When you publish a link you usually picture a specific audience: channel subscribers, newsletter customers, site visitors. Most often, people in your own country.
We checked that against 723,758 links. The median short link receives clicks from twelve countries. Fewer than four per cent of links stay inside a single country.
That changes several practical things, from the language of your landing page to the currency printed on it.
Method
Sample: 723,758 links with at least 50 redirects over the last 12 months. The threshold exists so that geography is computed over a meaningful number of clicks rather than two stray visits.
Country is derived from the IP address of the redirect.
Filtering: none. Test links, service links and automated traffic all remain in the sample.
One property of this method matters: each link contributes exactly one observation regardless of how many clicks it has. No single large link can therefore skew the result — unlike breakdowns weighted by traffic volume.
How many countries one link reaches
| Measure | Value |
|---|---|
| Median | 12 countries |
| 75th percentile | 18 countries |
| 90th percentile | 25 countries |
| Exactly one country | 26,687 links — 3.69% |
| Three countries or fewer | 74,156 links — 10.25% |
Half of all links gather an audience from at least twelve countries. One in ten reaches more than twenty-five.
The inverse figure matters more: only 3.69% of links are genuinely local. If you believe your link works inside one country, statistically you are almost certainly wrong.
Even the "home" country is not the whole link
The second measurement: what share of a link's clicks comes from its single largest country.
| Percentile | Share of top country |
|---|---|
| 25th | 50.9% |
| Median | 63.5% |
| 75th | 78.3% |
For the median link, the leading country supplies about two thirds of redirects. The remaining third comes from abroad. For a quarter of links, the main country does not even reach half.
So even when a link has an obvious home market, one click in three arrives from outside it.
Why this explains the flat daily curve
In the analysis of when people click short links, something odd turned up: links are created on a sharp office-hours curve (2.43x spread) while clicks are distributed almost evenly around the clock (1.38x). We had to leave a caveat there: perhaps that was just timezones adding up, and a local audience would show a sharper curve.
That caveat now has an answer. The flat curve is not a statistical artefact but a direct consequence of links genuinely travelling. The median link lives in twelve timezones at once, and its daily curve is the sum of twelve different daily routines.
A local audience does produce a sharper curve. That situation simply applies to 3.69% of links.
What follows
Check what someone who does not read your language sees. A third of redirects to the median link arrive from outside its main country. If the landing page exists in one language only and offers no way to switch, that third hits a wall.
Prices and currency. A page quoting a single currency shows it to roughly a third of visitors for whom that currency is foreign. Not always a problem, but always worth knowing in advance.
Do not plan around your own timezone. With an audience spread across a dozen zones, "morning" and "evening" stop being shared concepts. The detail is in the piece on click timing.
Availability restrictions. If the destination is unavailable in some countries — geoblocking, age requirements, local rules — you will not find out quickly, because in your report it still looks like an ordinary click.
Combine this with lifespan. Separately we measured that the median link keeps receiving clicks for 153 days. Together with geography that means: a link you think of as "a local one-week promotion" turns out in practice to be an international object with a six-month life. Both habitual assumptions — about reach and about duration — err in the same direction.
Limitations
IP geolocation is approximate. VPNs, corporate networks and mobile carriers routing through other countries all add countries that were never really there. How much they add cannot be estimated from this data.
Automated traffic was not filtered, and here it hurts most. Crawlers and messenger preview fetchers arrive from datacentres in various countries, and each such visit adds a country to the link. Part of the observed twelve may be exactly that. This is the study's most substantial weakness, and it biases the result upward.
The 50-click threshold skews the sample. A link that has gathered fifty redirects has already spread. Smaller links — the majority — are probably narrower geographically, but these figures do not describe them.
We abandoned the per-country breakdown. We attempted to build daily curves for individual large countries, in order to measure how much sharper a local audience's curve is. In the largest country a single link accounted for 34% of the sample's traffic, and the shape of the curve did not repeat from month to month. That is not a publishable result, so we are not reporting it.
This is one service's data. A service with a different audience will show different geography.
A short link is technically simple — an ordinary HTTP redirect. But it has neither borders nor a schedule: it travels further than its author expects, and outlives the occasion it was created for.
