Session replay is an add-on feature for paying customers only. You may try the feature for free for 14 days or 10,000 sessions (whichever comes first), after which you must contact Sales to continue use.
For example, your heatmap might show a lot of clicks on a blank spot on the page, or the clicks seem misaligned by some number of pixels.
The process for generating heatmaps selects an arbitrary version of the page, which can mean that the version of the page that Heap loads may not match up with the aggregates. Here are some common reasons why, and tips for troubleshooting.
You recently changed your UI
We recommend that you pick a Date range that doesn’t span multiple UI changes. If your UI changed significantly on January 11th, we recommend that you make the Date range as January 11th onwards. To see your old UI in a heatmap, the Date range should reflect the time before your change.
Heatmap background images are randomly chosen, and there’s no way to control which background image will show in your heatmap. For example, if you make a major change to your UI on January 11th and you make a heatmap for Jan 10–12th, then the background image will be from a session on Jan 10th, 11th, or 12th. Wether the image is of your new UI or your old one will be randomly determined.
Picking a date range with multiple UI iterations isn’t only counterproductive to having an accurate background image — it also makes that heatmap’s data less useful. It would be an aggregation of clicks from multiple different UIs, which isn’t an accurate picture of what your users experienced.
Page selection is too broad
You’re seeing clicks on content that doesn’t appear on the background image.
For example, you are an ecom store and you selected PDP/*/ because you wanted to view aggregated clicks across every product details page. However, your jeans PDP has extra sizing options (tall, short, standard) that your shirts PDP does not have. If the background image selected is from the shirts page, then you might see “ghost clicks” where the extra sizing dropdown is on the jeans page.
Try: Running a more specific URL or query to generate more targeted heatmap pages.
The background page is dynamic, or subject to change a lot
For example, you run a heatmap query on a page that updates frequently based on user selections. This page is dynamic and might change a lot; two users may have completely different looking pages, with clicks in completely different spots. This means that the background image might not match up with the click overlay effectively.
This could also happen if your website has many popups or interactive elements that load a few seconds after the primary page. For example, let’s say you serve users a “Share your email address!” pop up after the user browses for 10 seconds. The background image of your heatmap may not show that email popup, but you could see “ghost clicks” on the yes/no elements of that pop up.
Try: Using Explore events or Journeys to understand user behavior on highly dynamic pages instead.
You serve different versions of the page to different users
For example, let’s say you’re running an A/B test where paying customers see version 1 of the `/overview` page but free customers see version 2 of the `/overview page`. If you don’t filter for paying vs. free customers, then we’ll aggregate all those versions together. This means you might get a background image of version 1 with a hotspot of clicks that correspond to version 2.
Try: Filter down your heatmap results to a more precise subset of users.
Multiple Languages
For example, let’s say a marketer in the EU loads a heatmap of users who have visited her site’s homepage. There are 500 eligible sessions to look at; 50% of users saw the homepage in English, 25% in Spanish, 15% in German, etc. Heap will arbitrarily select one version of the homepage and render that as the visual.
Try: Filter down your heatmap results to a more precise subset of users.
You may have a low sample size of sessions
For example, let’s say you have a heatmap of only 20 sessions; 10 of them may be from windows that were 1600+pixels wide, but 2 of them were from windows that were just 1200 pixels wide. It’s possible that the clicks from the 1200 pixel wide clicks might look out of place if laid atop a background image that was 1600+ pixels wide.
Try: Expanding the date range for your heatmaps.