Nine shopping feeds key metrics and where the data comes from

What to take away

Most teams start with impressions and clicks because they are easy to pull, then wonder why the numbers never settle. Start with the decision the metric informs instead.
Nine metrics carry most of the weightimpressions, clicks, click-through rate, cost per click, conversion rate, return on ad spend, revenue per order, feed disapproval rate and non-branded share of conversions.
Every figure needs a named source and a date. Platform reporting, your own order system and consent-based analytics will disagree, and that disagreement is information.
Consent and advertising rules shape what you may collect before any dashboard is built. Check the ICO guidance on online tracking before wiring up tags.
Write down which system is the source of truth for each metric. Without that, two reports will always tell two stories.

Why the first metric choice usually goes wrong

Teams tend to pick metrics that are simple to export. Impressions sit at the top of the funnel, so they rise whenever you add products or broaden targeting. That movement says little about whether the catalogue is working.

Pick metrics by decision

Deciding whether to expand a category?

Yes

revenue per order and ROAS

check disapproval rate to fix product data

A better starting point is the decision you need to make. If you are deciding whether to expand a category, you need revenue per order and return on ad spend, not impressions. If you are deciding whether to fix product data, you need the disapproval rate.

The full set of measures and how they fit together is set out in the shopping feeds measurement and reporting guide, which is worth reading before you rebuild a report.

The nine metrics and their sources

Impressions. How often a product was eligible to appear. Source: the platform's own reporting interface. Compare like-for-like periods only.

Nine metrics and their sources

Impressionsplatform reporting interface
Clicksplatform reporting, reconciled with analytics
Click-through ratecalculated, not exported
Cost per clickplatform billing data
Conversion rateyour order system
Return on ad spendfinance system order values
Revenue per orderyour order system
Feed disapproval rateplatform diagnostics tab

Clicks. Sessions attributed to a product listing. Source: platform reporting, reconciled against your analytics.

Click-through rate. Clicks divided by impressions. Source: calculated, not exported. A fall can mean broader targeting rather than weaker titles.

Cost per click. Spend divided by clicks. Source: platform billing data, which is the closest thing to a hard number in the set.

Conversion rate. Orders divided by clicks. Source: your order system, matched to platform clicks where identifiers allow.

Return on ad spend. Revenue divided by spend. Source: order values from your finance system, not the platform's estimate.

Revenue per order. Total revenue divided by order count. Source: your order system. Useful for spotting discount-heavy periods.

Feed disapproval rate. Disapproved items divided by total items submitted. Source: the platform's diagnostics tab. This is the metric most likely to explain a sudden revenue drop.

Non-branded share of conversions. Conversions from generic search divided by total conversions. Source: search term reporting plus your own attribution model.

Where the numbers stop agreeing

Platforms count conversions on their own attribution window. Your order system counts orders. Neither is wrong, but they answer different questions.

Google's guidance on getting the full value from your web and app channels sets out how channel value should be assessed when a customer touches several surfaces before buying. Read it before you decide one platform overstates its contribution.

Consent also changes the denominator. If a share of visitors decline analytics cookies, your on-site conversion rate will look different from the platform's. Record the consent rate alongside the metric so the gap is explainable.

Reporting rules you should not skip

Advertising claims built on feed data are still advertising claims. The ASA and CAP regulatory overview explains who polices marketing communications and what the codes cover. A revenue claim in a campaign needs the same evidence as any other claim.

Keep a short data dictionary. For each metric, record the definition, the system of record, the refresh time and the person who owns it. That document prevents most arguments.

If you publish numbers internally, a shopping feeds reporting dashboard in England shows how to lay out the same metrics for a team that reads them weekly rather than daily.

A short glossary

Attribution window:: the period after a click in which a conversion is credited to that click.
Disapproval rate:: the share of submitted items the platform refuses to serve.
Feed:: the structured product file a platform reads to build listings.
Non-branded conversion:: a conversion from a search that did not include your brand name.
Source of truth:: the single system whose figure you accept when two reports disagree.

Common questions

How often should feed metrics be reviewed?

Weekly for disapproval rate and cost per click, monthly for revenue per order and non-branded share. Daily reviews tend to chase noise rather than signal.

Why do platform conversions exceed my order count?

Attribution windows, modelled conversions and cross-device credit all inflate platform figures. Reconcile at month end rather than daily, and keep the gap documented.

Can I report on clicks without cookie consent?

Aggregate platform click data does not need consent in the same way as on-site tracking, but any analytics tag you add does. Check the tracking guidance before you deploy it.

Which single metric predicts trouble earliest?

Feed disapproval rate. It moves before revenue does, and it points directly at a fixable product data problem.