The LinkedIn company page audit checklist that scores distribution, not decoration

Audit the four signals that govern distribution, not the banner

Four things decide whether a company page reaches anyone: how consistent its subject matter is, how many employees are formally attached to it, how regular its publishing rhythm is, and how large its following is relative to its size. Everything else on a page audit is decoration, and decoration only matters after somebody has arrived.

The free tools do not score these four for an unavoidable reason. They score what a public page shows a stranger, and three of the four signals require a number the tool cannot see, namely your actual headcount and your posting history in a form it can label. That is a fair limitation for a free tool and a bad reason to run your audit on their terms.

Signal one, topic concentration across your last twelve posts

Open your last twelve posts, give each one a single topic label, and count how many sit inside the top two labels. Divide by twelve and multiply by ten for the score. A page below six here has no subject, and a page with no subject cannot be surfaced to people interested in a subject.

The rule for labelling is one label per post and no hyphenated compromises. If a post genuinely needs two labels it is a post about nothing in particular, which is exactly what the signal is designed to catch.

A worked labelling of twelve posts. Top two topics account for seven, so seven divided by twelve times ten scores six.

Six out of ten is a common score and it looks harmless. In practice it means two thirds of the page's output is teaching the feed and the reader two different things about what this company is for. The fix is not to post less, it is to make the third and fourth topics into occasional posts rather than a third of the calendar.

Concentration also decides who you can reach. Reaching the whole buying committee requires the page to be legible to several different roles, and legibility starts with a page that is about a recognisable thing.

Signal two, how many employees have actually associated their profile with the page

Count the profiles on your page's People tab, divide by your real headcount, multiply by ten and cap the result at ten. This is the signal nobody scores and it is usually the cheapest one to fix, because it is a one time profile edit rather than an ongoing behaviour.

The mechanism matters. A profile that names your company as its current employer is what makes that person count as an employee of the page. The control that notifies employees about a new post reaches the profiles the page recognises, and the page's own posts are more likely to be surfaced to people connected to those profiles. An employee whose profile still lists their previous employer, or lists a free text version of your company name that does not resolve to the page, is invisible to all of it.

Association rate scoring, worked at four company sizes.

That contrast is the practical argument for doing association before advocacy. Advocacy programmes ask people to change a habit and most of them decline quietly. Association asks for a two minute edit once, and it raises the ceiling on every post the page publishes afterwards.

Signal three, cadence variance, which matters more than cadence volume

Measure the gaps between your last twelve posts in days, take the median gap and the largest gap, then score eleven minus the largest divided by the median, floored at zero and capped at ten. Two pages can publish an identical number of posts per month and score at opposite ends of this scale.

Volume is what every audit measures because it is the easiest thing to count. Variance is what a reader and a feed actually experience. A page that publishes six times in one week and then disappears for a month has trained nobody to expect anything, and each burst starts from a standing position.

Cadence variance scoring. The first and third rows publish a similar monthly volume and score sixteen points apart across the audit.

Read row two carefully, because it is the finding that surprises people. A page publishing once a week without fail scores full marks on this signal, and it beats a page publishing three times a week erratically. If your team can hold one post a week and nothing more, hold one post a week and stop apologising for it.

Signal four, follower to headcount ratio

Divide followers by headcount. Under five times scores two, five to fifteen scores five, fifteen to forty scores eight, and above forty scores ten. The ratio answers a question that raw follower count cannot: did this page earn an audience, or did it inherit one from the payroll and their immediate contacts?

Ratio scoring at four company shapes. The manufacturer has three times the follower count of the consultancy and half the score.

A forty person consultancy with three thousand followers is in a stronger position than a two thousand person manufacturer with twelve thousand, and every audit that reports follower count alone tells them the opposite. The ratio is also the only one of the four signals that moves slowly, so treat it as a direction of travel rather than a monthly target.

One caveat on inflation. A ratio built by inviting every contact regardless of fit produces a number that scores well and reaches nobody, because followers who are not your buyers do not engage and the page's output stops being shown to the ones who are. Invitation credits are a finite resource and spending them on volume is the most common way this score gets faked.

Score it out of forty and read the band

Add the four scores. The band tells you what to do next, and in three of the four bands the instruction is to do one thing rather than several. Audits fail at this point far more often than they fail at measurement, because a list of twenty findings is a list nobody acts on.

Bands and the single next action for each.

One further rule. Do not re-score more often than every six weeks. Three of the four signals are calculated from your last twelve posts, so a page publishing twice a week takes six weeks to fully refresh the sample. Scoring monthly measures the same posts twice and produces movement that is arithmetic rather than progress.

If your reach has visibly fallen rather than simply being low, run the score first anyway. A page whose reach dropped usually has a change in one of these four signals sitting behind it, and the score dates the change.

The thirty minute run order

Run it in this order and it fits in half an hour with a spreadsheet and two browser tabs. The order exists so that the two slowest tasks happen while you still have attention for them.

Two people can run this in parallel and compare, which is worth doing once. Topic labelling is the only subjective step, and two independent labellings that disagree wildly are themselves a finding about the page.

The cosmetic layer is worth ten minutes and no more

Cosmetics do not govern distribution, but they do govern what happens after somebody arrives, and arrivals are expensive. Run the list, fix what is broken, and do not let it grow into a redesign project that displaces the four signals that actually matter.

The last item is the only one on that list with a distribution effect, since a comment answered within the hour behaves differently to one answered next week. Everything above it is hygiene, and hygiene is a ten minute job that people somehow turn into a quarter.

Why a free page grader can score you high while reach falls

A grader scores what it can see from outside, which is the shell of the page. Reach is produced by what the page publishes, who is attached to it, and how regularly, and none of those three are visible in a public snapshot. A high grader score with falling reach is not a contradiction, it is the tool doing exactly what it was built to do.

What the public graders measure against what governs whether posts get delivered.

Use the graders for the cosmetic pass and nothing else. They are fast, free and accurate about the thing they measure. Handing them the strategic verdict is the mistake, and it is an easy one to make because a numeric score feels like an answer regardless of what went into it.

The four signal score has the opposite property. It is slower to produce, it needs two numbers from inside the company, and it tells you the one thing to do next. That trade is worth thirty minutes a quarter.