Read network exposure without calling it new followers

Keep the denominator beside the percentage

LinkedIn defines in-network as the percentage of impressions from members who follow or are connected to you, with an out-of-network category alongside it. These are exposure categories. Members reached is a separate estimate of distinct accounts; followers gained is a different event metric.

Record the exact labels shown in your owner analytics. A viewer outside the reported network category is not automatically a first-time reader, qualified prospect or new follower. The same account can receive several displays, and your relationship can change over time.

Compare two complete constructed observations

The original JSON exercise supplies post A with 1,000 impressions: 600 in-network and 400 outside. Post B has 100 impressions: 20 in-network and 80 outside. The outside shares are 40% and 80%. All values are invented exact inputs, not a creator benchmark.

B has the higher outside share, while A has more outside display events: 400 versus 80. Neither establishes unique people. A change from 40% to 80% does not by itself mean that your outside audience doubled; the denominator changed too.

Weight compatible observations instead of averaging shares

The simple average of 40% and 80% is 60%. It gives the 100-impression post the same influence as the 1,000-impression post. For these compatible constructed counts, the combined outside share is (400 + 80) divided by (1,000 + 100), about 43.64%.

Before using that operation on real records, match account, metric definition, observation age and distribution scope. Keep unknown or incompatible observations separate. This produces a share across selected post observations, not necessarily calendar-period activity or a count of distinct reached people.

Do not manufacture exact counts from rounded shares

If a screen shows only a rounded percentage, multiplying it by the total yields an approximate count. Preserve the displayed precision rather than inventing exact event identities. A rounded split may also not add to precisely 100 on the screen; do not repair the source by assigning the difference to a category.

Our exercise avoids that ambiguity by providing exact constructed counts first. The included runner checks the resulting shares and weighted result. Its success validates the exercise arithmetic, not LinkedIn’s internal classification or your source collection.

Use the split for a narrower decision

For an explainer intended for existing collaborators, in-network exposure can be relevant context. For an introductory lesson intended for unfamiliar readers, inspect whether the language explains its assumptions. Neither situation has a universal ideal outside percentage.

Read replies and appropriate downstream evidence separately. A higher outside share followed by no relevant enquiries is a different outcome from a useful conversation with a current collaborator. Do not choose a content format solely because one observed split is larger.

1. Record the actual metric labels, total and time.
2. Separate the percentage from the approximate or known count.
3. Check compatibility before aggregating observations.
4. State what the split cannot identify.
5. Choose one checkable editorial change and observe its actual result later.

Write a claim that preserves its scope

For the fictional packet, a defensible statement is: “Across these two constructed post observations, 480 of 1,100 display events are outside the declared network category.” It is not a claim about 480 new people or followers.

Save the source version and selection rule with the report. If the real fields are unavailable, leave that question unresolved. A public-profile recap cannot reconstruct this private exposure split from reactions alone.