How we count a click
The exact rules our redirect follows, so that a difference between our number and another tool's is explainable.
What counts as a click
Every request that reaches a short link and receives a redirect counts as a click, unless one of the rules below filters it. Counting happens at the redirect itself, not through a script on the destination page, so a click still counts when the browser never finishes loading the destination.
What we filter, and why
Known crawlers
Search engines, uptime monitors, and anything else that identifies itself as a bot. These visit real links constantly. Billing for them and reporting them as visitors would be dishonest.
Link prefetch
Chrome, Firefox and Safari all fetch a page ahead of the actual click in some cases: on hover, or when preloading a top search result. These are identifiable from the request itself (headers such as Sec-Purpose: prefetch) and are not counted, because nobody has requested the page yet.
No user agent
A request with no User-Agent header at all essentially never comes from a browser with a person sitting in front of it.
Datacenter IPs: not live yet
Traffic from known hosting and datacenter IP ranges is planned for filtering as well. That filter is not live, and is not claimed here as though it were. This page will be updated when it ships.
How we handle repeat clicks
No session window is applied, unlike some analytics tools. There is no rule deciding that a request is "the same visitor as five minutes ago, so it counts once." Every request not filtered above counts, including two from the same person in the same minute. The one case guarded against is our own machinery: the pipeline is built to tolerate retries, and if it ever processes the same click twice, the duplicate is dropped before it reaches the dashboard. A single real visit is never reported as two because of a fault on our side.
Why this will not match Google Analytics
It will not match, and that is expected. Analytics tools disagree with each other for the same reasons ours disagrees with all of them.
- Different vantage point. We count at the redirect, before the destination page loads. Google Analytics counts after the page loads and its own script runs, so ad blockers, tracking protection and slow connections cause it to undercount in ways that cannot happen here.
- Different bot lists. Every analytics tool draws its own line around what counts as a bot.
- Different session handling. Many tools fold repeat visits from the same person into a single session. We do not, as described above.
Neither number is wrong. They are measured from different places.
How long we keep it
Individual click detail, used for CSV export and for any range within the last couple of days, is kept for as long as the plan allows: 30 days, 12 months, or 3 years. Day-by-day totals are kept indefinitely once summarized, so upgrading a plan later does not erase history. It only leaves the per-click detail on days that have already aged out unavailable.