- Published
- Byline
- GoHighLevel Partners Editorial
- Filed under
- GoHighLevel Automation
The report reflects the events you collect and the order you declared, and both are things you configured rather than things that are true.
Google Analytics funnel exploration: reading where people stop
A funnel exploration is a report that puts a sequence of steps in order and shows how many people reached each one. It answers a question no other standard report answers well: not how many visitors you had, but where the ones who were trying to buy gave up. For an agency that is usually the most actionable single view available, because a step losing most of its traffic is a page you can rewrite this afternoon.
It is also easy to build in a way that produces a confident wrong answer. The report reflects the events you collect and the order you declared, and both of those are things you configured rather than things that are true.
The report only knows what you send it
Every step in a funnel exploration is defined by an event. If the event does not exist, the step cannot exist, and the most common cause of a nonsensical funnel is not the report but the collection underneath it. Before building anything, confirm the tag is present on every page in the sequence — tracking code that is not firing is usually a matter of code placed in the page header being on one step and not the next, which produces a funnel showing a catastrophic drop at exactly that point.
The second common cause is using a page view as a step when several steps share a page or a page is reachable out of sequence. A page view is a coarse event, and where the meaningful moment is an interaction rather than a load, send an event for the interaction instead. tracking clicks that leave the site is the same principle applied at the exit.
Open and closed, and why the choice changes the numbers
| Closed funnel | Open funnel | |
|---|---|---|
| Who is counted | Only people who completed step one | Anyone, entering at any step |
| Answers | How a defined journey performs | Where a page sits in real behaviour |
| Use for | A campaign with one entrance | A site people arrive at from search |
| Risk | Undercounts people who joined midway | Later steps can exceed earlier ones, which looks broken |
Neither is more correct. A closed funnel is right for a linear campaign and an open one for a site where entry points vary — which is precisely the difference between a funnel and a website as containers. Pick deliberately and say which you used when reporting the number, because the two produce different conversion rates from identical data.
Order, and the trap of enforcing one that does not exist
The report will happily enforce a sequence your visitors do not follow. If people commonly view the pricing page before the features page, a funnel declaring the reverse shows a large drop that is an artefact of your assumption. Look at the actual paths before fixing the order, and treat an implausible drop as a hypothesis about the report before it is a hypothesis about the page.
Where the sequence genuinely branches, several shorter funnels usually beat one long one. where leads drop out of a quiz funnel is the clearest case: a quiz that routes people down different paths cannot be modelled as one line, and forcing it into one hides the branch that is actually failing.
Elapsed time is the underused column
Beyond how many people moved on, the report can show how long they took. A step that people complete quickly is understood; one where the median time is long is one where they are hesitating, reading, or leaving to check something. That is a different problem from a step people abandon outright, and it calls for different changes.
Long times on an early step often point at load rather than content, which is where a page speed score is worth checking before rewriting any copy. Long times on a form step usually mean it is asking for something people need to go and find, which is the argument splitting a form across steps makes.
Breakdowns are where the finding usually is
An aggregate funnel tells you a step is weak. Splitting it by device, by source, or by campaign tells you for whom, and that is normally the actionable version. A checkout that converts acceptably overall and badly on phones is a layout problem; one that converts badly for a single traffic source is a targeting or expectation problem.
For source breakdowns to mean anything, UTM parameters have to be applied consistently and survive the whole journey. And be careful when the journey crosses a boundary: one funnel on several domains splits sessions unless the property is configured for it, and a funnel built across that boundary will show a drop where a visitor simply moved between hosts.
What the report cannot tell you
Three limits are worth stating plainly. Where a consent banner gates analytics, the report describes visitors who agreed rather than all of them, so absolute counts understate and comparisons across the install date are not like-for-like. Data can be sampled or thresholded on smaller segments, so a narrow breakdown may be directionally useful and numerically approximate. And the report describes correlation within a sequence — it shows where people stopped, never why.
That last point is the important one. A funnel exploration generates hypotheses; it does not test them. Testing is split testing a funnel step, and ending a test honestly is declaring a split test winner. Treating a drop-off number as a verdict is how a team spends three weeks rewriting a page that was never the problem.
Building one that survives review
- List the real steps first, on paper, before opening the report.
- Confirm an event exists for each, and that it fires once rather than on every render.
- Choose open or closed deliberately, and record which you chose.
- Confirm the final step is the outcome you care about — usually the thank-you step rather than a click.
- Check the totals against a source outside analytics: the number of orders, the number of records created. A funnel that disagrees with the CRM is measuring something else.
- Break the funnel down by device and by source before drawing any conclusion.
Point five is the one that catches the most errors, and it is the connection to pipeline conversion: the analytics view and the pipeline view of the same week should be close, and when they are not, the gap itself is the finding.
Frequently Asked Questions
What is a funnel exploration in Google Analytics?
A report that places a sequence of steps in order and shows how many people reached each one. It answers where the visitors who were trying to buy gave up, which is a different and usually more actionable question than how much traffic arrived.
What is the difference between an open and a closed funnel?
A closed funnel counts only people who completed the first step, so it measures a defined journey. An open funnel lets anyone enter at any step, so it measures where a page sits in real behaviour. Closed suits a campaign with one entrance; open suits a site people find through search.
Why does my funnel show a huge drop at one step?
Check the collection before the page. The most common cause is the tag missing from that step, which produces an apparent collapse exactly where the tag ends. The second is a declared order visitors do not actually follow, which the report will enforce regardless of reality.
Should I use page views as funnel steps?
Only where a page maps cleanly to one step. Where several steps share a page, or a page can be reached out of sequence, a page view is too coarse and the funnel becomes misleading. Send an event for the meaningful interaction instead.
Why do the funnel numbers disagree with my CRM?
Usually consent gating, cross-domain session splits, or a final step defined as a click rather than a completed outcome. Compare the funnel total against orders or records created as a routine check — when they diverge, the size of the gap is itself the finding.
What does a long time between steps mean?
That people are hesitating rather than leaving, which is a different problem from outright abandonment. Early in a funnel it often points at load time; on a form step it usually means you are asking for something the visitor has to go and find before continuing.
Can a funnel exploration tell me why people dropped off?
No. It shows where they stopped, never why. It is a generator of hypotheses rather than a test of them, and treating a drop-off figure as a verdict is how a team spends weeks rewriting a page that was not the problem. Test the hypothesis before acting on it.
We build your GoHighLevel account for you
CRM, pipelines, funnels, SMS and automation — configured for your business, not left for you to figure out.
- Setup (one-time)$850
- Ongoing (per month)$38.80
No obligation quote · reply within 1 business day
This site is independently operated and is not officially affiliated with, endorsed by, or sponsored by HighLevel LLC. We do not receive referral commissions — this is our own flat-fee implementation service.