LinkedIn Is Not Working: What to Fix First

On this page

The first silence is the only one worth fixing
Find your first silence before you touch anything
The fix everyone reaches for, and the one that fits
Seven routes to getting the fix done
The arithmetic of changing five things at once
The two-week single-variable test
Symptoms that look like LinkedIn and are not
What to do this week

The short answer

Fix the earliest silence in the chain and nothing else. If almost nobody sees your posts, rewriting your About section changes nothing, because the profile is downstream of the impression. Work out which of five stages is the first to go quiet, change one variable there for a fortnight, and only then decide whether to buy help. The routes below are ordered by which suits the most common reader first.

The first silence is the only one worth fixing

LinkedIn is a chain, not a set of independent levers. Someone has to see the post, then look at the profile, then reply, then agree to a conversation, then decide to buy. Each stage multiplies the one before it. If the first number is near zero, improving anything further down multiplies a near-zero number and produces a slightly different near-zero number.

Disclosure, before the comparison further down

Two of the services named in this article, Supersonify and Personeur, are run by the same team as this site. They are sister services, and they appear here because they match two of the routes described accurately enough to be explained rather than guessed at. Weigh that as you read them. Both entries carry a real downside like every other entry, and for most people reading this the correct first move is one of the routes that costs nothing.

That is why published fix lists disappoint. A list of twenty-two improvements presented in no particular order invites you to start at the top and work down. The top of most lists is the profile, because the profile is the easiest thing to write about. For a large share of people the profile is not the problem, and a fortnight spent on it produces a better page that the same eleven people see.

The useful question is narrower than what should I improve. It is where does the chain first go quiet, because that stage is the only place where effort converts into a visible change within a fortnight. Everything downstream of it is invisible until the upstream number moves.

Find your first silence before you touch anything

There are five places the chain can go quiet, and they fail in a fixed order. Walk them from the top and stop at the first one that is silent. That stage is your fix for this fortnight, and the four below it are none of your business until it moves.

The Five Silences

Read these in order. The first stage where the honest answer is no is the only stage you are allowed to work on. Numbers here mean your own numbers, taken from post analytics and profile views, not a benchmark from somewhere else.

Silence one: nothing is seenYour posts reach a few dozen people, most of whom are former colleagues. Nothing downstream can be diagnosed yet, because the sample is too small to read. The fix lives in distribution and cadence rather than craft: who you are connected to, how often you post, and whether you comment anywhere other than your own posts.

Silence two: seen but not openedImpressions are respectable and profile views are flat. People are reading the post and not becoming curious about the author. That is a hook and a headline problem, and it is the only silence where the profile genuinely is the fix, specifically the top two lines rather than the whole page.

Silence three: opened but no replyProfile views are healthy and the inbox is quiet. Something on the page is stopping a convinced reader, usually the absence of an obvious next action or proof they cannot check. This is a rewrite of the About section and the featured row, and nothing else.

Silence four: replies that never become conversationsPeople message you and it stalls after two exchanges. The failure is in what you offer next, not in the profile at all. A vague invitation to chat sometime dies quietly, while a specific, small, dated next step survives.

Silence five: conversations that never become workYou get calls and they end politely. LinkedIn is not the broken part. The offer, the price, the proof or the fit is, and no amount of posting fixes any of those. This is the silence that most often gets misdiagnosed as a marketing problem.

Two rules make this work. The first is that you take the earliest silence, not the most interesting one, because people gravitate to the stage they most enjoy working on. The second is that you do not skip a silence because you have already tried something there. Having tried is not the same as having tested, and the difference is the subject of the arithmetic further down this page.

The fix everyone reaches for, and the one that fits

Each silence has a fix that feels natural and a fix that works, and they are rarely the same. The instinctive fix is usually the one that requires no new information, which is exactly why it is comfortable and exactly why it fails.

Where the chain first goes quiet	The fix people reach for	The fix that fits	How you know inside a fortnight
Almost nobody sees the posts	Rewriting the profile and the headline	Commenting substantively on other people's posts daily, and posting on a fixed cadence	Impressions per post roughly double, or they do not, and either answer is useful
Impressions are fine, profile views are flat	Posting more often	Rewriting the first two lines of the post and the headline they click through from	Profile views per hundred impressions moves in either direction
Profile views are healthy, the inbox is quiet	Posting more often, again	One next action on the profile that a convinced reader can take today	First inbound message that references something specific on the page
Replies arrive and stall	Rewriting the About section	Replacing an open invitation to chat with a small, specific, dated offer	Reply-to-call rate over roughly ten conversations
Calls happen and end politely	More content of any kind	Changing the offer, the proof or the price, offline, without touching LinkedIn	Two or three calls where the objection changes rather than repeats
Everything looks fine and the pipeline is empty	Starting over with a new strategy	Checking that your audience composition matches your buyer at all	Follower and viewer job titles read against the titles you actually sell to

The instinctive fix versus the fix that matches the symptom.

Notice the second column. Two of the six rows point at posting more, and two point at rewriting the profile. Those are the two moves that get recommended for every symptom because they are the two moves everyone already knows how to make. A router exists to interrupt that reflex.

Seven routes to getting the fix done

Once you know which silence you are fixing, you have to decide who does the work. These are ordered by which route suits the most common reader of this page first, and that reader is a founder or consultant who has not yet diagnosed anything and has no budget allocated to finding out.

Change one thing yourself, for two weeks

You identify your earliest silence, change exactly one variable at that stage, hold everything else constant for a fortnight, and read the result. Then you either keep the change or discard it and try the next hypothesis.

Best forAlmost everyone reading this, and particularly anyone who has never run a clean test. It is also the only route that produces knowledge you keep, since you learn which lever moves your specific audience.

Costs youNo money. Perhaps two hours to set up and read, on top of whatever you were already spending on posting.

Watch outIt is slow, and it demands a discipline most people do not have, which is leaving four tempting improvements untouched for fourteen days. It also fails outright if you have no baseline numbers written down, so a fortnight of measuring has to come before the first test.

An AI writing tool

A subscription tool that drafts posts, rewrites profile sections and suggests hooks from a prompt or an outline you supply.

Best forSomeone whose silence is caused purely by volume, meaning they know what to say, have diagnosed the stage correctly, and simply never publish because a blank page stops them every Tuesday.

Costs youA small monthly subscription, plus your own hours editing, which are usually underestimated because the second draft takes longer than people expect.

Watch outIt fixes throughput and never diagnosis. Pointed at the wrong silence it will help you publish the wrong message more often and faster. Readers also recognise the default register more reliably than most users believe, and recognition costs more trust than the extra volume earns.

A course or structured programme

You buy a system, usually recorded, that teaches a full method for profile, posting and outreach, then you apply it yourself over some weeks.

Best forSomeone who will genuinely do the work but has no mental model at all, and who learns better from a complete framework than from assembling one out of articles.

Costs youMoney once, plus ten to twenty hours of your own time to work through it, plus the ongoing hours of doing what it says.

Watch outA course teaches the system that worked for its author's audience, which may not resemble yours. It also arrives undiagnosed, so it will happily have you rebuild all five stages when only one was broken, and you still do every hour of the work afterwards.

A freelancer for one bounded deliverable

You hire an independent writer or strategist for a single defined job, such as rewriting a profile, producing a month of posts, or auditing your audience composition.

Best forSomeone who has already diagnosed the silence, can write a two-paragraph brief describing it, and wants that one artefact produced properly without hiring anyone permanently.

Costs youMoney for the deliverable, at rates that vary far too widely across marketplaces to quote responsibly, plus about an hour of your own time on the brief and revisions.

Watch outThey will build exactly what you brief. A wrong diagnosis buys you a well-executed wrong fix at full price, and you will not find out for a month. Quality is also genuinely hard to verify in advance, so ask for previous work in your specific format before agreeing.

Personeur sister service

A LinkedIn ghostwriting studio for founders and executives, covering voice capture and a continuing publishing cadence rather than a one-off repair.

Best forAn executive whose silence is at stage one and whose honest constraint is time, meaning they know what they think, they have a market that would listen, and they have not published in eight months.

Costs youReal money on a continuing basis, plus several hours of your own at the start, since voice capture only works if you actually sit for the interviews.

Watch outIt does not fix positioning, audience or offer, and it is not a diagnosis. If your silence is at stage four or five, more well-written posts will change nothing and you will have bought a monthly cost against a problem that lives offline. It is also a continuing commitment rather than a repair, so it suits people who intend to publish for a year or more.

Supersonify sister service

Runs LinkedIn as a growth channel for B2B companies and teams, covering ads, the company page, employee advocacy and pipeline reporting, so the unit of work is a channel rather than a person.

Best forA company where the silence is measured in pipeline rather than impressions, where several employees post, and where somebody upstairs is asking what the channel produced last quarter.

Costs youA budget line rather than a fee, because ads and reporting need spend behind them, plus internal ownership from whoever runs marketing.

Watch outThe wrong size for a solo operator by a wide margin. If your problem is that one person's posts reach forty people, this is machinery you will not use. Reporting also needs a functioning sales process to attribute to, and generalist full-service agencies that treat LinkedIn as one line in a broad retainer usually produce even less, so compare within the category rather than against nothing.

Hiring someone in house

You add a marketing hire, full or part time, who owns the channel permanently and sits inside the business rather than beside it.

Best forA company that has decided LinkedIn is a permanent channel rather than an experiment, and that already has enough internal knowledge to brief and judge the hire.

Costs youThe largest fixed cost of any route here, since it is salary plus ramp time before anything ships, and ramp is measured in months rather than weeks.

Watch outA capable generalist marketer is rarely a good writer in a founder's voice, and those are different skills bought at the same salary. Hiring also converts a variable cost into a fixed one before you have evidence the channel works, which is the wrong order of operations for most companies.

The honest summary of that list is that the first entry serves most people and costs nothing, and that the paid routes only pay back once a diagnosis exists. If you finish this page and decide to spend a fortnight testing one variable yourself, the page has done its job.

What to take away

Every stage of LinkedIn is downstream of the one before it, so fixing a late stage while an early one is silent produces no measurable change and a lot of wasted effort.
The most common mistake is rewriting the profile when the real problem is that almost nobody is reaching the profile in the first place.
Changing five things at once means any improvement is unattributable, so your effective learning rate from a busy fortnight can genuinely be zero.
One change per fortnight gives you twenty-six clean tests a year, which is more than most people run in a decade of unstructured effort.
Buying help before you know which stage is broken means paying someone to execute a fix you have not diagnosed, and a well-executed wrong fix costs full price.

The arithmetic of changing five things at once

The most common way people waste a productive month is by fixing everything simultaneously. It feels like maximum effort and it is, but the information it returns is close to nil, and that is arithmetic rather than opinion.

Suppose you change five things in one week: the headline, the About section, your posting frequency, your post format, and the audience you comment among. Two weeks later, engagement has improved. Which change did it? There are thirty-one non-empty combinations of five changes, and you have run exactly one of them, with no control. You cannot attribute the result, so you cannot repeat it deliberately, and you cannot discard the four changes that did nothing. You have bought an outcome and no knowledge.

Now assume the disciplined version. One change per fortnight is twenty-six clean tests in a year. Most people never run twenty-six deliberate tests in a decade of unstructured effort, because unstructured effort produces changes rather than tests. Set against the hours already committed, the discipline is nearly free: three posts a week at forty minutes each is a hundred and four hours a year, and the router costs perhaps two of those hours in setup and reading.

The exception worth naming

There is one case where changing several things at once is correct, and that is when the earliest silence is total. If a profile has no photo, no headline worth the name and an empty About section, do not run a fourteen-day test on the headline. Clear the whole hygiene layer in an afternoon, then start testing from a working baseline. Single-variable discipline applies to optimisation, not to construction.

Around three percent of members post more than once a week, according to aggregate 2026 LinkedIn statistics reports. Cadence alone puts you in a small group, which is why stage one silences so often respond to consistency before craft.

Aggregate 2026 LinkedIn statistics reports

The two-week single-variable test

This is the procedure behind the first route in the comparison. It takes about two hours of setup across a fortnight and it is the cheapest information you will ever buy about your own audience.

Write down the baseline before you change anything

Take your last eight posts and record impressions per post, profile views for the period, and the number of inbound messages. Three numbers in a spreadsheet. Without them you will be comparing a feeling to another feeling in a fortnight's time.

Name the earliest silence out loud

Walk the five stages in order and stop at the first no. Write the stage number down where you can see it, because the temptation to drift to a more enjoyable stage arrives around day four and a written commitment beats a remembered one.

Choose exactly one variable and define what changes

Not improve my hooks, which is not testable. Something like every post opens with a specific claim about a named role, for fourteen days. If you cannot describe the change in one sentence a colleague could follow, it is too vague to test.

Hold everything else still for fourteen days

Same cadence, same formats, same profile, same commenting habits. This is the part people fail. Keep a list of the improvements you are not making, so they feel deferred rather than forbidden, and run them in later fortnights.

Read the same three numbers and make a decision

Keep the change, discard it, or extend the test if the volume was too small to read. Then return to stage one of the router and check whether the earliest silence has moved, because a fixed stage promotes the next one into your line of sight.

One caution about reading the result. A fortnight of posts is a small sample, and small samples swing. Treat a change under about twenty percent as noise, treat a doubling as real, and treat anything in between as a reason to run the same test one more fortnight rather than declaring victory.

Symptoms that look like LinkedIn and are not

Some silences have nothing to do with the platform, and no fix applied inside LinkedIn will touch them. Recognising these early saves the most money, because they are the ones that survive every route in the comparison above.

Your buyers are not on the platform in any number. Some markets genuinely are not, and a year of excellent posting will confirm it expensively.
The offer is one nobody is currently shopping for, which shows up as warm conversations that end with a request to stay in touch.
Your audience composition drifted, so your followers are peers and competitors rather than buyers, and engagement looks healthy while producing nothing.
Nobody is monitoring the inbox, which turns every upstream improvement into a slightly larger pile of unanswered messages.
The price is wrong for the proof you can show, which is a positioning problem that no amount of visibility can outrun.

The tell for all five is the same. Improvements at every stage produce movement in the metrics and no movement in revenue. If you have genuinely fixed two silences and the pipeline has not changed, stop optimising and go and examine the offer, because the platform has been telling you the truth for a while.

The audience composition case deserves a specific look, since it is the one most often mistaken for a content problem. If your profile is doing its job at every gate and the visitors are the wrong people, the gates are not the issue, which is the argument set out in treating the profile as a conversion step.

What to do this week

Nothing on this list takes longer than an evening, and running them in order is what turns a vague dissatisfaction into a testable question.

The first fortnight, in order

Record three baseline numbers: impressions on your last eight posts, profile views for the period, and inbound messages received.
Walk the five silences in order and write down the number of the first one that is quiet.
Check the hygiene layer once, and if the profile is genuinely unfinished, clear it in a single afternoon before you test anything.
Choose one variable at your earliest silence and write the change as a sentence someone else could follow.
List the four improvements you are deliberately not making this fortnight, so they stop competing for attention.
Put a date fourteen days out in the calendar to read the same three numbers and decide.
Decide in advance what result would make you buy help, and which route you would buy, so the decision is made calmly rather than in a bad week.

That last item is the one worth taking seriously. People buy help at the point of maximum frustration, which is the worst moment to judge a proposal, and they buy the route that markets to them most persuasively rather than the route that matches their diagnosis. Deciding the trigger and the route while you are calm is what makes the eventual purchase a good one. If the answer turns out to be a writer, what a profile writer actually costs covers what you should have settled before briefing anyone, and a checklist ordered by impact covers the hygiene afternoon.

Questions people ask next

What should I fix first if LinkedIn is not working?

Fix the earliest stage that has gone quiet. Walk the chain from impressions to profile views to replies to calls, and stop at the first stage where the number is near zero. That stage is the only place effort converts into a visible change, because every later stage multiplies whatever the earlier one produced.

How long should I test one change before deciding it failed?

Fourteen days at your normal posting cadence, which usually gives six to eight posts. Read the same three numbers you recorded as a baseline. Treat a change under roughly twenty percent as noise, a doubling as real, and anything between as a reason to run the identical test for one more fortnight.

Should I rewrite my profile or post more often?

It depends entirely on where the silence starts. If impressions are already low, rewriting the profile improves a page that almost nobody reaches. If impressions are healthy and profile views are flat, more posting simply repeats the failure, and the top two lines of the profile are the thing to change.

How do I know if the problem is my offer rather than LinkedIn?

Look for movement in metrics without movement in revenue. If you have fixed two stages, impressions and replies have both improved, and no work has resulted, the platform is functioning and the offer, price, proof or fit is not. That diagnosis is made offline and no posting schedule will change it.

Is it worth paying someone before I have diagnosed the problem?

Rarely, because a freelancer or a studio will execute the brief you give them, and an undiagnosed brief buys a well-made fix for the wrong stage at full price. Spend one fortnight establishing a baseline and identifying the earliest silence first. That fortnight costs nothing and makes every later purchase cheaper.

What if several stages look broken at the same time?

Then you are looking at construction rather than optimisation, and single-variable discipline does not apply yet. Clear the whole hygiene layer in one afternoon, get a working baseline, and start testing from there. Sequential testing is for improving something that already functions, not for building the thing.

Sources

Send your profile, keep the aha.

One email with your LinkedIn URL. One specific, actionable insight back, free. If it is obvious, we did not look hard enough.

Email your profile →

Or write to bksarkhedi78@gmail.com

LinkedIn is not working, so what do you fix first