Conversion and Funnel Tracking
A funnel built to answer every possible question usually answers none of them, because nobody opens it. Ivan Sp builds multi-step conversion funnel tracking scoped deliberately to what a client needs answered, rather than the most granular setup technically achievable. A funnel built to track everything usually produces a report nobody reads.
A Real Funnel, Built to Spec
One property’s funnel tracked page-view events with source, medium, country, city, and device attribution. It added form-abandonment calculated pragmatically, immediate previous-page referrer capture, and enhanced event tagging on a third-party booking tool to attribute bookings back to the page that produced them.
Every piece in that list answered a specific question the client actually asked. Nothing got added purely because it was technically possible to track.
Choosing the Simplest Reliable Method
Form-abandonment tracking illustrates the scoping discipline directly. Field-by-field interaction tracking would show more detail, and it also costs more to build and more to maintain. A simpler calculation, page views minus completions divided by page views, answered the actual question about where visitors dropped off, at a fraction of the setup cost.
The client was asked directly what they needed. Their answer, including skipping a more granular “form started” event, shaped the final scope.
Attribution Through Third-Party Booking Tools
A booking made through a third-party scheduling tool arrives in that tool’s own dashboard with no connection to the page that generated it, unless enhanced event tagging closes the gap. This attribution step turns a booking count into a page-level conversion metric a marketing decision can actually use.
Why Overbuilt Tracking Is a Real Cost
Tracking that gets built and never analyzed is not neutral. It carries an ongoing maintenance cost, since every site change risks breaking an event nobody is checking. Scope gets cut deliberately rather than maximized, keeping the tracking that survives worth the effort of maintaining it.
Funnel Data Feeding Into Reporting
Funnel tracking data feeds directly into ongoing reporting, covered under SEO reporting. It also feeds the conversion-by-service prioritization that directs future content investment toward what actually converts.
Where the Standard Funnel Advice Breaks Down
When did anyone last open the funnel report you paid to have built?
- The usual advice
- Track every step in detail, since granular data can always be aggregated later.
- What holds up in practice
- Every tracked step is a thing that breaks silently and a row nobody reads. The binding constraint is attention rather than storage, and an unread funnel has a real maintenance cost with no return.
Imagine a property with a nine-step funnel covering every interaction on a booking flow. The implementation is technically impressive.
Two steps are read each month. When a booking stops being attributed, finding the broken step among nine takes most of a day. The funnel had never answered a question anyway.
I build the smallest funnel that answers the stated question. Four steps that are trusted beat nine that nobody checks.
Tracking that gets built and never analysed is not neutral. It costs implementation time, maintenance attention, and the debugging hours when something eventually breaks.
A funnel nobody reads is a monthly cost with no return, and it is quietly hiding the one step where you are actually losing people.
Four Things Commonly Assumed About Funnel Tracking
Which of these describes your current funnel setup?
Ivan Sp meets these four assumptions on most measurement engagements.
| Commonly assumed | What actually holds |
|---|---|
| Granular tracking can always be aggregated later. | True of the data and false of the cost. Each step is another surface for silent failure and another row nobody reads. |
| Page views are enough to build a funnel. | Page views miss interactions that never change the URL, and they cannot distinguish a form start from a form abandon. |
| External booking tools cannot be tracked. | Cross-domain measurement and referral exclusions carry the session across. Without them the booking is credited to the tool. |
| A drop-off between steps means a problem at that step. | The drop-off marks where people left. Why they left needs session recordings or usability research. |
| Once built, a funnel keeps working. | A template change, a plugin update, or a booking tool release can break a step without any visible error. |
The step where your customers actually give up is worth finding. Nine tracked steps nobody reads will hide it just as effectively as no tracking at all.
What a Funnel Needs to Work
- A defined question
- The business decision the funnel exists to inform, written before any tag is built.
- Named steps
- Each stage as a distinct event, rather than inferred from page views alone.
- Attribution capture
- Source, medium, and campaign carried through every step.
- Cross-domain handling
- Configuration for any step that happens on an external booking or payment tool.
- Referral exclusions
- Set for those domains, so a handover does not start a new session.
- Exploration reports
- GA4 funnel and path exploration, which read the events once they exist.
- The simplest reliable method
- The least complex tracking that answers the question, rather than the most granular.
- A review date
- A scheduled check on whether anyone is reading the funnel at all.
A Worked Example: Nine Steps Against Four
Consider two funnels built for the same booking flow. The comparison below is illustrative, written to show what scoping changes. No client engagement is described here.
| Measure | Nine-step funnel | Four-step funnel |
|---|---|---|
| Events configured | 9 | 4 |
| Steps that answer a stated question | 4 | 4 |
| Reports opened per month | 2 | 4 |
| Time to locate a broken step | Most of a day | Under an hour |
| External booking attributed correctly | No | Yes |
| Maintenance surface | 9 things that can break | 4 things that can break |
I scope to the question rather than to the platform’s capability. Both funnels captured the same four useful steps, and one buried them under five more.
An overbuilt funnel is a nine-course tasting menu on a Tuesday. Everything is technically correct and nobody has time for it.
The fifth row is the real cost. Effort spent on steps nobody reads is effort not spent closing the attribution gap that mattered.
How Funnel Tracking Gets Built
I write the question down before opening any tag manager, since the question decides the steps.
- Write the business decision the funnel is meant to inform, in one sentence.
- Map the minimum set of steps that would answer that decision.
- Check whether each step changes the URL, since steps that do not need event tracking.
- Configure each step as a named event through Google Tag Manager rather than as a page view.
- Carry source, medium, and campaign values through every step of the flow.
- Configure cross-domain measurement for any step on an external booking or payment tool.
- Add referral exclusions for those domains, so the handover does not start a new session.
- Verify every step in Tag Manager preview mode and in GA4 DebugView before publishing.
- Build the funnel exploration report and confirm the step order matches the real flow.
- Set a review date to check whether the funnel is being read, and remove steps nobody uses.
How This Engagement Is Scoped
I scope funnel work by the number of steps that cross a domain boundary.
- What arrives
- A written measurement question, a configured funnel, verification evidence, and a review date.
- Typical duration
- Two to three weeks for a single flow with one external tool.
- What the client supplies
- GA4 and Tag Manager access, plus admin access to any external booking tool.
- Included
- Question definition, event configuration, cross-domain setup, verification, and documentation.
- Not included
- Server-side tagging infrastructure, and consent platform licensing.
- What changes the estimate
- Several flows, several external tools, or a booking tool with no tagging support.
- How it is quoted
- After a short scoping call, once the flow and external tools are listed.
Tools and Primary Sources
The references below document the configuration. Each opens in a new tab.
- GA4 funnel exploration documents how funnel and path exploration read the events a container sends.
- GA4 cross-domain measurement documents the configuration that keeps a session intact across two domains.
- Google Tag Manager is where each funnel step is defined as a tag and a trigger.
- GA4 DebugView confirms each step fires with the parameters expected.
Check Your Own Attribution
I suggest a check that exposes the most expensive funnel fault.
- Open GA4, then Reports, then Acquisition, then Traffic acquisition.
- Look for your booking, payment, or scheduling tool in the source list.
- If it appears, your conversions are being credited to it rather than to the real source.
- Open your funnel report and note how many steps you actually read.
A booking tool appearing as a traffic source is a cross-domain configuration gap. The gap makes organic search look unproductive, which is how budgets get cut wrongly.
Common Questions
How many steps should a conversion funnel track?
As many as the client’s actual decision-making needs, typically three to five key steps. A funnel with more than that usually tracks detail nobody ends up acting on.
What is the simplest way to measure form abandonment?
Page views minus completions, divided by page views, is a pragmatic method most clients find sufficient. It avoids tracking every individual field interaction while still answering the core drop-off question.
Does funnel tracking need to be rebuilt if a site’s design changes?
Often yes, at least partially, since a redesign can move or rename the elements tracking depends on. A funnel audit after any major site change catches breakage before it goes unnoticed for months.
Should every interaction on a page be tracked?
No. Each event is another thing that can break silently and another row nobody reads. The right number is the smallest set that answers the questions the business actually asks.
Discuss This Work
The first call is a free 30-minute consultation with no pitch attached. You receive an honest assessment of where the site stands and what can realistically be built.