WhatsApp group chat statistics

The 90-9-1 rule

Nearly every online community repeats a proportion: a small group generates most of the content, a mid-sized group participates occasionally, and most people only read. WhatsApp groups are no exception, though in groups under ten people the distribution tends to be less extreme.

The interesting part isn't confirming the rule — it's seeing where each person falls, and whether that changed over time.

The roles that show up in almost every group

What to look at in a group that doesn't apply one-on-one

Who opens and who closes in a group

In a one-on-one chat, who opens and who closes is a question with two possible answers and it reads fast. In a group it's one of the metrics that clarifies the most, because the two roles rarely land on the same people, and almost never on the most active ones.

Opening means breaking a long silence: writing when nobody has written for hours. It tends to concentrate in one or two people, and those are the ones holding the group together even when they aren't the ones sending the most messages. Closing means having the last word before the next silence, which is a different thing and usually spreads out differently.

A group where a single person opens almost every time is a group that depends on that person. When they ease off, the drop shows up here first, and only later in the message totals.

System messages

A group export contains lines with no author: "X added Y", "X left", "X changed the subject". They aren't anyone's messages and don't count as such.

One important limitation: the export doesn't say how many people were there when the group started. It only records changes from the first line of the file. That's why the net change is shown rather than a member total.

Comparing two eras of the same group

A five-year-old group isn't one thing — it's several, one after another. The ribbon at the top of the dashboard doubles as a range filter, so you can narrow to a year, a month, or whatever window matters, and every section recalculates on that slice.

It's the most direct way to see a change. Set it to last year, note who was on top and how concentrated the conversation was; then set the year before and compare. A group that faded shows it right there, in the gap between two ranges, long before the running total says anything.

It works the other way too: a group that started slow and found its footing, or one that changed people and subject at the same time.

Sharing the results without exposing anyone

Group stats usually end up back in the group, and there's a catch: the names on screen are the ones saved in your phone, not the ones people use themselves. Sending the screenshot as-is means sending your address book.

That's what anonymous mode is for. It replaces every name with "Person 1", "Person 2" and so on, numbered by volume. The charts stay just as readable and the proportions don't move. It covers the people who only appear in a join or a leave, which is exactly where a name slips through unnoticed.

With anonymous mode on, the shareable card also drops the file name from its title — and in a group that file name is usually the group's name.

Large groups

Past twenty people the analysis changes meaning: individual rankings stop mattering and distribution becomes the point. There are usually three or four people generating more than half of everything.

A group of hundreds processes fine, though per-person detail stops being useful at that scale.

Analyze your group

Export the group as a .txt and upload it. It's processed in your browser — the file is never sent to a server. You can filter by date range to compare how the group changed between periods.