By Nick Quirk
October 12, 2026
8 min read
Most websites lose the majority of their visitors in two or three predictable spots, not evenly across every page. Analyzing the user journey means rebuilding the path people take from first click to conversion, then measuring how many survive each step so you can see exactly where the leak starts. The method below uses GA4 funnel reports, session recordings and a bit of segmentation, and it works whether you sell products, book appointments or collect leads.
What a Drop-Off Point Actually Is
A drop-off point is any step where the percentage of users continuing to the next step falls sharply compared with the steps around it. If 100 people land on a product page, 42 add to cart, 38 start checkout and 11 finish, the checkout stage is the leak, not the product page.
The number alone is not the insight. A 70% drop between a blog post and a service page might be completely normal, while a 25% drop between “shipping details” and “payment” is a serious problem worth a week of work.
Context comes from three comparisons: the step before it, the same step last quarter, and the same step on a different device. Without at least one of those, you are guessing.
Map the Journey on Paper First
Open a blank document before you open analytics. Write the five to seven steps a customer actually takes, in order, using the URLs or events that represent them.
A typical lead-gen path looks like this:
- Entry: organic landing page, paid ad click, or direct homepage visit
- Orientation: scroll past the hero, read the offer, check pricing
- Evaluation: visit a services page, case study, or reviews section
- Intent: click a CTA, open the contact form, start a quote tool
- Action: submit the form, complete the booking, call the number
E-commerce usually adds a cart and a multi-step checkout, which is where the sharpest losses tend to show up. Writing the map first keeps you from fishing through reports and calling random numbers “insights”.
Building a Funnel Exploration Report in GA4
GA4’s Funnel exploration is the fastest way to turn that paper map into numbers. In the Explore section, create a new Funnel exploration, then add one step per stage using page paths or events such as view_item, begin_checkout and purchase. Google’s own funnel exploration documentation covers the setup options in detail.
Two settings change what you see more than anything else. Switch between open and closed funnels: a closed funnel only counts users who entered at step one, which is stricter and usually more honest for paid traffic.
Then turn on Show elapsed time. A step that takes users an average of 90 seconds when the previous step took 20 is almost always a confusion problem, not a desire problem.
Your funnel is only as accurate as your event tracking, so confirm that form submissions, call clicks and add-to-cart actions actually fire before you trust a single percentage. It also helps to understand how bounce rate and engagement rate work in GA4, since an “engaged session” threshold of 10 seconds can mask a lot of early abandonment.
The Four Leaks That Show Up Most Often
Across audits, the same handful of stages account for most lost conversions:
- The first screen: visitors leave within 3 to 8 seconds because the headline does not match the ad or the search query they used.
- The pricing or service detail page: people want a number, find “contact for pricing”, and go back to Google.
- The form or checkout start: every extra required field costs completions, and surprise shipping costs are the single most cited reason for abandonment, according to Baymard Institute’s research, which puts the average documented cart abandonment rate near 70%.
- The confirmation or payment step: slow processing, failed validation, or a payment method the customer does not use.
If the leak sits at step three or four, the fix is usually mechanical. Our walkthrough on reducing friction in your checkout process covers the field-by-field changes that recover the most revenue per hour of dev time.
Segment Before You Draw Conclusions
A single blended funnel hides more than it reveals. The same checkout can convert at 4.1% on desktop and 0.9% on mobile, and the blended 2.2% tells you nothing about which one to fix.
Run your funnel again with these segments applied one at a time:
- Device category: mobile versus desktop, since mobile drop-off is often a layout or thumb-reach issue
- Channel: organic, paid search, paid social, email and direct behave very differently at the top of the funnel
- New versus returning: returning users usually convert at two to three times the rate of first-time visitors
- Landing page: one underperforming entry page can drag an entire channel’s numbers down
Paid social traffic deserves its own funnel entirely, because the intent level is lower and the drop-off happens earlier. If you are weighing platforms, the comparison of TikTok Ads versus Instagram Reels Ads explains why identical landing pages can produce very different post-click behaviour.
Clean source data makes all of this possible, which is why consistent UTM parameters matter more than most teams realise. Mismatched tags turn half your campaign traffic into “direct” and quietly ruin the segmentation.
Finding the Why Behind the Number
Analytics tells you where people left; it rarely tells you why. Pair every drop-off you find with at least one qualitative source before you start changing things.
Session recordings filtered to users who reached the problem step and then exited are the quickest win. Watch ten of them and patterns appear fast: rage clicks on a non-clickable element, repeated scrolling past a form field, a validation error that never clears.
Scroll maps answer a different question, which is whether people ever saw the thing you want them to click. If 60% of mobile users never reach the pricing table, the pricing table is not the problem, its position is.
Exit-intent surveys with one question (“What stopped you from booking today?”) typically collect enough responses in a week to confirm or kill a hypothesis.
Prioritising Fixes by Expected Impact
Not every leak is worth plugging. Rank candidates by the number of users affected multiplied by the size of the drop, then divide by implementation effort.
A step with 8,000 monthly users and a 30% drop is worth far more attention than a step with 300 users and a 70% drop, even though the second number looks alarming. Work top-down from traffic volume.
Test changes one at a time where you have the traffic for it. Our breakdown of A/B testing versus multivariate testing explains the sample sizes each approach realistically needs, and most small-business pages only have enough traffic for the simpler option.
Reviewing the Funnel on a Schedule
Journeys shift as your traffic mix shifts. A seasonal campaign, a new paid channel, or a batch of posts that rank for broader queries will all change who arrives and how ready they are to act.
A monthly 30-minute review of the same funnel, same segments, same date comparison catches regressions before they cost a quarter. Content strategy affects this too: the balance between evergreen and trending content determines whether your top-of-funnel visitors have any purchase intent at all.
Businesses serving a defined geographic area often see different patterns again, which is one of the benefits of working with a local agency that understands how local search behaviour feeds the funnel.
Frequently Asked Questions
What Is a Normal Drop-Off Rate Between Funnel Steps?
Losing 50% to 70% between the landing page and the next meaningful step is common, while losses above 30% inside a checkout or form flow usually signal a fixable problem. Compare each step against the steps immediately before and after it rather than against an industry average.
How Much Traffic Do I Need Before Funnel Analysis Is Useful?
Roughly 300 to 500 users entering the funnel per month gives you enough data to spot a genuine pattern rather than noise. Below that, lean more heavily on session recordings and direct customer feedback, since small-sample percentages swing wildly week to week.
Can I Analyze the User Journey Without GA4?
Yes, though you will need two tools instead of one: a product analytics platform such as Matomo or Mixpanel for step counts, plus a recording tool for behaviour. GA4 remains the most common choice because funnel exploration and segmentation are included at no cost.
How Long Should I Wait Before Judging a Fix?
Give any change at least two full business cycles, typically 14 to 28 days, before calling it a win or a loss. Shorter windows get distorted by day-of-week effects, payday timing and campaign launches that have nothing to do with your edit.
Should I Fix the Biggest Drop or the Earliest Drop?
Fix the earliest significant drop first, because every user you keep there flows into every step below it. A 10% improvement at the top of the funnel compounds through the remaining stages in a way that a late-stage fix cannot.
Get a Second Pair of Eyes on Your Funnel
If your funnel report is showing a drop you cannot explain, a focused audit of tracking setup, segments and on-page behaviour usually surfaces the cause within a few days. Talk to SEO Quirk about where your visitors are leaving and what it would take to keep them.