What “Unique Readers” Actually Counts — and Why It Has an “At Least” Sign

Your document has been opened nineteen times by “≥ 7 people”. The first number is a count of events and means exactly what it says. The second is an inference, and it is worth knowing precisely which inference before you repeat it to anyone.

A person, not a browser

The naive version of this number counts browsers, which is easy and wrong in an obvious way: one person with a laptop and a phone becomes two people, and your audience doubles without a single extra human being involved.

So the count merges wherever there is evidence that two visits are the same person, and falls back to the browser only where there is none. There are three cases, in order of strength:

• Signed in. The reader had a Quixli account and was logged in. Every visit from every device is one person, and this is the only case that is proof rather than inference.

• Typed an address into the email gate. Visits carrying the same address are one person, whatever they were made from. Nobody verified that address, so it is labelled self-reported wherever it appears — what a gate does and does not buy is the longer answer.

• Neither. The reader is distinguished by their browser. This is the fallback, it covers most readers of most business documents, and it is the one that drifts.

The address you mailed it to is deliberately not in that list

It looks like the most obvious identity available and it is the most dangerous one. A link forwards. Count people by the address a link was mailed to and everybody a forward ever reached collapses into a single person — the figure stops moving at precisely the moment your document starts spreading, which is the moment you needed it to move.

A count that goes flat when things are going well is worse than no count, so that address is excluded from the identity by design rather than by omission.

Three things that push the number up

• Cookies cleared or blocked. A reader whose browser keeps nothing between visits looks like a new arrival every single time they come back.

• Private windows. Same effect, deliberately, and common for anything somebody considers confidential — which a proposal about their own company usually is.

• A second device, without signing in or passing a gate. The laptop at work and the phone on the train are two readers unless something ties them together.

And one that pushes it down

Two people, one browser. A shared workstation, a borrowed laptop, a machine in a meeting room, a colleague who reads it on somebody else’s screen while they are away from their desk. Rarer than the three above in most situations, and not rare at all in a hospital, a warehouse, a school or a small office with a shared front desk.

Why the sign is “at least”, and not modesty

The four effects above pull in both directions, and the direction that dominates depends on who your readers are. There is a fifth, though, that only ever goes one way, and it is much larger than any of them.

A document can be seen by somebody who never loads the link at all.

• Projected in a meeting: one visit, and a room of people who read every word of it.

• Screenshotted and pasted into a chat, where the part that mattered reaches a dozen people and the link reaches none.

• Exported to PDF and mailed on, at which point it is a file on somebody’s disk and produces no events anywhere, ever.

• Printed, and read on a train.

• Read over a shoulder, which is how a surprising amount of professional work is actually reviewed.

Every one of those is a human being reading your work, and not one of them is a visit. So whatever the browser-level counting does, the number of people who have seen your document is the number you can see plus an unknown, unmeasurable and never-negative amount. That is what the “≥” is admitting to. It is not a hedge about the arithmetic; it is a statement about the part of the world your analytics cannot reach.

How to use a number that behaves like this

1. Never present it as a fact. Not to a client, not in a report, not in a board deck. The moment it appears next to a real number it inherits a precision it does not have.

2. Watch it move, not its value. Seven versus eight is noise. Two becoming nine over a Tuesday afternoon is a document being circulated inside a company, and that is worth acting on.

3. Read it against reading time. A count that climbs with no real reading behind it is a document being passed around rather than read — and reading time is only worth reading against if it was measured properly in the first place, which is not a given.

4. If you need an actual headcount, ask. “Who else should I make sure sees this?” is a normal question, it costs nothing, and it returns something no analytics screen can.

One practical note: counts, unique readers and average reading time are shown in full on every Quixli plan, including the free one. It is the identity of the person behind a visit that is withheld without a subscription, not the size of the audience.

If a rising count has you wondering who the new readers are, what a forwarded link can and cannot tell you is the next piece. The full set of what is and is not measured is on the features page.