How it works
One script, four sources, and a schedule. What gets read, how often, and where the numbers come from.
How it works
Two kilobytes, deferred, first-party. It records the page, the referrer, the browser, the OS and a country derived server-side. Your IP is used at the moment of the visit and never stored.
One Google connection covers Search Console and GA4 — same consent, granted once, read-only. Google publishes its data on a two-to-three day delay, which is their pipeline and not a lag here.
Your tracked terms are checked in real search results, and the domains that appear above you become the competitor list — the rivals a searcher actually sees, not a list you guessed at.
ChatGPT, Claude, Gemini, Perplexity and Grok get the questions your buyers ask, generated from your own top pages and search terms. Answers are stored with who was cited.
Uptime and response time every five minutes. A crawl of your own pages weekly, checking titles, headings, alt text, broken links and redirect chains — weekly because these change when you deploy, not overnight.
A ranking move past your threshold opens a Slack message or a Linear issue. A digest summarises traffic, movement, mentions and anything health flagged. Repeat sends are prevented by a ledger, not by the schedule, so a retry never means a second email.
Not for the tracker. It sets nothing on the visitor's device and stores no identifier that lasts beyond the day — the identifier is a hash of IP and user agent salted with a secret replaced every night, so yesterday cannot be linked to today by anyone, including us. That is a description of the default, not legal advice. Switching a site to identified mode changes it, and that consent becomes your responsibility.
They will not, and neither is wrong. GA4 is removed by content blockers on a large share of traffic, samples big reports, and withholds rows for privacy thresholds. A first-party script is not blocked the same way and counts sessions by different rules. The GA4 tab shows both on one axis; the useful reading is whether the ratio between them stays steady.
Raw events are kept 30 days on the free plan and one to five years on a paid one, depending on the plan. Daily rollups — counts with no per-visit detail — are kept indefinitely, which is what lets a chart span years without keeping a log of individual visits.
Not the kind that records the page. A site can switch on click heatmaps and a replay of the pointer, both off by default: they keep positions as numbers and nothing about what was on the screen, what was clicked or what was typed, so there is no video of anyone’s visit. If you need to watch the page itself, this is the wrong tool.