Why adding post reach does not count your total audience

State which audience you want to count

“How many accounts saw post A?” and “How many different accounts saw at least one of my three posts?” are different questions. LinkedIn’s member definition of members reached concerns distinct members and Pages for the post, without repeat views. It is still an estimate.

A recap often moves from a post table to a year-wide claim. That is where duplication can enter unnoticed. Keep the post-level metric label and avoid relabeling its sum as your whole audience. Even a correctly deduplicated account total would not count completed reads or customers.

Follow three overlapping fictional sets

The original exercise gives post A accounts P1, P2 and P3; B has P2, P3 and P4; C has P3, P4 and P5. Each post reaches three constructed accounts. Summing the post counts gives nine, but the union contains only five distinct IDs.

P3 appears in every post audience and must be counted once in the union. These fictional IDs make set membership visible for teaching. They are not an export of people who viewed a LinkedIn account, and this guide does not collect silent viewers’ identities.

See why counts alone cannot recover the union

Now hide the identities and keep only three totals of three. If all posts reached the same three accounts, the union is three. If none overlapped, the union is nine. The original packet includes both alternatives, each matching exactly the same per-post totals.

For exact sets under a compatible scope, the union is at least the largest individual set and at most their sum. Those simple bounds do not reveal the actual overlap. Estimated, rounded or differently scoped platform metrics need additional caution; do not turn an exact exercise bound into a precise measurement claim.

Use an explicitly defined aggregate when available

If an authorized analytics source supplies a distinct total across a period, read its unit, account scope, date basis and deduplication definition before using it. A field called total reach is not enough on its own. If the needed definition or field is unavailable, report the per-post observations without inventing the union.

Do not try to fill the gap by identifying everyone who silently viewed a post. Normal post analytics contain aggregate exposure and limited breakdowns, not a complete permission to collect individual viewing histories. A reaction list is another population and cannot repair a missing exposure union.

1. Define one-post versus across-post audience scope.
2. Keep metric definitions, account and observation clocks compatible.
3. Check whether the source actually supplies the needed distinct aggregate.
4. If only per-post counts exist, label their sum as a sum.
5. State overlap uncertainty and avoid claiming unique readers or customers.

Write the narrow claim you can support

For the visible fictional sets, you can say “the three post counts sum to nine, and the known union contains five accounts.” If you had only the three totals, the second clause would be unavailable. Preserve that difference in both chart titles and captions.

For a real report with only post-level estimates, use wording such as “sum of reported per-post members reached; accounts may repeat across posts.” It is less dramatic than an invented unique-audience figure, and it answers a reproducible question.

Keep a reproducible explanation without private identity data

Run the included exercise checker to reproduce nine summed counts, the five-account union, and the three/nine alternative unions. No packages, account or private dataset is required. This validates the constructed arithmetic rather than any platform’s deduplication behavior.

Keep the original scope and dictionary version with your own report. A later collection may describe a different window or changing relationships; do not silently combine it with an older total. Leave the unavailable distinct result explicitly unresolved.