population; reconciling platform impressions and retailer totals is a separate task.
Ask who is inside the reported population
Ask the retailer for their definition of an eligible customer and transaction. Determine whether the analysis includes linked in-store purchases, online orders, or both. Identify which loyalty, account, or other identifiers are required. Find out if unmatched shoppers are excluded, modelled, or reported separately. Note whether the number is observed sales among eligible customers or an extrapolation, and keep the method description.
Nectar360 reports in-store and online sales among exposed Nectar customers for its off-site campaigns. That example defines a reportable population, but not every shopper or sale is visible. Ask the actual report provider for its implemented rule.
Fix the product and transaction boundary
Agree the SKU or product set before reading the result. Differentiate promoted items from any wider brand or category sales. State whether the figure is order value or units, gross or net of returns and cancellations, and whether stores or channels are omitted. Record the reporting cut-off as later returns or linked purchases can alter a subsequent extract.
Population question / Evidence to retain
- Who could be matched?
- Eligible identity or account definition and any coverage limits.
- Which purchases qualify?
- Store and online channel rules, product set and transaction status.
- What is estimated?
- Any modelling or extrapolation method, labelled separately from observed sales.
- When is the count final?
- Exposure and purchase windows, refresh date and treatment of returns.
Do not use the media audience size as the denominator for a retailer sales rate unless the report establishes that the same people and period are covered. A large gap between exposed people and reportable shoppers can alter how a sales figure is interpreted, even if the underlying campaign delivery was sound.
State what the figure can support
Label the result with its population and product boundary: for example, sales recorded among eligible exposed customers under the retailer's stated rules. A purchase linked to exposure is not, on its own, evidence that the advert caused an extra sale.
If coverage is too narrow or undisclosed for the decision at hand, request a labelled sensitivity view or an alternative measurement design. Keep the original population definition with every exported figure so that a later presentation does not turn a limited outcome into a claim about all UK shoppers.
