Brazilian bio-link audiences click twice as often as American ones

Brazilian visitors to a link-in-bio page produced 61.6 link clicks per 100 page views. American visitors produced 30.8 — almost exactly half. Turkey sat close to Brazil at 57.5, Germany at the bottom at 23.6. Across all countries the figure was 54.4 clicks per 100 views.

That gap is large, it holds on the two biggest samples in the data, and it is not explained by the thing you would expect to explain it.

The numbers

Window: 16 July – 14 September 2026. Linkhiver pages only. Demo accounts excluded.

Brazil and Turkey are the two largest samples here, and they are the two highest. The United States is third largest and roughly half their rate.

What "clicks per 100 views" means, and what it does not

This is not a click-through rate. One visitor can click several links on the same page, so the figure can exceed 100 — four of the pages in our sample do, topping out at 118.8. It is simply link clicks divided by page views, expressed per hundred.

We report it that way because it is what the data actually supports. We have no session identifier, so we cannot tell you how many distinct people visited, and we will not dress a per-view number up as a per-person one.

Why we think the gap is real

The obvious objection is bots. Automated traffic loads pages and does not click, which drags clicks per view down — so a country with more bot traffic would look disengaged when it is merely polluted. If American traffic were dirtier than Brazilian traffic, that alone could manufacture this entire result.

We checked, and the evidence points the other way.

Traffic we are confident is automated — requests we could not resolve to any country — shows a distinct signature: 78.7% arrive with no referrer, 88.4% are desktop, and they click 2.0 times per 100 views. That is the shape of non-human traffic in this dataset.

American traffic looks the least like that of any country we measured: 16.4% with no referrer, 19.9% desktop. It is the cleanest large sample in the set. Brazilian traffic is 28.3% no-referrer and just 4.2% desktop — overwhelmingly mobile, which is what genuine social traffic looks like.

So the low American figure is not an artifact of dirtier data. If anything, the American number is the one we should trust most.

Indian traffic is the exception worth naming: 84.0% no-referrer and 62.9% desktop puts it close to the automated signature, and we would not draw conclusions from its 30.5.

What we cannot tell you

• Views are page loads, not people. No session identifier exists in this data. A reload counts again, and a creator previewing their own page is counted.
• Bot filtering is behavioural, not identity-based. We have no user-agent record. Our filter removes profile-days where views spike to at least 200 and at least 20× that page's ordinary day. It cannot see steady low-volume automated traffic, or a page whose baseline is already inflated. Between 0% and 94.7% of weekly raw views were removed by it, depending on the week.
• If ad blockers suppress page views more often than click events, every figure here is biased upward. We cannot measure that from our own database.
• This describes Linkhiver pages, not the category. 60 days, one platform, a user base weighted towards Brazil and Turkey.
• We are not claiming to know why. Plausible explanations — WhatsApp-led sharing, mobile-first browsing, different norms around following a link — are hypotheses we have not tested. We are reporting the gap, not its cause.

One finding we are not publishing

We looked at whether the number of links on a page affects clicks per view. The shape was tempting: a peak in the 4–6 link range and a collapse past ten.

We are not publishing it, because the collapse rests on two pages. Two pages cannot support a claim about link count, and a chart of it would be more persuasive than the evidence deserves. We will revisit it when that group has at least thirty.

Method

Every figure above is the output of a single query against our analytics table, filtered to exclude demo accounts and suspected automated traffic. The bot filter is a spike detector applied per page per day: a day is excluded when it records at least 200 views and at least 20× that page's median ordinary day.

Thresholds do not change the headline. Requiring a page to have 1, 10, 25, 50, 100 or 200 filtered views before it counts moves the all-country figure only between 52.0 and 56.1, while the sample shrinks 36-fold.

Data collected 16 July – 14 September 2026. Published 18 September 2026. We intend to re-run this quarterly; figures will change as the user base does.