Visitor tracking that tracks behaviour, not people.
See every visit as a journey: where it came from, what it touched, how far it read, and whether it converted. Identities rotate daily by default, so there is nothing following anyone around.
What is website visitor tracking?
Website visitor tracking records what individual visits do on your site: the pages seen and in what order, the source that brought them, the clicks and events along the way, and whether they converted. It is behaviour at the level of a visit, rather than site-wide totals like monthly pageviews.
The word “visitor” is doing a lot of work in that sentence, and different tools mean very different things by it. Some mean a browser they can follow for a year. Some mean a company they inferred from an IP address. Mrkr means one person’s activity, identified well enough to be readable as a story and not well enough to follow them tomorrow.
Every visitor, as a list.
Each row is one visitor in the selected range: when they were first and last seen, how many sessions and pageviews they accumulated, and whether they ever completed a conversion. The list loads a hundred at a time, most recently active first.
In the default cookieless mode a visitor shows as a short handle rather than a name, because a stable-looking name would imply a long-term identity that does not exist. In cookie mode they get a consistent, deterministic handle instead, so the same person reads the same every time you look.
Open a row and the visit becomes a timeline: pageviews, autocaptured clicks and custom events in order, with the deepest scroll point reached and a link straight into the replay.
What visitor tracking is actually good for.
Explaining an average
Aggregates tell you where to look. Ten timelines that all stall on the same field tell you what you are looking at.
Reproducing a support ticket
Open the visitor, read the sequence, watch the recording. Browser, viewport and event order are all there.
Checking a launch honestly
The first fifty journeys through a new page say more in ten minutes than a week of waiting for significance.
Seeing what people read
Deepest scroll point per session, per page. With engaged time it separates skimmers from people who stopped halfway.
Two different things called visitor tracking.
Person-level identification
- Resolves an IP address to a company, or a form fill to a named individual.
- Persists an identifier for months so one browser can be followed indefinitely.
- Enriches visits with third-party contact data bought from a data broker.
- Mrkr does none of this, and it is not on the roadmap.
Behaviour-level tracking, what Mrkr does
- Identifies a visit by a salted hash of request signals, rotated every day.
- Shows the full journey through the site, in order, with timestamps.
- Ties the visit to its source, campaign, scroll depth, events and revenue.
- Stores no IP address and no name, because it never collects one.
- before a cookieless visitor id rotates and cannot be rejoined
- 24hbefore a cookieless visitor id rotates and cannot be rejoined
- IP addresses stored: it goes into the hash and is discarded
- 0IP addresses stored: it goes into the hash and is discarded
- live window for anyone currently on the site
- 5 minlive window for anyone currently on the site
- tracker script doing the collecting
- <6 kBtracker script doing the collecting
Rotation is what separates a daily salted hash from a fingerprint. A fingerprint is built to persist; this identifier is built to stop working.
What each mode can answer.
Cookieless is the default because most teams want the banner gone more than they want the longer window. Cookie mode is one switch per site. The honest difference is this.
| Question | Cookieless (default) | Cookie mode |
|---|---|---|
| How many people visited today | Yes | Yes |
| What did this visitor do in this session | Yes | Yes |
| Where did they come from, and on what campaign | Yes | Yes |
| Did they convert on this visit | Yes | Yes |
| Is this the same person who came last week | No, by construction | Yes |
| How many distinct days was this person active | No | Yes |
| Consent banner required for it | No | Yes |
Nothing on this page is legal advice. The reason cookieless needs no banner is that the ePrivacy rule governs storing and reading on a visitor's device, and cookieless mode does neither for measurement.
Open a real visitor
The live demo has a populated visitor list with journeys, scroll depth and replays attached. No signup, no card.
Open the live demoThe trade-off, stated plainly.
Cookie mode is a per-site setting that gives persistent identity across sessions and weeks, and it means you need consent for that site. Mrkr will not pretend the switch is free in either direction.
No identity across days
The salt is thrown away every 24 hours. The same person hashes to a different value tomorrow, and nothing can join the two.
Questions this rules out
No multi-week timelines and no how-many-visits-before-they-buy. That is the direct price of not needing a banner.
No names, in either mode
No enrichment step and no IP-to-company lookup anywhere in the product. Cookie mode adds persistence, never identity.
Questions, answered.
Keep reading
- Session replay without the cookie banner
Watch the visit instead of reading it, with input masked in the browser.
- How analytics works without cookies
The daily-rotating hash in detail, and why it is not fingerprinting.
- Engagement metrics
Scroll depth, engaged time, and stickiness across days.
- Visitors documentation
The list, the drawer, and exactly how identity works in each mode.
- Cookieless vs cookie mode
What changes when you flip the switch, and what it costs.
- Every Mrkr feature
The full index of what the product measures.
Your first visitor is already here.
Drop in the script and watch them land. It takes about a minute.