TLDR: A funnel journey map is not analysis and not a data pull. It is a manual walk through everything your site lets a customer do, every path, every button, every action. You do it before the audit, because you cannot judge whether add-to-cart is tracked correctly until you know every place a customer can add to cart. The map is the reference truth the audit measures against.
Click to discuss this article with your favourite AI
Ask most teams where a customer adds something to their cart and the answer is instant: the product page. It is such an obvious answer that nobody checks it. And it’s not wrong, but it’s probably badly incomplete.
On one site I audited, add-to-cart fired from the product page, yes, but also from the category grids, from saved wishlists, from a “you may also like” strip on the cart page and from promoted products on the homepage. The product page was the biggest source at about 78%. It was nowhere near the only one. Roughly a fifth of all add-to-cart activity happened somewhere else.
Now imagine your tracking only fires on the product page, which is the default on a lot of setups. Your add-to-cart number is instantly a fifth too low, and here is the dangerous part: it does not look too low. It looks like a number. Nobody questions a plausible figure. You would build a funnel on it, calculate a conversion rate from it and make decisions against it, none the wiser.
You do not find this by staring at reports. You find it by walking the site funnel, and the first ten minutes of that walk will show you more than months of the dashboards ever did.
The map is a walk, not a data pull
The funnel journey map is one of the least technical steps in this whole series, and one of the most important. It is not an analysis and it is not a data export. It is me going through the site by hand with the browser console open and GA4’s DebugView in a second tab, writing down every way a customer can do each thing that matters and watching whether an event fires each time I do it.
Every route to add to cart. Every way to start and complete checkout. Every path into search, every way to save an item for later, every point where the site asks someone to log in. You click every button. You complete each key action by every route you can find. And you do it on mobile as well as desktop, because the mobile site often hides the wishlist behind a menu or drops the “you may also like” strip entirely, and that changes what can even fire.
You are not measuring anything yet. You are building an inventory of what is possible: what a customer can do here, and in how many ways. That inventory is the thing everything after it gets checked against.
Why it has to come before the audit
Here is a number from that same site. On the search results page, over ninety days, there were 45 add-to-cart events against roughly 1.7 million page views.
What does 45 mean? On its own, nothing you can act on, because it has at least three explanations and they do not agree. Maybe there is no quick-add button on search at all, so there is nothing to track and 45 is just noise. Maybe there is a button, customers use it, and the tracking never fired, in which case 45 is hiding a real and busy surface. Or maybe there is a button, customers genuinely ignore it and click through to the product page instead, in which case 45 is exactly right and search simply is not a place people add to cart.
The easy move is the last one. You trust the number, conclude that nobody adds to cart from search and move on. And you might be right. But you might also be reading a broken tag as a fact about your customers, and you would never know, because a true zero and a silent failure look identical in a chart.
You cannot tell them apart from the data. You can only tell by walking the page: looking for the button, clicking it and watching whether anything fires. Is there a button, and does it do anything. That ten-second act is what separates a real behaviour from a broken tag, and it is the whole reason the map comes before the audit.
This is why the map comes first. Everything the audit does depends on knowing what should be there. A tracking number only means something against an expectation, and the map is where the expectation comes from. Skip it, and every low number in your data is ambiguous. You will not know whether you are looking at customer behaviour or a broken tag, and you will quietly assume behaviour, because a small number looks like small demand.
You also see the site the way a customer does
There is a second payoff, and it is a happy accident of doing the walk properly. When you move through the site by hand, completing real tasks by every route, you meet the friction a customer meets. The wishlist that makes you log in before it will save anything. The checkout that will not let you change a quantity without going back to the cart and starting again. You are not analysing this yet, but you are seeing it, and it goes on the map next to the actions it gets in the way of.
By the end you have something no report gives you: a complete picture of what the site lets a customer do, what it makes hard and every surface where a tracked action could be firing. It is written down, it is specific to your site, and it is the reference you will hold everything else up against.
Which is the next step. By the end of the walk you already know which actions fire an event and which stay silent. The audit is where you follow each of those events the whole way through, from the click to the number in the report, and check that it is not just firing but firing correctly, carrying the right detail and set up to answer the decisions you started with. Firing is the low bar. Answering the question is the real one. That is next.
Not sure what your funnel actually looks like?
If your funnel looks clean but you have never actually walked every path a customer can take through your site, there is a good chance it is missing something you would want to know about. Mapping that is where I start an audit.