Why did my LinkedIn reach drop 2026: find the signature, not the list

Take three readings before you take any advice

Three readings taken from analytics you already have will narrow a reach drop to one of six causes, and each cause has a different first move. The standard article on this question hands you a list of reasons: external links, engagement pods, posting inconsistently, a profile that does not match what you write about. Every item on that list is real. None of it is diagnostic, because a list gives the reader no way to tell which entry is theirs, and choosing wrong costs a month of effort pointed at the wrong layer.

The three readings are the serve ratio, the layer split and the onset shape. Each takes about three minutes to pull and none of them needs a paid tool. Together they produce a signature, and a signature points at one cause instead of six.

Pull all three before you interpret any of them. One reading on its own is ambiguous.

None of these three are LinkedIn features. They are relationships between numbers LinkedIn already shows you, which is why no dashboard surfaces them and no vendor blog explains them. The platform reports quantities. Diagnosis lives in the ratios between quantities, and the ratios are yours to calculate.

The serve ratio tells you whether the pool shrank or the rounds did

A serve ratio close to 1.0 means almost everyone who saw your post saw it once, and a ratio climbing above your own normal means the feed is topping up your distribution by showing the same people the same post again. That distinction is the difference between a content problem and an audience problem, and it is invisible in the impression count alone.

The mechanics are simple. Impressions count every time your post is rendered on a screen. Members reached counts the unique people it was rendered to. Divide the first by the second and you have the average number of times a person who saw your post was served it. The figure cannot fall below 1.0, and its normal level is a property of your account and your audience rather than a published platform benchmark.

Two readings matter. If members reached is falling faster than impressions, so the ratio is rising, your reachable audience is contracting and the feed is compensating with repetition. If impressions and members reached fall together at roughly the same rate, so the ratio holds steady, the pool is intact and your posts are getting fewer distribution rounds inside it.

Those two findings send you in opposite directions. A rising ratio is fixed by adding audience surfaces, which means commenting where your buyers already read, being tagged by other people, and turning up in search results for your topic. A steady ratio with falling impressions is fixed inside the writing, because the post is not earning its second round of distribution. Posting more often repairs neither, and it makes the first one worse by asking a shrinking pool to absorb more.

Profile views separate an account problem from a content problem

If your post impressions fell over the last 90 days while your profile views held roughly steady, the problem sits in the posts and not in the account. Profile views are fed by surfaces that have nothing to do with how well a given post performs: comments you left under other people's posts, your name appearing in search results, mentions, tags, and people clicking a byline. When those hold up and post reach does not, your name is still being distributed and your recent writing is not earning rounds.

The reverse pattern reads differently. Profile views and post impressions falling together across the same period points at something account level, and the honest list of account level causes is shorter and more boring than the folklore suggests. You went quiet for three weeks and came back. You changed your headline and your posts no longer match the topics your profile is associated with. You moved country or switched the language you write in. You stopped commenting, which removes the surface that produced most of those profile views in the first place.

Run the split over 90 days, not 30. Profile views are noisy at short ranges because a single post that travels can produce a week of them, and a single quiet fortnight can halve the count without meaning anything. Ninety days smooths out the noise while still being recent enough to describe your current account rather than your account from last year.

The shape of the fall tells you whether to look for a date or a drift

Write your last twenty posts in date order with their impression counts beside them, and look at whether the numbers step down or slide down. A step has a cause you can name and usually undo. A slide has a cause you have to reason about, and it is rarely one thing.

A step down between two consecutive posts means something changed in the window between them. Find the last normal post and the first abnormal one, then write down everything you altered in the days between the two. That list is short, and the answer is almost always on it.

• You changed your headline, your About section, or the topic you write about, so the profile no longer matches the posts
• You switched your default format, for example from text posts to video or from text to documents
• You started putting outbound links in the body of your posts, which is worth reading about before you assume the size of the effect in what the link penalty evidence actually shows
• You joined a comment group or a pod, or a group you were already in changed how it operates
• You were away, and your first post back landed to an audience that had stopped expecting you
• You changed cadence sharply in either direction, from weekly to daily or from daily to monthly

A slide instead of a step points at drift, which means the audience changed underneath you rather than the distribution changing above you. Followers accumulated from one viral post about a topic you no longer write about. Your writing moved from a narrow subject to a broad one, or from a professional register to a personal one. The people who followed you for the first thing are still counted as followers while quietly declining to engage, and a follower who does not engage lowers your reach rather than raising it.

The Ra-Aha Reach Signature Grid

Six signatures cover almost every reach drop an individual account experiences, and each one is defined by what the three readings do together rather than by any single number. Find the row your numbers match and read only that row. The grid exists so that you stop applying fixes designed for someone else's problem.

Two of the six signatures end with the instruction to change nothing. That is deliberate. A diagnostic that always concludes the writing needs work is not a diagnostic, it is a sales page for writing help, and the reason so many reach articles reach the same conclusion is that the people publishing them sell the same service.

A worked example: the same account, two quarters

Here is the arithmetic on one account across two quarters, with every input labelled as an assumption you should replace with your own figures. Assume a consultant posted twelve times in each quarter, kept the format constant, and pulled the impression and members reached figures from each post before the panel got stale.

Illustrative arithmetic on assumed figures. Substitute your own two quarters and the reading method stays identical.

Read it in order. Impressions fell by a quarter, which is what prompted the panic. Members reached fell by nearly half, which is a much bigger fall than the headline number suggested. The serve ratio therefore rose from 1.56 to 2.02, meaning the average person who saw a post in quarter two saw it twice. Profile views barely moved, so the account is still being distributed normally. That combination is signature two, and it says the audience pool contracted while the writing carried on working.

Consider what the wrong diagnosis would have cost. Reading only the 25% impression fall, this consultant would have concluded that the posts had gone stale and spent six weeks rewriting hooks. The hooks were fine. The pool had shrunk, and the repair was to spend twenty minutes a day in other people's comment sections rebuilding the surface that brings new people in. Same effort, different layer, and only one of them works.

The first move for each signature, and how long to wait

Each signature has one first move, and the discipline is doing that move alone until you have measured it. Changing your writing, your cadence and your commenting habits in the same fortnight guarantees that whatever happens next teaches you nothing, because three variables moved and one outcome changed.

One move per signature. Hold everything else constant for the stated window before judging it.

The waiting periods are not arbitrary politeness. Distribution decisions are made per post, so a change in your writing can only be evaluated across enough posts for the noise to average out, and six is roughly the point at which one outlier stops dominating. Audience surface work is slower still, because a comment you leave today produces a profile view next week and a follower who engages the week after.

What your analytics cannot tell you, and who profits from that

Three things you would want to know are simply not available to you, and being clear about that is more useful than pretending otherwise. LinkedIn does not expose dwell time to creators, so every dwell time benchmark you have read is a number you cannot verify against your own account. It does not expose impressions over time on a post, so the shape of a single post's distribution has to be recorded by hand as it happens, which is the method described in the two clock test for a post with no impressions. It does not tell you who was in the audience, only how many of them there were.

That absence is why so much of this category is folklore. When a number cannot be checked, the number that spreads is the one that sounds most decisive, and the pages publishing decisive numbers are usually selling scheduling software, engagement services or writing help. The three readings in this article are worth the effort precisely because they use data you can see and arithmetic you can repeat.

Run the triage before you read another article about the algorithm, including this one a second time. The readings take longer to describe than to perform, and once you have them the entire category of generic advice becomes filterable, because you can ask of any tip whether it addresses your signature or somebody else's.