You select the lead generation objective, you optimise for qualified leads, you leave the bidding on automatic, and you are billed per impression. Not per lead. Never per lead.
That is not an interpretation. It is the pricing-by-objective table LinkedIn publishes, and the lead gen form documentation states it separately in plain words: ads carrying one of those forms can be purchased on a cost-per-click or a cost-per-impression basis, and cost per lead is not available.
Widen it out and the picture gets sharper. There are three bidding strategies, and two of them bill on impressions across almost every objective. Only manual bidding charges for the event you selected, and even then no setting anywhere in the table charges you for a lead or for a conversion.
The consequence for anyone signing off a budget is one sentence: the delivery risk is entirely yours. The channel is paid whether the mechanism produces anything or not.
This page covers the mechanics of the channel itself. For the cost question, and the one English-language benchmark that declares its sample, see LinkedIn Ads vs Google Ads.
One more thing worth having up front, because almost nobody accounts for it: this account is not the same product in every market. If your audience is in the United States or the United Kingdom you have targeting a European competitor cannot buy, and that competitor can see things about your delivery that you cannot see about theirs. We get to that below, with three live ads opened side by side.
What you are billed on, strategy by strategy
The pricing-by-objective table is the most useful document in the whole of this channel’s help centre, and the least circulated. It separates the objective you select from the event you are charged for, and the two almost never coincide.
There are three bidding strategies, and two of them bill on impressions.
Maximum delivery is the automated strategy, and you enter no bid at all. On brand awareness, website visits, engagement, video views, lead generation, website conversions and job applicants, it is billed on impressions. The single exception is the sponsored messaging formats, which are billed on sends.
Cost cap lets you set a ceiling on your cost per result. It is also billed on impressions.
Manual bidding is the only strategy where you pay for the event you selected: landing page clicks, engagement clicks, video views, or sends.
So the rule, stated plainly: unless you bid manually, you are buying impressions. And even when you bid manually, there is no setting anywhere in the table that bills you for a lead or for a conversion.
The documentation says the same thing in words. On lead gen forms: ads carrying one can be purchased on a cost-per-click or a cost-per-impression basis, and cost per lead is not available. The cost per lead in your report is a ratio you computed after the fact. It was never a purchase price.
One table entry is worth reading twice. Lead generation, optimising for qualified leads, on maximum delivery, is billed on impressions. That is the configuration most accounts run. The delivery risk sits entirely with the advertiser, and the channel is paid whether the form is ever filled in or not.
A note for anyone working from an older guide. The strategy formerly called target cost has disappeared from the current table. Any guide that still presents four bidding strategies is describing a product that no longer exists.
And “click” is not one thing
This matters more than it sounds, because it is where quoted costs come apart.
Engagement clicks include landing page clicks, clicks to the LinkedIn Page, social actions and Page follows. For lead generation, clicks include the form, the LinkedIn Page and the social pill. Only landing page clicks mean somebody arrived on your site.
The size of that gap is measurable. On one agency panel of roughly 950 accounts, covering 60 million dollars of spend between September 2025 and September 2026, the median cost per click on LinkedIn’s own definition was 4.19 dollars, while the median cost per landing page click on the same accounts over the same period was 17.54 dollars. Same panel, same months, a factor of four between two numbers both called a cost per click. That is an agency’s own book and not a market reference, which is exactly why it is useful here: it shows what the definition does, not what the market costs.
For the cost question itself, and for the one benchmark in English that declares its sample, see LinkedIn Ads vs Google Ads.
And a definition that decides what you buy when you run video: a video view is two seconds or more of continuous playback with the video at least 50% in view. Two seconds, half visible, is a billed view.
Sponsored messaging is billed on send. Not on open, not on reply. On send. The channel is paid at the moment the message leaves, whatever becomes of it.
One change that breaks your own history. As of October 2026, sponsored messaging impressions and eligible clicks are reported in Campaign Manager for the first time. For any account mixing messaging with Sponsored Content, that shifts CPM, CPC, click-through rate, reach and frequency. A year-on-year comparison of those metrics breaks at October 2026, and it will look like a performance change when it is a counting change.
The floors that are published, and the one everybody gets wrong
What is published comes to a short list. The minimum daily budget is 10 dollars, for any format. The minimum lifetime budget for a new campaign is 100 dollars, and after launch it becomes 10 dollars multiplied by the number of days scheduled. The minimum audience size for a campaign to deliver at all is 300 members, with 50,000 recommended. And the minimum unspent budget required to launch a new ad set is 100 dollars, or the local equivalent.
That last one is where the dollar figures stop being universal. The channel does publish the currency table for it: 80 pounds, 100 euros, 100 Canadian or Australian dollars, 100 Swiss francs, 8,000 yen, 5,000 rupees, 200 reais, 150 Singapore or New Zealand dollars, 800 rand, 600 Norwegian or Danish kroner, 700 Swedish kronor, 800 Hong Kong dollars, 600 yuan. Note the pound figure: 80, not 100. For the daily and lifetime minimums, no such table exists.
Now the floor bid, because this is where the advice you have read is wrong in both directions.
LinkedIn publishes an absolute floor, and it is far lower than anything in circulation: the lowest bid that can be placed on manual bidding is one cent, or the equivalent in local currency. The two dollar minimum bid that appears across agency blogs does not exist in the help centre.
What is not published is the bid that actually wins anything. The floor is one cent; the clearing price is set by an auction against everyone else targeting the same members, and the channel states that the cost required to win depends on the desirability of your target audience. So both of these are true: there is a published minimum bid of one cent, and there is no published figure for what delivery costs. An article quoting a fixed two dollar floor is wrong. An article quoting an average cost per click is quoting a third party.
Two more caveats on the floors themselves.
The daily budget is not a ceiling. The documentation states that a single day’s actual spend can run up to 100% above the daily budget, the only guarantee being that a week will not exceed seven times it. A 50 dollar daily budget can produce a 100 dollar day.
And the 10 and 100 dollar figures live on a marketing page, not in the help centre. That page still uses the old vocabulary of campaigns rather than the current ad sets, and carries no visible date. The current help articles do not repeat the figures: they say only that the minimum depends on the duration of the campaign, and the interface tells you the number when you type in something too low. Treat 10 and 100 as the published figures they are, and read the real one off the interface.
The same ad account is a different product depending on where your audience is
Here is the part of this channel that English-language coverage almost never states, and it runs in the direction most readers will not expect: an advertiser targeting the United States has more targeting in the account than one targeting Europe, and discloses far less of what it is running.
Both halves come from the platform’s own documentation, and the second half we checked by opening the ad library ourselves.
The vocabulary first, because it is doing real work. LinkedIn’s term for the restricted perimeter is “Designated Countries”, and it means the European Union, the European Economic Area and Switzerland. The United Kingdom is named separately, which places it outside that perimeter. Every restriction below that is documented for Designated Countries therefore does not apply to a British audience.
Group member targeting was removed in Europe, and still exists elsewhere. The documentation is dated: member groups are no longer available as a targeting attribute for ad sets created from 15 May 2024 onward, for the European Economic Area and Switzerland. It remains one of the few attributes the platform states uses no inferred or modelled data at all, which is what made it valuable.
The origin of that removal is worth stating exactly, because it is routinely misreported. The European Commission did not find an infringement. Civil society organisations filed a complaint arguing that the feature could allow members to be targeted on the basis of special categories of personal data, such as racial or ethnic origin, political opinions, religious beliefs or trade union membership, revealed by which groups a person joins. The Commission sent a request for information, and LinkedIn disabled the feature of its own accord. Commissioner Breton recorded the step on 7 June 2024: “LinkedIn has fully disabled the feature allowing advertisers to target LinkedIn users based on their LinkedIn Group membership in the EU Single Market.” No penalty was imposed.
Connected television exists here, and not for everyone. Its targeting is limited to the United States and Canada. It is not available for a European or a British audience. And it has just become easier to buy for those who can: since 22 October 2026 the format is also available through Google’s Display and Video 360, for American targeting.
Message ads carry a consent gate in Europe and none elsewhere. They were unavailable to European members: new campaigns became impossible on 15 December 2021, existing ones were suspended on 10 January 2022. Since mid-October 2024 the channel supports European Economic Area and Switzerland targeting again, on a condition it repeats across four separate documentation pages: “only EEA and Switzerland members who have opted in”. A consent screen has been offered since August 2024, where members choose between receiving sponsored messages in their inbox and seeing a banner above it. The opt-in rate is published nowhere, so a European advertiser knows the format is open and cannot know what share of the market it reaches. An American or British advertiser has no such gate and no such unknown.
Several attributes carry an explicit European caveat. For location, gender, interests, traits, age and matched audiences, the documentation states that reach may be limited when targeting within the European Economic Area or Switzerland. The extent is never quantified. You know your reach is reduced. You do not know by how much.
Audiences built from your own file are subject to consent there. Only members who have given consent are matched, and the matched audience size may be limited in the European Economic Area and Switzerland. The consent rate is not published. A file of 5,000 contacts therefore does not produce an audience of 5,000 people, and you will never learn how much of the shortfall is refusal and how much is simple absence from the platform.
Minors are handled differently too. No audience including under-18s can be selected anywhere. But in Designated Countries and in Brazil, the platform additionally tries to exclude from audiences those members whose profile suggests they are under 18. That proactive exclusion is not announced for the United Kingdom or the United States.
And the exchange for all of this is disclosure, which runs the other way. We opened the ad library and looked at three live ads from the same advertiser, on 3 November 2026. The ad targeted at the United Kingdom showed the format, the advertiser and the payer, and nothing else. The ad targeted at the United States showed exactly the same three fields. The ad targeted at Germany showed those three plus the delivery period, the total impressions in a bracket, the breakdown by country, and the targeting parameters used.
So the United Kingdom takes the fuller targeting product and publishes nothing about it, which is the position an American advertiser is in as well. A European advertiser has the thinner product and a public record of everyone else’s. If you run one account across the United States, the United Kingdom and the European Union, you are running three products through one interface, and you should stop comparing their numbers to each other.
Two British specifics while we are here, because they catch people out. Financial services can only be promoted by advertisers authorised by the Financial Conduct Authority, and British advertisers are held to the CAP Code. Prescription medicines, by contrast, can only be promoted in the United States and Canada.
Measurement, and the three attribution windows that disagree with each other
The conversion tag does three jobs: it tracks conversions, it builds retargeting audiences from site visits, and it feeds audience analytics. On retention the documentation is precise: members’ “direct identifiers are removed within seven days in order to make the data pseudonymous”, and “this remaining pseudonymized data is then deleted within 180 days.”
A measurement asymmetry worth knowing if you also run Microsoft ads. LinkedIn can receive information from Microsoft’s UET tag, and its own FAQ states that data from the European Economic Area and Switzerland is excluded from what it receives. Same tag, same setup, and a European audience produces a thinner signal than an American or a British one.
Two obligations are worth knowing before you place the tag. The first: it is the advertiser’s job to obtain the necessary consents, and the platform states explicitly that it does not provide legal advice. The second is read far less often: you are not permitted to install this tag on pages that collect sensitive data, which the documentation names as consumer health or finance pages, account management and medical appointment booking. A healthcare business that places the tag on its appointment funnel is in breach of the channel’s own terms, independently of whatever local law says.
The server-side interface exists. It requires generating a token in Campaign Manager, creating a conversion rule, then wiring the integration. That means development resources, or a partner. It is the same trade-off as on every other channel, and we have set it out in pixels, browsers and server-side conversions.
Now the part that catches people out: the default attribution windows depend on which of the platform’s own pages you read.
The page the channel devotes to attribution windows lists four possible values, 1, 7, 30 or 90 days, and announces 90-day click and 90-day view as the Campaign Manager default. Another page from the same channel still gives 1, 7 or 30 days. And the developer documentation gives 30-day click and 7-day view.
Those are three answers to one question, published by one publisher. The “30 days and 7 days” that circulates everywhere in English is the value of the programmatic interface, not the value of your account. Read the figure your own Campaign Manager actually applies rather than taking it from an article, this one included.
Two attribution models are offered, both last-touch, one at the level of each ad set and one at the last ad set. A repeated conversion inside the window counts once, except for purchase and add-to-cart, which count every event. And a share of conversions is modelled rather than observed.
Finally, brand measurement is out of reach for a small company, and the figure is published. The tool that measures a campaign’s effect on awareness imposes a published minimum cumulative budget: 60,000 dollars for one question, 90,000 for two, and 45,000 more per additional question, over 14 to 90 days, for one to six questions. In other words, the only brand effect measurement this channel offers starts at a budget most companies do not commit in a full year.
And in the European Economic Area, even paying does not buy complete measurement: since 24 February 2024 it does not cover a small share of traffic from members who saw dynamic, text or message formats.
The ad library, and what it shows less of outside the European Union
This is the most underused competitive research tool on the channel, and it has a peculiarity: it shows more about ads targeting the European Union than about any others.
What is exposed everywhere: the ad preview, its format, the advertiser name, the payer name, and a notice where an ad is restricted. For restricted ads, the channel states that it does not display the preview, the advertiser name or the payer name, “in order to comply with the law”.
What is exposed only for ads targeting the European Union: the impression range, the targeting parameters that were used, and the delivery dates. The channel writes it without hedging: this additional information is provided “to comply with the law”.
Put plainly: the transparency you enjoy on your European competitors is an effect of European law, not a favour from the platform. An American advertiser researching American competitors sees the creative and learns nothing about how it was delivered or to whom.
The depth is bounded at both ends. The library covers ads delivered after 1 June 2023, and an ad stays in it for one year after its last impression. Past that, it is gone. Serious competitive monitoring on this channel is therefore sampled at intervals and archived, because what you do not record today will have disappeared in twelve months.
And payer search is the filter nobody uses. It finds campaigns run on behalf of a brand by an agency, which do not necessarily appear under the brand’s own name.
How many people, exactly
Start with the fact that decides how to read every figure that follows: LinkedIn publishes no member count for any individual country. Not for the United States, not for the United Kingdom. What it publishes is a worldwide total and a regional breakdown, and it publishes those in three incompatible versions.
Version one, for investors. The results release of 28 October 2026 gives more than 1.4 billion members, and the press pages break that into more than 296 million for North America, more than 450 million for Europe, the Middle East and Africa, more than 427 million for Asia-Pacific and more than 233 million for Latin America.
Version two, still live. The corporate about page continues to say more than 1 billion members.
Version three, the one an advertiser reads. The audience page inside the advertising section gives North America 233 million, Europe 257 million, Asia-Pacific 277 million, Latin America 156 million, and the Middle East and Africa 60 million. It carries no date.
Set the first and the third side by side: the page written for advertisers puts North America 63 million members below the page written for investors. Neither number is an audience. “Members” counts accounts that were created. An account opened in 2015 and never reopened is still in there.
So for a media plan, use measurement, and there are three sources worth having.
The European Union: 56.4 million. That is the number of average monthly active recipients the platform declares under the European Digital Services Act, for January to June 2026, published on 31 August 2026. The country detail runs France 12.6 million, Germany 7.9 million, Italy 6.5 million, Spain 6.1 million, the Netherlands 5.4 million. One reading note, and it holds for every platform under that text: the figure is declared by the platform itself, and the Commission reproduces it as such. It is not an audited measurement.
The United Kingdom: 18.4 million, and two minutes. The regulator measures this one. Ofcom’s Online Nation 2025 report puts LinkedIn at 18.4 million monthly visitors in May 2025, which is 38% of online UK adults, and 4.7 million visitors on an average day. Measurement is by Ipsos iris, on UK online adults.
Two things in that table matter more than the headline. The channel is not growing in the UK: 18.9 million in 2023, 18.3 million in 2024, 18.4 million in 2025, a change of plus one per cent across three years. And each visitor spends two minutes a day there, with 58% of total minutes on the platform spent by men. For comparison, the same report has Reddit at 29.4 million monthly visitors and growing 28%, and TikTok at 27.3 million growing 14%. More British adults visit Reddit, or TikTok, in a month than visit LinkedIn. It ranks seventh.
The United States: 31% of adults. The Pew Research Center survey of 5,511 American adults, run between 30 January and 18 June 2026 with a margin of error of plus or minus 1.8 points, finds 31% saying they ever use LinkedIn. The breakdown is the part worth keeping: 41% of 18 to 29 year olds, 37% of 30 to 49, 32% of 50 to 64, and 13% of those 65 and over. By education, 51% of college graduates. By income, 45% of households at 100,000 dollars and above.
What to do with all this. The American figure is the only one that states the argument for the channel in a form an executive can test: the platform concentrates in the graduate, higher-income part of the population, 51% against 31% overall, and that comes from an independent survey rather than from a platform deck. The British figure is the warning that belongs next to it: wide monthly reach, two minutes of daily attention, and no growth in three years. Reach and attention are not the same purchase, and this channel sells a lot of the first.
And one figure to throw out on sight. Several widely cited pages put LinkedIn’s American user base at around 287 million, which is 82.5% of the population of the United States. That is cumulative advertising reach presented as people, and it cannot be reconciled with a survey of 5,511 adults finding 31%. If a source hands you a platform penetration above 80% of a national population, it is reporting an ad tool’s estimate, not usage.
What this advertising model has already cost the platform
An executive putting budget into a channel has an interest in knowing whether that channel’s advertising model has already been ruled on, and on what ground. Here the answer is yes, and the amount is large.
310 million euros. That is the penalty imposed by the Irish Data Protection Commission, notified on 22 October 2024 and announced on the 24th. It breaks down into three fines, together with a reprimand and an order to bring processing into compliance.
The ground is precisely the subject of this article. The platform had no valid legal basis for behavioural analysis and targeted advertising: consent was found to be neither freely given, nor informed, nor specific, nor unambiguous, and neither legitimate interest nor contractual necessity was accepted as a fallback. Failures of transparency and of the fairness principle were found alongside.
The origin is French. The case starts with a complaint filed in 2018 by La Quadrature du Net with the French data protection authority, passed to the Irish authority as lead supervisory authority because the platform’s main European establishment is in Ireland.
The file is not closed. The platform appealed. A preliminary Irish High Court ruling in April 2026 settled a procedural question: the appeal goes to the amount of the fine only, and not to the finding of infringement itself. The merits have not been retried.
Why this matters to an advertiser outside Europe. Not because the jurisdiction reaches you. Because the thing that was found to lack a lawful basis is the targeting mechanism itself, in a market of some 450 million people, and because product changes follow findings of this kind. The group targeting removal above is what one such finding looks like from inside the ad account.
And a correction, because it circulates. This penalty is not Luxembourgish. We found no decision from the Luxembourg authority against this platform; the confusion probably comes from a well-known decision by that same authority against a different company.
What gets your ad rejected
The channel reviews every ad before delivery, “usually within 24 hours”, and recommends creating the campaign group at least 48 hours before the start date. Worth knowing before you pin a launch to a date.
The published rejection reasons are concrete and the list is short: non-standard spelling or grammar, excessive capitalisation or emoji, unrelated hashtags, pop-ups on the landing page, naming the platform in the creative or implying affiliation with it, test ads, a price or offer that does not match the destination page, creative language different from the audience language, a displayed address different from the real one, a broken back button, unclear billing, and no privacy policy.
Restricted sectors deserve a check in advance if you fall into one: alcohol depending on the country, dating, contests, fundraising, pharmaceuticals, medical devices, financial services in the broad sense, and crypto assets. That last case is the most tightly controlled: prior authorisation, targeting limited to around ten countries, and only to financial professionals.
Political advertising is prohibited worldwide on this channel. In the European Union the policy aligns with the 2024 European regulation on the subject, and any ad targeting the Union requires the advertiser to declare by checkbox whether it is political. An ad declared as such is rejected. The documentation states conversely that job postings and promoting the organisation as an employer are not treated as political.
On targeting, finally, sensitive categories are prohibited: political opinions or affiliation, racial or ethnic origin, health, genetic and biometric data, religious or philosophical beliefs, criminal record, sex life or sexual orientation, trade union membership, and income. Gender and age bracket remain available, after ticking an attestation of non-discrimination in employment, housing, education and credit.
What does not hold up
“You can buy leads on LinkedIn.” You cannot. The documentation says so in as many words: lead gen form ads are purchased on a cost-per-click or cost-per-impression basis, and cost per lead is not available. Any cost per lead you report is a ratio you built after the fact.
“The LinkedIn cost per lead is X.” There is no primary source for any such figure. The channel publishes no amount and does not sell per lead. We have gone through where the English-language figures come from, and which single benchmark declares its sample, in LinkedIn Ads vs Google Ads.
“Message ads are banned in Europe.” Out of date since October 2024. They can be delivered in the European Economic Area and Switzerland again, to opted-in members only. Any advice repeating the ban predates that reopening.
“The default attribution windows are 30-day click and 7-day view.” That is the value published in the developer documentation. The channel’s own attribution window page announces 90 and 90 in Campaign Manager. Read your own account.
“The ad library lets you see competitors’ budgets.” It does not. It exposes an impression range, and only for ads targeting the European Union, never a spend figure. No mainstream advertising channel publishes its advertisers’ spending.
“LinkedIn Ads is only for big budgets.” The published floors say the opposite: 10 dollars a day and 100 dollars lifetime for a new campaign. The real constraint is not the entry ticket, it is measurability. With a long sales cycle and low volume you will not accumulate enough events to decide anything. That is a reading problem, not a price problem.
“Connected TV is available on LinkedIn.” Only if your audience is in the United States or Canada and you set English as the audience language. For a European or British audience the format cannot be bought.
“LinkedIn was fined by the Luxembourg authority.” It was not. The 310 million euro penalty of October 2024 came from the Irish authority, lead supervisory authority because the platform’s main European establishment is in Ireland.
How we work on this
We come to this subject in two situations: a company spending on LinkedIn without knowing what it is being billed for, or a company hesitating to start because it has been quoted a cost per lead it cannot find anywhere.
That is the day job of our paid acquisition work, and the method is four checks.
We establish the billed event campaign by campaign. It is the first thing to know, and almost nobody knows it on their own account.
We rebuild what the reported cost per lead actually covers. It is a ratio calculated after the fact. We reconstruct it from the real events, to see what it counts and what it leaves out.
We separate declared targeting from inferred targeting. An audience built on deduction is not steered like an audience built on declared data, and the gap between the audience you were quoted and the audience you reached often comes from there.
We record your competitors’ ad library entries and we archive them. Since ads drop out after a year, a dated record beats an occasional look.
Where to start
Open one campaign’s detail and find the billed event. Impressions, clicks or sends: that is what you are buying, and it is not the objective on the label.
Rebuild your cost per lead. Spend divided by leads, over a period long enough to read. It is an indicator, not a price.
Read the attribution window your own Campaign Manager applies. Not the one in an article. Three different values are in public circulation from the same publisher.
List your targeting criteria and mark the inferred ones. Company name, job title, company size and gender are deductions.
Record your three main competitors’ ad library entries, and date the record. In a year, those ads will not be there.
If your audience is in the United States, add one more: check whether the extra transparency you have read about applies to you. Impression ranges and targeting parameters are shown for ads targeting the European Union. On American competitors, you get the creative and nothing else.
The same profile data is sold inside Microsoft’s search channel, and the market list there excludes the United Kingdom: the Microsoft Advertising file.
In short
- Billing per lead does not exist on LinkedIn. The documentation states that lead gen form ads are bought per click or per impression, and that cost per lead is not available.
- The lead generation objective on automated bidding is billed per impression. So is website conversion on automated bidding. The delivery risk is entirely yours.
- A video view is two seconds of continuous playback with the video at least 50% in view. Message ads are billed on send, not on open and not on reply.
- The published floors are 10 dollars a day, 100 dollars lifetime for a new campaign, and 300 member accounts. A single day’s spend can exceed the daily budget by 100%.
- There is a floor bid, and it has no fixed value. It is computed per campaign. Any article quoting a fixed minimum bid is inventing a number.
- The same account is three products. Group targeting is gone in the European Economic Area, connected TV exists only for the United States and Canada, and message ads reach only opted-in members in Europe.
- The United Kingdom gets the fuller targeting and the thinner measurement, because it is outside the EEA for targeting rules and inside the exclusion list for the conversion tag’s pseudonymous identifier.
- Three different default attribution windows are published by the same publisher. The 30 and 7 everyone quotes is the API value.
- Brand lift measurement starts at 60,000 dollars of cumulative budget for one question.
- The ad library shows impression ranges and targeting parameters only for ads targeting the European Union, and the platform says plainly that this is to comply with the law.
- The targeting model has already been found to lack a lawful basis, in a 310 million euro Irish decision of October 2024 whose appeal concerns the amount only.
The billed event is the one fact about this channel that is knowable, free to check, and almost never known on the account that pays for it. Request a quote, or see how we approach B2B paid acquisition.