Revue hebdomadaire des statistiques LinkedIn

LinkedIn analytics should end with a publishing decision, not another screenshot of a dashboard. A useful weekly review answers one question: what should you repeat, change or stop next week?

LinkedIn's own post analytics separate discovery, profile activity, social engagement and link engagement. Impressions are estimated views, while members reached is an estimate of distinct viewers. Treat both as directional signals, then connect them to profile visits, followers, clicks and conversations before deciding what worked. LinkedIn Help

Step 1: Build a small review set

Review your last 5 to 10 posts, not your lifetime totals. For each post, record the topic, audience, format, opening promise, call to action, impressions, members reached, profile viewers, followers gained, clicks, saves, comments and reposts.

Do not treat every field as equally precise. LinkedIn states that several analytics numbers are estimates, and availability varies by content type and age. Record the date of the review so a later comparison uses the same window. LinkedIn Help

Step 2: Separate distribution from intent

Impressions answer how often a post was shown. They do not answer whether the right people cared. Members reached is closer to unique distribution, but it is still an estimate.

Use four layers:

1. Distribution: impressions, members reached and in-network versus out-of-network views.
2. Response: reactions, comments, reposts, saves and sends.
3. Intent: profile viewers from the post, followers gained, link visits and direct replies.
4. Outcome: a qualified conversation, call, trial, referral or opportunity.

These layers are not a funnel with guaranteed conversion. They stop you calling a high-reach post successful when it produced no useful next step. A post with modest reach and a relevant conversation can be more valuable than broad passive distribution.

Step 3: Tag the signal, not just the number

Numbers become useful when each post has enough context to compare it with another post. Use a short tag set:

• point of view: the belief or trade-off made explicit;
• proof: observation, example, result, screenshot or customer story;
• reader: the role or situation addressed;
• format: the container used to deliver the idea;
• next step: comment, profile visit, click, reply or no action.

Keep the tags stable for four weeks. Changing the taxonomy every review makes patterns look more precise than they are.

Step 4: Turn one pattern into one test

Choose one decision for next week: repeat a topic with a new example, change the format, rewrite the opening when reach was good but profile activity was weak, remove a low-quality call to action, stop a topic with no relevant response, or answer the strongest comment in a follow-up post.

Write the decision as a test: “For the next two posts, I will keep X, change Y and look for Z.” Do not change topic, format, audience and call to action at the same time. You will not know what caused the result.

A practical weekly review template

What repeated? Name one topic, promise or format that appeared in more than one post.

What moved the right people? Note the strongest profile, follower, click or reply signal and who created it.

What was only reach? Mark the post that travelled furthest without a meaningful next step.

What will change? Choose one variable for the next two posts.

What evidence will count? Define the signal before publishing, such as “two replies from people in my target role”, rather than “more engagement”.

How Orsana fits into the review

Orsana works at the decision layer: group posts by the context you choose, spot repeated patterns and turn the review into a next-post brief. It does not make LinkedIn's estimates more precise, and it cannot infer a qualified business outcome that was never recorded. Your judgment and a clear definition of “qualified” still matter.

FAQ: LinkedIn analytics action plan

What should I track every week on LinkedIn?

Track distribution, responses, profile activity, link visits and the business outcome you can actually observe. Also record each post's topic, reader, format and call to action so you can compare like with like.

Are LinkedIn impressions the same as people reached?

No. Impressions count how often a post was shown. Members reached is an estimate of distinct members and Pages that saw it, without repeat views. LinkedIn says both figures may be estimates.

How often should I review LinkedIn analytics?

Once a week is a practical default for most creators and small teams. Review a consistent set of recent posts, then make one test. Daily checking is useful only for a specific launch or experiment.

Should I optimize for impressions or leads?

Neither metric should win automatically. Use impressions and reach to understand distribution, then prioritize the signal closest to your goal: a relevant reply, profile visit, click, call or opportunity.

Read next: LinkedIn analytics metrics · best LinkedIn analytics tools · LinkedIn dashboard