9 reports you already pay for and almost nobody opens

Before I buy anything, what can I already check with the tools I have?

Most attribution problems get diagnosed by buying something. A surprising share of them can be diagnosed by opening a report you already have, in a tool you already pay for, that nobody on the team has ever clicked.

These nine are the ones that earn their reading. I have named the report and what it answers rather than a click path, because platform menus move constantly and a listicle full of exact menu trees is wrong within a quarter.

1. Meta: the Purchase event diagnostics, specifically redundant purchase events

Most stores send the purchase twice on purpose, once from the browser pixel and once server-side from the order webhook, because the browser event is lost whenever an ad blocker, tracking protection or an off-domain checkout gets in the way. What makes that redundancy safe rather than harmful is that both events carry the same event ID, which is the key the platform uses to collapse them into one.

This diagnostic is Meta telling you whether that is working. If it reports redundant purchase events, the two sides are not agreeing on the ID and every sale is arriving as two. Check it before concluding anything else about your numbers, because it inflates everything downstream.

2. Meta: event match quality, read per event and not per account

Match quality is how well Meta can recognise the people in the events you send. The account-level average is close to useless, because it is dominated by whichever event fires most and that is almost always PageView, which matches easily.

Read the score for Purchase. A healthy overall number sitting on top of a poor Purchase score is the normal shape, and Purchase is the only one that matters for attributing revenue.

3. Google Ads: the conversion actions list, with the primary and secondary column

This one report explains most cases of a Google conversion count that looks like a multiple of your order count.

The Conversions column sums every action marked primary. If the account tracks purchases, add-to-cart, newsletter signups, phone calls and a PDF download, all five are inside the number you have been comparing against orders. Nothing in the interface warns you, because from Google's side that is correct: those are the outcomes the advertiser said mattered, and Smart Bidding optimises toward their sum. Open the list and the multiple usually resolves in one screen.

4. Google Ads: Model Comparison, and know what it can still compare

It compares data-driven attribution against last click. That is the entire list as of today: the first-click, linear, time-decay and position-based models were removed and the conversion actions using them were migrated to data-driven, so much of what is written about this report describes options that no longer exist.

Narrower, but still worth ten minutes for one question: how much credit each campaign gains or loses when the model stops handing everything to the final interaction. Upper-funnel campaigns that look weak under last click are exactly where the difference shows up.

5. Google Ads: Path Metrics, which is how you should choose your attribution window

Path Metrics reports how many days and how many ad interactions passed before your conversions happened. That is the measurement almost nobody makes before setting a window, and it is the only honest basis for the decision.

Plot the distribution and find where it flattens. If most conversions close within two days, a long window is mostly collecting coincidences, people who saw an ad three weeks ago and would have bought anyway. If your buyers research for a month, a short window is discarding influence that genuinely happened. Either way the number comes from your account rather than from a default someone else picked.

6. Google Ads: the offline conversion upload diagnostics

If anything in your stack uploads conversions to Google, this is where Google tells you what it did with them. It is more direct than the equivalent on Meta: the status is a literal verdict and the alerts name the specific errors.

Four of those error names mean duplicate, including a repeated order id. A run of those is an idempotence failure somewhere upstream, which usually means order webhooks are retrying and nothing is recording that the sale was already sent. Zero rows is also a verdict, and a useful one: it means nothing is uploading at all.

7. GA4: Conversion Paths

Conversion Paths shows the actual sequences of touchpoints people went through before converting, rather than a single winning channel.

It is worth opening precisely because GA4's acquisition reports do the opposite. They credit the last non-direct click, which gives one channel the whole order and every other channel nothing. Seeing the real paths next to that is the fastest way to understand why GA4 and your ad platforms can never agree even when every underlying event is captured perfectly.

8. GA4: Traffic acquisition and User acquisition, open side by side

This is less a report than an exercise, and it takes two minutes.

Traffic acquisition uses session-scoped dimensions, so it tells you which channel brought the session in which the purchase happened. User acquisition uses first-touch dimensions, so it tells you which channel first brought that person to the site, possibly months earlier. Put them next to each other and you are looking at one property giving two different answers to what your team thinks is one question. Add the Advertising section, which applies the property's own attribution model, and it is three.

Two people quoting GA4 at each other in a meeting are frequently quoting different reports, and this is how you find out.

9. Shopify: sessions and sales attributed to marketing, plus the model it uses

Every plan shows sessions attributed to marketing, so traffic by channel is always available. Switching the attribution model and seeing sales credited to a channel sit behind a higher plan tier, and those tiers get renamed periodically, so check your own plan rather than an article.

The part worth internalising is the default: Shopify credits the last referral source before the purchase. That is a third model, different from your ad platforms and different from GA4's, and it is why the store dashboard and Ads Manager disagree about which channel closed the same sale.

What these nine cannot do

They will tell you whether something is wrong. Between them you can find out if your events are duplicating, whether your conversion count contains things that are not purchases, what your real buying cycle is, and whether a platform's claims exceed the orders you took.

What none of them will do is stitch one specific order back to the click that produced it, or keep that reconciliation current without a person redoing it every week. That is a different job and it is the one worth paying for, if and only if these nine have told you there is something to reconcile. Running them first is not a delay. It is how you avoid buying a dashboard to solve a tagging problem.