Start with the decision, then choose the metric
A weekly trading meeting can contain dozens of charts and still leave the team without a clear priority. The problem is often the relationship between a number and a decision. Before adding another KPI, ask what you would do differently if it moved. If the answer is unclear, the metric may belong in a supporting report rather than at the centre of the meeting.
For a fashion business, a useful starting point is the customer journey: attracting relevant shoppers, helping them choose confidently and giving them a reason to return. Each stage needs a small number of consistently defined measures, supported by the detail required to investigate a change.
Read acquisition and conversion together
A rise in traffic can look positive while conversion falls. That does not automatically mean the storefront is getting worse. A collection launch might bring more first-time mobile visitors with a different level of intent. Compare the channel, device and customer mix before deciding which team owns the problem.
Start with qualified traffic and commercial outcomes, then investigate the journey between them. If product views are stable but add-to-cart activity falls, look at product-page confidence and measurement quality. If checkout starts are healthy but purchases fall, the next investigation belongs further down the journey.
Keep revenue definitions consistent
Gross sales, net sales and attributed revenue answer different questions. A campaign platform may include sales within its attribution window that are also claimed elsewhere. Your storefront may apply refunds, taxes and shipping differently from your analytics. Adding those reports together does not produce a trustworthy commercial total.
Agree on the source of truth for revenue and document the definition. Keep currency, timezone and reporting period consistent. When a number differs between systems, record the reason where possible and make the remaining uncertainty visible.
Bring repeat purchase and returns into the picture
Acquisition is only part of growth. A promotion that attracts a large first-time audience may perform differently once returns and second purchases are considered. Review cohorts over a period that fits your purchase cycle, and avoid comparing customers with very different time to return.
Returns deserve a clear definition too. Order-level and item-level return rates tell different stories, especially when a customer buys several sizes. Where the data supports it, investigate collection and variant patterns before treating a store-wide average as a diagnosis.
End the review with an owned next step
A useful priority contains five things: the observed change, its likely business relevance, the supporting source, the uncertainty and a next step with an owner. The action might be an experiment, a content improvement or simply fixing measurement before deciding.
Keep the list short enough to act on. At the next review, return to the previous priorities and ask what was learned. This is how reporting becomes a working rhythm rather than a weekly presentation.
