Page analytics without recording the reader.

The share link is only half the job

An agent writes a report in twenty seconds. You send the link. Then you are guessing. Did the client open it? Did they read past the first screen? Did the pricing section land, or did they bounce at the intro? For a one-off artifact that guess is harmless. For a deliverable, it is the whole game, and a plain host leaves you with nothing but a URL and hope.

So we set out to close that gap. The moment you decide to measure how a page is read, though, you hit a fork, and most tools take the wrong branch.

Why not session replay

Session replay records the viewer: mouse movement, clicks, scrolling, sometimes keystrokes, stitched into a playback of their visit. It is powerful, and it is the standard answer to "how did they use this". It is also the wrong answer for a private client deliverable on a host that leads with private-by-default links.

Recording a client reading a confidential proposal is a privacy liability and a consent problem, and it quietly changes what the product is. A host that promises private links should not turn around and surveil the person on the other end of one. So we drew a hard line: measure the page, never the person.

What the beacon actually collects

The engagement beacon is a small, first-party script. Per viewing session it keeps three things, and nothing else:

1. Engaged time. A one-second tick that only counts while the tab is visible and focused, so a page left open in a background tab does not inflate anything.
2. Scroll depth. The furthest point the reader reached, as a percentage of the page.
3. Dwell by page depth. A ten-bucket array. Each engaged second adds to the bucket for the part of the page currently in view. Summed across a session, this is the shape of where attention went.

That is the complete list. No URLs, no page text, no clicks, no mouse path, no identity, no cookie that follows anyone anywhere. The owner reads an aggregate across all sessions: a typical time on page, the share of readers who reached the end, and the heatmap. It is an extension of the "opened by ~N people" count every Stacktree page already carries, not a new tracking product bolted on.

Reading the heatmap

The attention heatmap is the part people find most useful. It is ten bars, top of the page to the bottom, each as long as the time readers spent there. A long bar near the top and short ones below means people skim and leave. A spike in the middle means a section is doing work. A flat tail means the closing call to action is never being seen, which is the kind of thing you want to know before you send the next one.

Because it is aggregate, one reader who leaves a tab open does not distort it, and no single visit can be reconstructed from it. You learn the pattern, not the person.

The loop, end to end

Engagement is one of three layers, and they compound. Viewers can react and leave a private note right on the page, with no account. An activity feed and an optional daily email tell you who opened, reacted, and commented while you were away. And the engagement view tells you how the page was read. Publish, share, and then, for the first time on most static hosts, actually find out how it landed, without recording a soul.