The most expensive click you buy, and the nine months before it pays
LinkedIn is the one paid channel where the person who clicked is very probably the person you wanted. It is also the channel whose reporting is least able to tell you whether that person ever bought — because the sale happens months later, through a committee, in a system LinkedIn cannot see.
Contents
An industrial software company spent £48,000 on LinkedIn in the first half of the year. The campaign manager reported 3,900 clicks, 214 leads from lead gen forms, and eleven conversions — eleven, because a conversion on LinkedIn is a demo request submitted within thirty days of a click, and most of their demo requests arrive later than that.
Their CRM told a different story, once somebody went and looked. Four deals closed in the same six months whose first recorded touch was a LinkedIn campaign. Contract value: £310,000. None of the four appeared as a LinkedIn conversion, because the shortest of them took 104 days from click to signature and the longest took nine months. The channel with the worst cost per lead in the marketing report had the best return in the accounts, and no report anyone was reading said so.
This post is about that gap: why it exists on LinkedIn more than on any other channel, what LinkedIn's own measurement can honestly tell you, and how to close the gap yourself with the two exports you already have.
Why the click costs what it costs
LinkedIn Ads are priced on who the person is, not what they are searching for. You pay for a job title, a seniority, a company size and an industry, and a small audience of plausible buyers is expensive by construction.
Search advertising sells intent: somebody typed the words, and you bid on the words. Social advertising in the consumer sense sells attention at scale: a large audience, cheap impressions, and targeting by interest and behaviour. LinkedIn sells something narrower than either. Its targeting runs on the fields people fill in on their own profiles — title, function, seniority, company, industry, company size, skills — which are the exact fields a B2B seller uses to describe a qualified account.
That is the reason a LinkedIn click can cost five or ten times a search click for the same product. The audience of “finance directors at manufacturers with 200 to 1,000 employees in the UK” is a few thousand people, and every vendor who sells to finance directors at mid-sized manufacturers is bidding for the same few thousand. Scarcity sets the price. It is also the reason the click is worth paying for: the person who arrives is far more likely to be somebody who can sign than the person who arrives from anywhere else.
The consequence for measurement is that the usual efficiency metrics are the wrong instruments. Cost per click and cost per lead are designed to compare channels that deliver roughly similar people at different prices. LinkedIn delivers different people. Judged on cost per lead it loses to every other channel almost every time; judged on cost per closed deal it frequently wins, and the second number is one the platform cannot produce.
On LinkedIn, cost per lead measures the price of the audience. Only cost per closed deal measures the value of it, and that number lives in your CRM, not in the campaign manager.
What the Insight Tag can see, and what it is for
The LinkedIn Insight Tag is a script on your site that recognises visiting LinkedIn members. It reports who visited in aggregate, builds retargeting audiences, and counts on-site conversions. It sees nothing that happens off your site, and it never tells you which individual came.
LinkedIn describes the tag as the piece of code that enables conversion tracking, website audiences and website demographics. Each of those is worth understanding on its own terms, because each one is honest about a different thing.
Website audiences are retargeting lists: everyone the tag saw on a given page or set of pages, available to advertise to again. Conversion tracking is a count: a rule you define (a page visited, an event fired) is credited to a campaign when a member who saw or clicked an ad triggers it within the rule's window. All three run on the same recognition, and all three stop at the edge of your website. The tag has no idea what happens after the visit, and it was never meant to.
There is a privacy consequence worth stating, because it bears on the data you will have. The Insight Tag has no cookieless or consent-denied mode: it either loads or it does not, and in the EEA it may load only after a visitor has accepted advertising cookies. Every visitor who declines is invisible to it. On a B2B site with a European audience that is a large share of traffic, and it is a share the campaign manager will report as if it did not exist.
The Insight Tag answers “are the right kinds of people visiting” well and “did this account buy” not at all. Use it for the first question and stop expecting it to answer the second.
Lead gen forms capture well and attribute nothing
A LinkedIn lead gen form opens inside LinkedIn, pre-filled from the member's profile, and delivers a lead with a real name, company and email. It solves capture. Attribution — connecting that lead to the deal it becomes — is left entirely to you.
The mechanics are the same as Meta's lead ads, which we wrote about last week: the form is native to the platform, the fields are filled from data the platform already holds, and completion rates are far higher than a landing page achieves because nobody types anything. On LinkedIn the pre-filled data is unusually good. The email is often a work address, the company is the one on the profile, and the title is the one the person chose to be known by.
The failure mode is the same one every native form has. The lead arrives in a spreadsheet or a CRM record with a campaign name attached, and from that moment its fate depends on process. If sales works the lead in the CRM and the deal is eventually created against the same contact, the thread survives. If the deal is created against a colleague of the person who filled in the form — which, in a buying committee, is the common case — or created fresh by a salesperson who never looked at the lead source, the thread is cut, and the closed deal has no LinkedIn on it anywhere.
Nothing in the campaign manager will tell you this happened. The lead was delivered; LinkedIn's part is done. Whether 214 leads became four deals or forty is knowable only by carrying the leads forward yourself.
Every attribution window is shorter than a B2B sale
LinkedIn conversion rules carry a post-click window of up to 90 days and a view-through window of up to 30. A considered B2B purchase routinely takes six to twelve months. The deal that justifies the spend falls outside every window the platform offers.
For an e-commerce purchase, 90 days is generous. For a B2B contract it is barely the discovery phase. The industrial software company in the opening paragraph had a median sales cycle of 160 days; a 90-day window caught the demo requests and missed every signature. What the campaign manager reported as eleven conversions was eleven demo requests, which are a real signal and not the one the finance director was asking about.
There are two ways to respond to this, and one of them is a mistake. The mistake is to define the conversion as something that happens quickly — a form fill, a content download, a demo booked — and then treat conversion counts as if they were revenue. They are early signals, valuable for optimising the campaign, and they say nothing about whether the deals close. A campaign can win on demo requests and lose on contracts, and vice versa, and the platform will never see the difference.
The better response is to keep the platform conversion for what it is good at, which is feeding the campaign's own optimisation with a fast signal, and to measure the deals separately, from the system that records deals. That is a reconciliation, not a tracking problem, and it is the subject of the rest of this post.
| Measurement | Answers | Cannot answer |
|---|---|---|
| Website demographics | Are the right job titles and industries visiting? | Which company visited, or whether it bought |
| Website audiences | Who should see the next ad? | Anything about revenue |
| On-site conversion rules | Did a click lead to a form or page within the window? | Anything after 90 days, or anything off-site |
| Lead gen forms | Who submitted, from which campaign? | What the lead became |
| Conversions API | Can I report an off-site event back to a click? | Whether that event was a closed deal unless you say so, inside the window |
Reconcile deals to campaigns from the two exports you have
The honest LinkedIn number comes from matching closed deals in your CRM against the leads and clicks LinkedIn delivered, at the account level, with no time limit. It takes two exports and a matching step, and the result is the one your finance team will accept.
Start from the deals, not from the leads. Export closed-won opportunities for the period, with the contact, the company, the close date and the value. Then export every LinkedIn lead — from lead gen forms, and from your CRM's own record of web form submissions where the source was LinkedIn — with the contact, the company, the campaign and the date. The order matters: starting from the leads counts what LinkedIn sent; starting from the deals counts what the business earned, and asks which of it LinkedIn was present for.
Then match. Email addresses first, normalised: lower-cased, aliases resolved, and treated as one identity per address. Phone numbers second, normalised to a single international format, because a CRM phone field and a form phone field agree on the digits and on nothing else. And then the step LinkedIn makes uniquely necessary: match on the company as well as the person.
A buying committee is three to seven people. The one who read the ad and filled in the form is often a manager researching options; the one on the signed order is a director who never clicked anything. Person-level matching credits LinkedIn with nothing for that deal, which is wrong. Account-level matching lists every touch from anyone at the company before the close date and asks which channels were present. When LinkedIn is the first touch on the account and the deal closed seven months later, that is a LinkedIn-influenced deal, and the platform's 90-day window has no say in the matter.
- Match deals to leads by email, then by phone, then by company domain — and keep the confidence of each match visible, because a company-domain match is weaker evidence than an email match.
- Do not cap the lookback. Use the earliest touch on the account, however long ago it was, and report the elapsed time as a finding of its own.
- Report the deals with no confident source as exactly that. On a B2B pipeline that bucket is large, and it is the honest denominator for every channel's claim.
- Put the cost beside the result: LinkedIn spend for the period against contract value of the deals LinkedIn was present for, and compare that ratio to the same ratio for search and for email.
This is the reconciliation CloseRev runs. It takes a closed-sales export and a lead export — including a LinkedIn lead gen download — normalises the phone numbers and email addresses, matches at the confidence each piece of evidence supports, and reports revenue by channel with the unmatched share shown honestly as Direct / Unknown. Nothing is installed on your site and no window applies: a deal that closed nine months after the click is credited to the click.
Two numbers to stop reporting, and two to start
Cost per lead and platform conversions flatter every channel except LinkedIn and mislead about all of them. Cost per closed deal and time to close, from your own records, are the numbers that survive a finance review.
Stop reporting cost per lead as if it compared channels. It compares the price of audiences, and LinkedIn's audience is the most expensive because it is the most specific. Stop reporting platform conversions as if they were revenue. They are early signals inside a window, and on LinkedIn the window ends before the deal begins.
Start reporting cost per closed deal by first-touch channel, from the reconciliation above, over a period long enough to contain your sales cycle. And start reporting time to close by channel, because it is the number that explains the other one: a channel whose deals take nine months will look like a failure for eight of them in any report with a shorter horizon.
The channel with the worst cost per lead in the marketing report had the best return in the accounts, and no report anyone was reading said so.
None of this is an argument that LinkedIn always pays. Plenty of campaigns deliver expensive leads that never close, and the reconciliation will say so as plainly as it says the opposite. The argument is that the campaign manager cannot tell you either way, that the tools it offers measure capture and early intent rather than revenue, and that the revenue answer is sitting in an export you can pull this afternoon.
Judge LinkedIn on the deals it was present for, matched at the account level with no window — or do not judge it at all, because every other number it gives you was never about revenue.
Questions people actually ask
- Why are LinkedIn Ads so much more expensive per click than other channels?
- Because you are paying for who the person is rather than what they are looking for. Targeting by job title, seniority, company size and industry narrows the audience to people who can plausibly buy, and a small audience of buyers is priced accordingly. The click is expensive; the question is whether the deal behind it justifies the price, and LinkedIn's own reporting cannot answer that.
- What does the LinkedIn Insight Tag actually measure?
- It records visits to your site by LinkedIn members and reports them in aggregate — which job functions, industries and company sizes visited — and it lets you build retargeting audiences and count on-site conversions. It does not tell you which individual visited, and it cannot see anything that happens off your site, including the deal that closes in your CRM.
- Do LinkedIn lead gen forms improve attribution?
- They improve capture. The form is pre-filled from the member's profile, so completion rates are high and the lead arrives with a real name, a real company and a real email. Attribution is then a matter of carrying that lead through your CRM to the closed deal, which is your job, not LinkedIn's — the form only tells you the lead came from LinkedIn.
- What are LinkedIn's conversion attribution windows?
- Conversion rules on LinkedIn carry a post-click window and a view-through window that you set when you create the rule; the click window can run up to 90 days. A B2B purchase that takes six to twelve months from first click to signed contract falls outside any window LinkedIn offers, so the closed deal never appears as a LinkedIn conversion however the campaign performed.
- How do you attribute a B2B deal when several people from the same company clicked?
- At the account level, not the person level. Buying committees are the norm, and the person who clicked the ad is often not the person who signs. Match the closed deal to the company, list every touch from anyone at that company, and credit the channel that was present — rather than looking for the one individual whose click matched the one email on the contract.
- Is LinkedIn worth it if the platform cannot prove the ROI?
- Often yes, and the proof comes from reconciling closed deals against LinkedIn's lead exports rather than from the campaign manager. When a channel brings in ten leads a month at a high cost per lead and two of them close at a five-figure contract value, the cost per lead was never the number that mattered.