Heatmap

A heatmap is a visual representation of aggregated user behavior on a webpage, using color gradients, typically ranging from cool colors such as blue for low activity to warm colors such as red for high activity, to show where users click, move their cursor, or scroll to across many sessions overlaid directly on a single image of the page. The three most common types are click maps, showing where users click or tap regardless of whether that location is an actual clickable element, move maps, tracking cursor movement as a rough proxy for visual attention on desktop devices, and scroll maps, showing what percentage of visitors reach each vertical section of a page as they scroll down toward the footer. A fourth, less common variant, the attention or hover map, is sometimes distinguished separately on mobile devices, where tap and scroll events replace cursor movement entirely as the primary behavioral signal being captured.

Heatmaps matter because they translate abstract behavioral data into an intuitive visual format that is immediately understandable to designers, marketers, and stakeholders without requiring any real familiarity with raw analytics tables or event reports. A scroll map showing that only 20 percent of visitors ever reach a page's pricing section, positioned near the bottom of a long page, communicates a clear and actionable problem far more effectively than the equivalent numeric scroll-depth report buried inside a spreadsheet, and this visual accessibility is a major reason heatmaps are so widely used in stakeholder presentations and client reporting throughout CRO and UX work.

Heatmaps are generated by tools such as Hotjar, Microsoft Clarity, or Crazy Egg, which inject a lightweight tracking script that records cursor and touch coordinates along with scroll position across visiting sessions, then aggregates that raw data into the visual overlay once a sufficient sample size has accumulated, typically at least 1,000 to 3,000 sessions per page for a reasonably stable and representative pattern to emerge, since pages with lower traffic can produce heatmaps skewed by a small number of atypical sessions. Analysts generally segment heatmaps by device type, since click and scroll patterns on mobile differ substantially from desktop, and by traffic source, since visitors arriving from different channels often carry different intents and exhibit different behaviors even on the exact same page and layout.

A frequent misconception is that heatmaps show what users are looking at, when in reality move maps only track cursor position, which correlates with visual attention on desktop but only imperfectly, since users often keep their cursor still while reading with their eyes elsewhere on the screen; genuine gaze tracking requires dedicated eye-tracking hardware, which is far less commonly used in commercial CRO work than software-based heatmaps that rely purely on cursor and touch data. Another common pitfall is drawing firm conclusions from a heatmap in isolation without cross-referencing it against click targets and page structure, for example misinterpreting heavy clicking on a non-interactive image as positive engagement when it may actually indicate user confusion about whether that image was meant to be clickable at all.

In CRO and UX consulting, heatmaps are typically used early in an audit to generate hypotheses about where friction or confusion may be occurring, for instance identifying that users are repeatedly clicking on a static banner image expecting it to link somewhere, or that a scroll map reveals a key call-to-action sits well below the point where most visitors actually stop scrolling. Because heatmaps are descriptive rather than causal, meaning they show what happened but not definitively why it happened, they are almost always paired with session recordings for deeper qualitative context and followed by controlled A/B testing to confirm whether a proposed fix genuinely improves the metric it was intended to affect. Refreshing a heatmap after a design change also serves as a quick, low-cost sanity check that a new layout is being engaged with as intended before waiting for a full experiment to reach statistical significance.