Shopify shows 17 sales. Google Ads reports 5 conversions. You expect some difference, but not that much.
Or 95% of visitors appear to reject tracking.
Or you inherited a GTM container and nobody is quite sure what half of it does.
Or performance changed and the question is whether the business changed, the tracking changed, or both.
Those are not monitoring alerts. They are questions about whether the measurement itself can be trusted.
There is no useful fixed checklist for every tracking problem. We follow the evidence.
That can mean reviewing GTM, consent configuration, browser and server-side events, platform diagnostics, network requests, dataLayer behaviour, ecommerce events, attribution settings or the differences between the systems you are comparing.
The goal is not to produce a long report proving we looked at things. The goal is to answer the question you actually came with.
We start with the symptom: the mismatch, suspicious number, implementation concern or change you are trying to understand.
Tracking problems are unknowns by definition. We investigate first rather than pretending every problem fits a fixed repair package.
If nothing is wrong, we will tell you. If something is wrong, you will know what it is, why it matters and what needs to happen next.
Sometimes your developer can make the change. Sometimes you will want us to help. The diagnostic does not exist to manufacture a larger project.
A monitoring baseline can stay perfectly stable and still be wrong. Assurance establishes confidence in the implementation and the baseline; Monitoring can then protect that known-good state from future silent failures.
The scale and cause of a tracking problem are unknown before investigation. We'll make the diagnostic conversation easy, then scope any implementation work from what is actually found.
Tell us what you are seeing, what you expected to see and what has already been checked. We'll start there.