Your campaign manager says LinkedIn brought in forty leads last month. Your salesperson looked at the list, called four of them, and quietly gave up on the rest. Nobody is lying here. LinkedIn genuinely believes it delivered forty good leads, because forty people filled in a form and filling in a form is the only thing anyone ever told it to care about.
That gap is not a reporting problem. It is a training problem. And it is costing you money every day it stays open, because LinkedIn is not sitting idle between your reports. It is actively out there looking for more people like the ones you counted.
The Algorithm Copies Whatever You Count
Every ad platform learns the same way. You tell it what a win looks like, it goes and finds more of that. This is the part that sounds obvious and then quietly ruins accounts.
If the only thing you have ever reported back to LinkedIn is "someone submitted the form," then LinkedIn's entire model of your ideal customer is built from form-submitters. Not buyers. Not people your salesperson enjoyed talking to. Form-submitters. So it studies the forty people who filled in your form, finds the patterns they share, and goes hunting for more of them. If thirty-six of those forty were students, job seekers, competitors having a nose around, and someone who wanted the free guide and nothing else, then that is the audience LinkedIn is now getting better and better at finding for you.
The frustrating part is that you already know which four mattered. That information exists. It is sitting in your CRM, or your inbox, or your salesperson's head. It just never made the trip back to LinkedIn, so the platform keeps optimizing in the dark and you keep paying for the results.
We have written before about why LinkedIn often looks like it is failing when it is really being measured wrong. That piece is about how to judge the channel fairly. This one is about the wiring underneath it, because judging it fairly is only half the job. The other half is making sure LinkedIn can actually learn.
Decide What a Good Lead Is Before You Wire Anything Up
Skip this step and everything after it is expensive noise. Before you touch a single setting, you need a straight answer to one question: what does a lead worth having actually look like in your business?
Big companies turn this into a project. There are committees, definitions of a marketing qualified lead, arguments about what separates that from a sales qualified lead. You do not need any of that. You need to be able to finish this sentence: "I would take ten more of the inquiry we got from ___."
It helps to see the whole ladder, because most businesses are reporting from far too low down it:
Impressions and clicks tell you the ads are running. That is all. A form submission tells you the page works, and then the story stops dead. What actually matters sits further up. Did the inquiry turn into a real conversation? Did that conversation turn into a quote or a proposal? Did the quote turn into money? Each rung is a better description of a good customer than the one below it, and each rung is one LinkedIn has probably never seen.
You also need a rough value on each type. A newsletter signup and a request for a quote are not the same event and should not be reported as though they were. If you sell a service where the average job is four thousand dollars and you close one in three proposals, then a proposal request is worth roughly thirteen hundred dollars to you. That number is not academic. It is the number that lets LinkedIn understand that one of those is worth more than fifteen ebook downloads.
Everyone Installs the Insight Tag and Then Stops
The LinkedIn Insight Tag is a small piece of code on your website. It is the same idea as the Meta pixel or a Google tag: it watches for the actions you told it to watch for and reports them back.
Installing it is genuinely the easy part. Google Tag Manager will do it in about ten minutes, and most website platforms have a field where you paste the ID. Once it is live it does two useful things straight away. It lets you build retargeting audiences from people who visited your site, and it reports conversions that happen on the site so LinkedIn can start learning.
Here is the catch nobody mentions when they hand you the install instructions. The Insight Tag reports from the visitor's web browser, and the browser has become an increasingly unreliable narrator. Privacy settings, tracking prevention built into browsers by default, ad blockers, cookie banners where people click reject, someone switching from their phone to their laptop halfway through deciding. Every one of those breaks the chain. The conversion still happened. Your business still got the inquiry. LinkedIn simply never found out.
So the tag is not wrong, it is incomplete. It is a decent first signal and a terrible last one. Businesses that install the tag, see conversions appearing in the dashboard, and conclude they are done are the ones most likely to be flying on partial information without ever realizing it.
Tell LinkedIn What Happened After the Form
This is where the fix starts. LinkedIn accepts conversion data reported directly from your own systems rather than from the visitor's browser. It is called the Conversions API, and the plain version is this: instead of relying on a browser to tell LinkedIn what happened, your own business tells LinkedIn what happened.
Your systems already know. Your CRM knows an inquiry arrived. Your booking system knows a call got scheduled. Your invoicing knows a job got paid. None of that depends on someone's cookie settings, which is exactly why it is sturdier.
You want both running, not one instead of the other. The tag still does useful work, particularly for building audiences. The server-side reporting fills in what the tag missed and, more importantly, carries information the tag was never able to see in the first place. A browser can tell LinkedIn that a form was submitted. Only your CRM can tell LinkedIn that the form submission turned into a customer eleven weeks later.
That distinction is the whole ballgame. Filling the gaps is nice. Enriching LinkedIn with outcomes it could never have observed is what actually changes who your ads get shown to.
The Step Eight in Ten Businesses Skip
Reporting outcomes back from your CRM is called offline conversion tracking, and by the estimate of agencies who spend all day inside these accounts, only around two in ten businesses have it set up.
Sit with that for a second, because there is a strange asymmetry hiding in it. Almost every business running LinkedIn ads happily exports data out of their CRM to build audiences. They will upload a customer list to exclude existing clients or target lapsed ones without a second thought. The pipe runs one way all the time. Almost nobody runs it the other way and tells LinkedIn how the story ended.
Mechanically, you are sending a small amount of information back: this particular lead, which arrived on this date, reached this stage and was worth this much. LinkedIn matches it up with the click that produced it and finally learns something true. Do this for a few months and the optimization has real evidence to work with rather than a pile of form submissions.
If that sounds like something requiring a developer, it does not have to. There are two ways to do it and only one of them is technical. You can wire up a live server-to-server connection that fires automatically whenever a deal changes stage in your CRM, which is the tidy long-term answer. Or you can export a spreadsheet of your closed deals and upload it to Campaign Manager yourself on whatever schedule suits you. Both send the same information and both count the same way. For a business closing a handful of deals a month, a monthly spreadsheet is a perfectly respectable answer and it needs nobody's help but your own.
There is a practical prerequisite worth knowing before you start. For LinkedIn to match an outcome back to a click, the identifying information has to survive the journey. In practice that means your website form needs to pass through whatever identifier arrived with the visitor, and your CRM needs a field to hold it. If that sounds like the sort of thing your web person can do in an afternoon, that is because it usually is.
Connect the CRM and You Finally See the Air Cover
Once the CRM is talking to LinkedIn, a report opens up that most advertisers never look at, largely because it lives in Business Manager rather than the campaign screen where everyone spends their time.
Connecting it is genuinely a matter of authorising an integration rather than a build. LinkedIn currently connects to Salesforce, Dynamics 365 and HubSpot, you need admin rights on the Business Manager account, and the data can take up to seventy two hours to show up once it is linked. That waiting period catches people out, so link it and leave it alone for three days before deciding it did not work.
What it does is straightforward and genuinely clever. It looks at the deals in your pipeline, finds the company attached to each one, and checks whether LinkedIn showed your ads to people at that company before the deal opened. Not whether they clicked. Whether they saw.
An example from an agency running this for enterprise clients makes the point better than any explanation. One account had spent twenty thousand dollars and produced eight leads. Read as a direct response campaign, that is a disaster and you would switch it off on Monday. Then they overlaid the pipeline data and found eighteen opportunities had opened at companies LinkedIn had been advertising to. Sales had opened and closed almost all of them through their own outreach. But those companies had been seeing the ads first. Marketing had been providing air cover, and the campaign that looked like a failure was quietly warming up the market that sales was walking into.
For a small business this reframes what LinkedIn is even for. If you sell something considered, where people take weeks to decide and often ring you rather than filling in anything, your ads are doing work that the click column will never show you.
One honest caveat, because we would rather you know it now than feel misled later. This report tends to be generous to LinkedIn. It takes the deal values from your CRM and attaches them to LinkedIn as though LinkedIn were solely responsible, which of course it was not. Do not read the revenue figure as a verdict. Read the account list. Knowing precisely which companies reached opportunity stage after seeing your ads is the useful part, and it is useful whether or not you accept LinkedIn's arithmetic about who deserves the credit.
The Companies Tab Is Free and Almost Nobody Opens It
There is a second report worth your time, and this one asks nothing of you at all.
Inside Campaign Manager there is a Companies view, which replaced the old Company Engagement Report at the end of 2024. It groups the companies your content has reached by how engaged they are and lets you look back over windows from a week up to six months.
The upgrade that matters is what it now counts. The engagement figure takes in both your paid ads and your organic content, so the posts going out from your company page are feeding the same picture as the ads you are paying for. If you have been posting consistently and advertising to the same set of firms, you are seeing the combined effect rather than two disconnected halves.
For a small business chasing a defined set of target accounts, this is the closest thing to free intelligence on the platform. You can see that a company you have been trying to reach for six months has moved from ignoring you to repeatedly engaging. That is not proof of anything, but it is a far better reason to pick up the phone than the alphabet.
Its real limitation is worth stating plainly: it tells you the company, not the person. You will know that a firm is paying attention without knowing who inside it is doing the paying attention. Since you sell to people rather than to buildings, that leaves the last step to you. In a small business, where you probably already know the two people who matter at that company, that gap is a lot easier to close than it is for an enterprise with a target list of a thousand.
This Breaks for Human Reasons, Not Technical Ones
Ask the people who install this for a living why it so often stalls and none of them say the technology. They say the organization.
In a larger company the website belongs to one team, the CRM belongs to another, and the definition of a good lead belongs to a third who have never fully agreed on it. Getting all three in a room is the project. Once they agree, the actual connecting takes a couple of weeks.
Read that again as a small business owner, because it is unusually good news. You are all three teams. Nobody needs to be convinced, no meeting needs scheduling, and there is no committee that owns the definition of a qualified lead except you. The thing that turns this into a quarter-long initiative at a company of five hundred is an afternoon of decisions and a call with whoever looks after your website.
That is a genuine structural advantage, and it is rare enough in advertising that it is worth taking.
Where to Start on Monday
In order, because the order matters more than the speed:
Write down the one outcome that actually matters. Not the newsletter signup. The thing that means a real customer conversation started. If you cannot name it in a sentence, nothing downstream will work, and no amount of tracking will rescue a vague answer.
Check the Insight Tag is live and actually firing. Plenty of accounts have one that was installed during a website rebuild two years ago and quietly stopped reporting.
Add "how did you hear about us?" to your inquiry form. It is the cheapest measurement you will ever install, and agencies keep finding accounts where LinkedIn reports nothing while that box is full of people naming LinkedIn. The common worry, that an extra field will hurt your conversion rate, has not shown up in practice.
Get your closed deals reported back. This is the one that changes what LinkedIn does rather than just what you know. Start with the single outcome you defined in step one and add more later.
Then open the companies report once a month. Not daily, there is nothing to see daily. Monthly, looking for accounts that have warmed up since last time.
None of this is glamorous and none of it will fix a bad offer or a weak ad. What it does is stop you from paying LinkedIn to get better at finding the wrong people, which is the default outcome for anyone who installs a tag and calls it done. Teach it what a good lead looks like and it will go and find you more of those instead. That is genuinely all this is.
We wire this up for small businesses most weeks, so if you would rather hand the whole thing over, that is what we do with LinkedIn ads. And if you would rather do it yourself, do the first three steps this week. They cost nothing and they will tell you more about your LinkedIn ads than the dashboard has managed all year.
