LinkedIn Conversions API: Send CRM Deals Back to Ads

Two things to hold on to. The Conversions API replaces nothing: the Insight Tag still builds your retargeting audiences and captures the click identifier. And it only pays off if your CRM stages are honest, because whatever you send becomes the thing LinkedIn optimises towards.

The Insight Tag is JavaScript on your site. It records page views and on-site conversions in the browser, builds retargeting audiences and gives you the demographic reporting LinkedIn is known for. It is also a browser tag, so ad blockers, tracking protection and a declined consent banner each take a piece out of it, and it can only see conversions that happen on your website.

The Conversions API is a server to server connection. Your backend, your CRM or an attribution platform posts the event straight to LinkedIn. No script runs, so nothing can block it, and the conversion does not have to happen on your site at all: a stage change in HubSpot, a signed contract, a call your sales team logged. That is the point for B2B, where the meaningful conversion happens weeks later in a system LinkedIn cannot see.

• Insight Tag

  Browser JavaScript. Page views, on-site conversions, retargeting audiences and demographics. Blockable.

• Conversions API

  Server to server events from your CRM or backend. Nothing to block, and the conversion can happen anywhere.

• Both together

  The normal shape: the tag captures the click identifier and the audiences, the API carries the pipeline back.

Keep each conversion rule on one source, though. A rule fed by both the tag and the API can count the same action twice, and the simplest way to avoid that is to give every rule exactly one job.

Every event has to carry at least one identifier LinkedIn can resolve to a member. Two matter at person level, and a third layer works at company level.

• The li_fat_idLinkedIn's click identifier, appended to your landing page URL when first-party ad tracking is on. Capture it at the first visit and store it on the lead, the way you would a gclid.
• The hashed emailSHA-256 of the normalised address. The workhorse for CRM events, because your CRM has an email on every contact and a click identifier on far fewer.
• The account levelLinkedIn's revenue attribution reporting matches CRM accounts to companies, which is how influenced pipeline shows up when the person who clicked is not the person in the deal.

The B2B catch sits in the second one. Members sign up to LinkedIn with whatever address they like and plenty use a personal one, while your form collects the work address. Those two do not hash to the same value, so a clean setup still returns a lower match rate than a consumer business would see. Send the li_fat_id wherever you have it, and send both identifiers when you can.

The API is the same in all three cases. What differs is who maintains the connection, and how much of the journey is already joined before the event is built.

• A direct integration

  Your engineers post events to LinkedIn's API. Total control, and you own the token refresh, the retries and the identifier storage forever.

• LinkedIn's CRM integrations

  Salesforce, HubSpot and Microsoft Dynamics connect to Campaign Manager directly. Quick to switch on, limited to what the connector chooses to send.

• A platform that holds the journey

  An attribution tool already stores the click identifier and reads your CRM, so the event becomes a mapping decision rather than a build.

The question that decides it is where the join happens. A connector reading your CRM can only match on what the CRM holds, so if the click identifier was never stored on the lead record, nothing downstream can invent it. That is why the capture step matters more than the sending step, and why teams end up with a working integration and a disappointing match rate.

Everything on the LinkedIn side is a conversion rule. One per stage you intend to report, created before a single event is sent.

1. In Campaign Manager, open conversion tracking (under the analyse menu in the current interface) and create a conversion.
2. Choose the Conversions API as the data source rather than the Insight Tag, so the rule expects server events.
3. Name it after the stage, not the campaign. Opportunity created, not Q3 webinar: campaigns come and go and the rule stays.
4. Set the conversion type from LinkedIn's list, the closest one to what the stage actually is. The type shapes how the conversion is reported.
5. Set the attribution model and the click and view windows. LinkedIn currently offers click windows up to around ninety days, and the default is far shorter than a B2B sales cycle.
6. Associate the campaigns the rule applies to. A rule with no campaigns attached records nothing, and that is the most common reason a correct integration shows zero conversions.

What the event carries

• The conversion rule the event belongs to.
• At least one user identifier: the hashed email, the li_fat_id, or both.
• The timestamp of the stage change, in the format LinkedIn expects.
• The value and the currency where the stage has one. An opportunity has an amount, a qualified lead usually has an expected one.
• Your own reference for the record, so a value can be restated later without creating a second conversion.

Send the time the stage actually changed, not the time your job ran, or every conversion in the account clusters at 3am and the attribution window measures the wrong gap. Send one test event before wiring the CRM to the rule: the rule's status in Campaign Manager is what tells you the connection is live, and it moves once the first valid event lands.

The build is small. The decision is which stages leave your CRM at all, and that one deserves an argument before anybody opens Campaign Manager.

• The form fill or MQLSend it and keep it as a secondary rule. It is the volume number: useful for diagnosis, wrong as a bidding target.
• The qualified leadSales confirmed the fit. For most B2B accounts this is the rule worth reporting on, because it arrives often enough to be a signal.
• The opportunityA real deal with an amount and a close date. The first stage where a value on the event means something.
• Closed wonThe outcome, with the contract value. Send it always, and expect some of them to land outside LinkedIn's window.

Once the stages land, Campaign Manager stops being a lead counter. Cost per opportunity by campaign is the column that changes budgets, because it reorders them: the campaign with the cheapest leads is rarely the campaign with the cheapest opportunities, and on LinkedIn the gap between the two is wide.

Revenue by campaign follows, and with your CRM fields beside it you can cut the pipeline by job title, seniority and company size. LinkedIn's demographics tell you who saw and clicked the ads. Your CRM tells you who bought. Reading the two together is where B2B media planning actually happens, and no lead form report gets you there.

The account rollup problem

One caveat no integration removes. In B2B the person who clicks is rarely the person who signs: a manager downloads the guide, the VP takes the meeting, procurement signs the contract, and only the first of those three touched your ad. Match the event on the contact who actually clicked and the campaign gets credit correctly. Roll the deal up to the account and you see influenced pipeline instead, which is a different question with a different answer.

• The window is shorter than your cycle

  LinkedIn's longest click window is around ninety days. A deal closing in month six is real and Campaign Manager will not credit it. Report the earlier stage there.

• Lead Gen Forms never touch your site

  The form opens in the feed and submits without a page load, so there is no li_fat_id on your side. You get the lead through lead sync or a CRM connector, with the campaign attached, and you join on that.

• Member privacy limits what comes back

  Demographic breakdowns only appear once enough members sit behind them, and you never receive member data LinkedIn did not already hand you through the form.

• Demographics are not per-lead reporting

  The demographics report describes the audience that engaged. Per-lead reporting comes from your CRM. Mixing the two produces a slide nobody can reproduce.

The first two quietly break projects. A team maps closed won to a LinkedIn rule, watches nothing land, and concludes the integration failed when the attribution window simply expired. See attribution for a long sales cycle for the reporting shape that works around it.

Most B2B accounts run all three. The table is about what each one can and cannot tell you, not about picking one and dropping the others.

The hard part of all this is not the API call. It is holding a click identifier from an anonymous first visit until a deal closes months later, in a system that also knows the CRM stage. That is LeadJourney's normal job. One script on the site or in your GTM container, live in about 21 minutes without a developer, captures li_fat_id at the first visit along with the UTMs and the landing page, then joins that click to the person at the form fill, the call or the booking.

From there the CRM stages drive it. HubSpot, Salesforce, Pipedrive, Close, Attio, GoHighLevel, ActiveCampaign and Odoo connect natively, anything else by webhook or Zapier. You pick which stages go back as conversions, qualified lead, proposal sent, closed won or all three, and map each to a conversion rule. The deal amount and currency travel with the event, sent server to server over the LinkedIn Conversions API, and the same mapping feeds Meta, Google Ads and Microsoft Ads.

On the reporting side that gives you cost per qualified lead and cost per closed deal by campaign, revenue per campaign and landing page, and the full journey behind any single lead. Attribution models switch without re-tracking, and LeadJourney's own reports carry no window, so a deal closing in month six still shows the LinkedIn click that started it even when Campaign Manager can no longer credit it. Tracking is server-side on your own domain at 95%+ accuracy, first-party by design and hosted in Frankfurt with a DPA signed.

Read verified reviews on Trustpilot, G2 and leadjourney.io/testimonials.

4:35

Next: the best LinkedIn ads tracking tools, the best LinkedIn revenue attribution tools and attribution for a long sales cycle. For the definition, see conversions API; for the product side, the LinkedIn Ads integration and the leads behind your conversions.