- 20%of clicks only on the target button
- 40%of the traffic went to the wrong destination
- 1A/B test to confirm the visitor's intention
- +25%of traffic redirected to the right destination
The context
An insurer offers a hospitalisation cover that many Belgians receive through their employer. The online service used to consult it and file a claim is called My Healthcare. The product page has a simple goal: get the visitor to click the My Healthcare button, which leads to that service.
The monthly report showed a stable click-through rate on that button, around 20%. Not disastrous for a product page. But the page received significant traffic, largely from people already insured, and nobody knew what the remaining 80% were doing.
The finding
First step, trace the page's destination paths in GA4: every link clicked was scanned, and the result surprised the team.
- 20% of visitors clicked My Healthcare, the target button.
- 40% of the total clicked "Log in" in the site header, which opens the general customer area.
- The rest scattered or left the site.
The most clicked link on the page was therefore not its main button, but a secondary navigation link.
The analysis
Read the journey after the click
A product page that feeds the login page in volume is a signal. Either visitors have finished reading and want to manage their contract, or they are looking for something they did not find on the page. Session tracking settled it: once logged in to their customer area, these visitors got lost in the pages.
Several links for the same intention
The visitor wants to log in to check their insurance contract. Only, they do not recognise the name "My Healthcare" as the area dedicated to their insurance.
For the insurer, the distinction is obvious: the My Healthcare button opens the health service, the "Log in" link in the header opens the customer area. For the visitor covered through their employer, the name My Healthcare means nothing. They know they have hospitalisation insurance, they look for somewhere to log in, they pick the "Log in" link.
The hypothesis
Visitors who click "Log in" from the hospitalisation page are actually looking for My Healthcare, but do not recognise the name. If we offer them the choice at the moment of the click, the majority will choose the hospitalisation insurance.
The analysis made this hypothesis very likely. But it meant rewriting a product page and touching a service's vocabulary. Before rolling it out, the team wanted a quantified confirmation. This is the typical case where an A/B test validates an intention, rather than comparing 2 layouts.
The test
A variant was set up in Kameleoon, on the hospitalisation page only. On clicking "Log in" in the header, a popup offered 2 clear paths:
- "My account", to the general customer area.
- "My hospitalisation insurance", to My Healthcare.
Each choice was measured. The popup was never meant to stay: its job was to validate the visitor's intention.
The result
2 visitors in 3 who clicked "Log in" chose "My hospitalisation insurance". The hypothesis was confirmed: most of that traffic had the right need, just the wrong link.
The decision
The popup was removed. It had done its job as a probe, it would have been friction in production. The page and its links were rewritten. The flow of traffic to the general customer area shrank and the target link went from 20 to 45%.
What to remember
1. Identify the CTR of the links on the page. Analysing the most clicked links reveals how visitors behave.
2. Analyse the journey of the abnormal clicks. Do those sessions show a normal engagement rate, or does it reveal a drop-off?
3. An A/B test can act as a probe. When the analysis points to a confusion, confirming the visitor's intention at the moment of the click gives a quantified answer within weeks, without committing to a redesign before being sure.
Tools used
- Google Tag Manager for click tracking with an identifier per link.
- Google Analytics 4 for destination clicks.
- Contentsquare for journey visualisation.
- Kameleoon for the popup variant and choice tracking.
For the full method, see the CRO and behavioural analysis expertise. For the click naming convention, the tracking plan.