Dead Clicks: When Less Makes More

Originally posted: August 24, 2020

Last updated: September 5, 2020

Introduction

Dead clicks are clicks on a website element that don’t result in any action being taken by the browser. For example, say a user clicks on an unlinked promo graphic. They likely expected to either get directed to a details page with more information about that promo or to get the promo applied to their order. However, nothing happened since the graphic had no link. That user may think the website is broken, becoming distrustful and less inclined to purchase. Dead clicks can indicate instances of confusion, mismatches in user expectation, or enhanced user focus.

We use FullStory to do regular behavioral analysis and monitoring of user behavior on client websites. FullStory has a neat dashboard feature that shows you the top dead click areas on a site for whatever segment you’re currently viewing.

One of our clients is an online mattress retailer. The consistently highest dead click element on their site was a firmness bar graphic. This graphic indicated how soft or hard a mattress is on a scale of 1 to 10 with color gradient swatches. It looked something like this:

The bar was designed to provide a visual representation of the firmness of a mattress in an easy to digest format. It was shown on all mattress products in category display blocks and product detail pages, like so:

The high rate of dead clicks on the bar indicated that users were consistently misinterpreting the graphic as an interactive selector. This behavior was confirmed by watching session replays and soliciting feedback. User feedback indicated that some users thought the feature was broken, while others became frustrated by the inability to select a firmness. Both scenarios were causing site abandonment and loss of orders. We tested the following changes to try and make it more obvious that the firmness bar was not an interactive element:

changing the firmness bar from being assembled with HTML to a single image
replacing the firmness bar with only text
adjusting the styling of the firmness bar so it looked less button-like

None of these approaches worked, so we changed tactics to try and leverage the click activity instead. We iterated on a series of tests that added an accordion dropdown that provided direction on how to place a custom order. This second set of experiments were not successful either.

Testing

Hypothesis

Comprehension: From behavioral analysis with FullStory and learnings from previous experiments, we observed that the firmness bar is a high dead-click target. Repeated attempts to leverage or curb this behavior have so far been unsuccessful.

Response: We want to remove the firmness bar from category and product pages on all breakpoints for all users.

Outcome: Removing a point of distraction/frustration should improve focus on primary CTA points for targeted pages.

Test Execution

Tech Stack

Behavioral Analysis Tool: FullStory
Split Testing Tool: Convert Experiences
Data Analysis Tool: Google Analytics

Planned Estimates

Run Duration: 35 days
Sample Size: 28,000 unique visitors
Minimum Conversions: 200 orders
MDE: 15%
Confidence Threshold: 95%

Executed Thresholds

Run Duration: 35 days
Sample Size: 29,727 unique visitors
Minimum Conversions
- Show Bar: 208 orders
- Remove Bar: 238 orders
Improvement: 16.00%
Confidence: 94.48%

All executed parameters met their planned estimates within an acceptable threshold.

Analysis

Goal Graphs

Summary

On average, about 8% of users were dead clicking on the firmness bar. Removing it (obviously) eliminates those dead clicks entirely. This behavior change was verified through sampling session replays for users placed in both variations. Users that did not see the firmness bar tended to have greater session time and lower abandonment rates, as well as being more likely to purchase within their initial session.

Removing the firmness bar caused a relative 16% improvement in overall order conversion; an absolute lift of +0.22% from 1.39% to 1.61% (+0.02% over the ±0.2 confidence interval with 94.48% confidence). This translates to just under an additional order per day overall (+0.94), or about +$300 per day based on current AOV.

The site is built with templated components, so removing the firmness bar is fairly low effort. Removal can be completed and deployed within an hour or two. Implementing the successful result will cover the expense of testing and site development within a week. The change will have a long-term positive return because it affects a core display element of the site, something that was a permanent feature shown to all users across multiple page types.

Conclusion

Dead clicks cause user frustration, leading to increased site abandonment. In order to understand the reason behind a dead click, you need to look at element importance and user context. Element importance is the visual priority level of the dead clicked element. Was it a dead click on the site background? That was probably an accidental click and not all that important. Dead click on the text in a paragraph? Maybe there’s something interesting in the text that the user was hoping to explore further. Use a session replay tool to help determine user context; see what the user was doing before they triggered the dead click.

List of Revisions

September 3, 2020: Added publish and revision dates after headline. Added tech stack list to Test Execution section.
September 5, 2020: Added screenshot of FullStory dead clicks UI panel to Introduction section.