What the Insight Tag does and what a conversion is
The LinkedIn Ads hub describes the Insight Tag as a small piece of JavaScript placed on every page of the website, usually through a tag manager. It is LinkedIn's tracking tagAn ad platform's browser code that reports page views and events on a site back to the platform., and it does several jobs at once: it records visitors so they can be reached again through Matched Audiences, it reports website demographics (the professional make-up of the people visiting, without identifying anyone) and it powers conversion tracking. The targeting and Lead Gen Forms guide explains how the visitors it records become website retargeting audiences, which is LinkedIn's form of RemarketingAdvertising to previous visitors or engagers..
A conversion is a definition you write in Campaign Manager that tells it which recorded moment counts as a result. Each one is a Conversion ActionA Google Ads definition of one action to count, such as a purchase or lead, with its source, category, counting method, value and conversion window.: something a visitor did that the business values, such as a demo request, a trial start or a download. The hub sets out the two ways to define one and the settings each carries.
| Part of the setup | What it is | What decides it |
|---|---|---|
| Base tag | The Insight Tag itself, loaded on every page | The tag manager trigger and the consent rules |
| URL conversion | Counts a visit to a page whose address matches a rule | The real address the browser shows after the action |
| Event-specific conversion | Counts an event the page fires when the action is confirmed | A trigger on the confirmed outcome, usually a data layer event |
| Conversion type | Labels what kind of action it is | What the action means to the business |
| Value | An optional figure attached to the conversion | Whether the business can put a fair value on it |
| Attribution settings | Which clicks and impressions get credit, and for how long | Options LinkedIn documents in its help centre |
| Campaign attachment | Which campaigns the conversion is reported against | The campaigns built to drive that action |
| Conversions API | A server route that sends the same events | Consent, browser restrictions and the size of the gap |
The attribution windows, their defaults and how long reporting takes to update are documented in LinkedIn's Marketing Solutions Help, listed in the sources, and change from time to time, so this page does not quote them. When a conversion column already reads zero, start from the Insight Tag diagnosis page, which repairs a broken setup cause by cause. This tutorial builds one that does not break in the first place.
Before you start
- The ad account, owned by the business. The Campaign Manager setup tutorial opens it under a business login and names an owner for the tag; do that first.
- Publish access to the website's tag manager container, or a developer who has it. The mechanics of tags, triggers and variables are in the Tag Manager guide and are not repeated here.
- A list of every template and subdomain that can host a conversion: the main site, a booking tool, a trial sign-up, a separate blog, a form provider's confirmation page.
- The consent position, agreed with whoever owns privacy. The privacy and consent guide sets out what has to be decided.
- A measurement plan, even a short one. The measurement plan framework turns business goals into the actions worth counting.
- A clean browser without content blockers, and access to its developer tools, for the tests.
The steps
Write down which actions count as conversions
Start on paper, not in Campaign Manager. List the actions that tell the business a LinkedIn campaign did its job: usually a demo or contact request, a trial or account start, and perhaps a download of a gated guide. For each, write where it happens, how you know it succeeded and what it means to the sales team. Keep the list short. A conversion nobody will ever optimise or report against adds noise to every column and invites the wrong campaign to claim credit. Put the list in the measurement plan, because it is the document the agency, the developer and the analyst will all read. If a lead comes through a LinkedIn Lead Gen Form instead of the website, it is counted inside LinkedIn and does not need the tag; the targeting guide covers that route.
Confirm the tag belongs to the business's ad account
Open the Insight Tag area of the business's own ad account and take the tag from there. A tag copied from an agency's account or a former employee's sends its visitors and conversions to that account, not the business's, and nothing on the website shows the difference. If the site already carries a LinkedIn tag, find out which account it reports to before adding another; two tags from two accounts make every audience and every count harder to read. The setup tutorial treats the tag as an owned asset for this reason, and the record it asks for should now list the tag and its account.
Add the base tag on every page from one tag manager trigger
Create the LinkedIn tag in the tag manager (GTMTag management system. or whichever the site uses) and fire it from a single trigger that covers every page, rather than one trigger per section. One trigger is easy to check and hard to break by accident; a set of page-specific triggers fails silently the day a new template is added. If the site already sends a confirmed event for analytics and other platforms, the tutorial on ad platform tags from one data layer shows how to hang the LinkedIn tag off the same events, so every platform counts the same moment. Preview the container, then publish it, and write down the version number you published.
Hold the tag until consent where the law requires it
Where consent is needed, the tag must wait for it, and must fire as soon as it is given. Wire it to the consent state the banner passes to the container, using consent modeA Google tag feature that adjusts what tags collect according to the visitor's consent choices. or the banner's own signals, so that the tag fires on the consent update as well as on later page loads. A banner that records acceptance but never tells the container is one of the documented reasons a conversion column stays at zero. Test both branches: with consent refused nothing should leave the browser, and with consent accepted the tag should load on the same page without a reload. The consent banner QA checklist covers the banner itself; this step only connects the LinkedIn tag to it.
Check the tag on every template and subdomain
Walk through each template and subdomain on the list from the preparation stage with the browser's developer tools open on the network tab, filtering for LinkedIn requests. The home page proves very little. What matters is the page where each conversion happens: a confirmation page on a booking subdomain, a trial sign-up on an app domain, a thank-you page served by a form provider. Where a page carries no container, add it, or move the conversion to a page that does. Record each page checked with the date; the tracking verification checklist adds the wider checks for other platforms in the same pass. When the whole site is about to change platform, templates or address structure, the tutorial Moving a Website Without Breaking LinkedIn Ads covers this template and subdomain walk, the conversion rules and the retargeting rules for that move.
Define URL conversions against real addresses
For each action that lands the visitor on its own address, complete the action yourself and copy the address the browser shows at the end. Write the URL rule against that address, not against the one in the project brief or the one somebody remembers. Real addresses carry query strings, language folders, trailing paths and subdomains, and the rule's match type has to tolerate them without matching pages that are not conversions. A rule that matches a thank-you address shared by several forms counts all of them; give each conversion its own address where you can.
Use event conversions where the address never changes
Some actions never change the address: a form that shows a success message in place, a booking widget, a sign-up step inside an app. For those, create an event-specific conversion and fire it from the tag manager on the confirmed outcome, normally a data layer event the developers push when the form or account creation succeeds. Never fire it on the click of the submit button, which counts attempts, failed validations included. Agree with the developers that the success event is part of the form's contract and is tested whenever the form changes.
Set each conversion's type, value and attribution and record them
Give each conversion the type that describes it and decide whether it carries a value. A value is useful only when the business can defend it, for example an average worth of a demo request agreed with sales; a guessed value makes every return figure in the report a guess too. Then choose the attribution settings: which touches count, clicks or also impressions, and how long after each a conversion is still credited, within the options the help centre documents. Impression credit is the territory of view-through conversionsA conversion credited to an ad that was shown but not clicked, within a set view window., and the comparison of view-through and click-through conversions explains what each kind of credit claims, and the diagnosis of LinkedIn conversions higher than the CRM shows traces what generous credit settings do to the count. Write every choice and the date into a conversion register and the measurement change log, and keep the settings stable, so that a change in the count reflects a change in results rather than in the rules.
Attach each conversion to the campaigns it measures
A conversion reports only against the campaigns it is attached to. Attach each one to every campaign meant to drive it as you build the campaign, not afterwards, and attach nothing else: a download counted beside demo requests in a lead campaign hides how few demos it brings. The LinkedIn Ads launch QA checklist makes the attachment part of every campaign's sign-off, which is the easiest way to stop it being forgotten when campaigns are copied or rebuilt.
Run a test conversion end to end
Clear cookies, open the site in the clean browser, accept the consent banner, reach the site the way a member would and complete each conversion for real. Watch the network tab for the base tag on the confirmation step and, for event conversions, the separate request when the event fires. Then wait for the reporting delay the help centre describes and check that each test appears against its conversion in Campaign Manager. A test that passes on the site but never appears in the account points to the rule, the attachment or the account the tag belongs to; the diagnosis page takes those in order. Mark each test in the register as passed, with the date, and remove or flag test submissions in the CRM so nobody follows them up.
Decide whether a server route is needed
The hub describes LinkedIn's Conversions APIServer-side event sharing with Meta. as a way of sending conversion events from a server rather than the browser, which matters as consent choices and browser restrictions thin out what the tag alone can see. Make the decision deliberately once the browser setup is proven. Compare the conversions Campaign Manager reports with LinkedIn-sourced leads in the CRM over a few weeks; a steady gap is measurement, as the guide to platform and analytics discrepancies explains, and only a gap the business cannot live with justifies the work. The comparison of client-side and server-side tagging sets out the trade-offs. How a server route is integrated, matched and deduplicated is documented by LinkedIn and left to its help centre and to whoever owns the server.
The order matters more than any single setting. Each step rests on the one before it: a value means nothing on a conversion that never fires, and a test means nothing on a tag from the wrong account. TikTok's equivalent runs the same sequence with its own names in the TikTok pixel and Events API tutorial, which helps when one team installs both.
What the tag gives you besides conversions
Conversions are the reason most advertisers install the tag, and the least of what it does over time. The recommended setup for LinkedIn traffic campaigns puts the tag on every page so that each visit adds to a website retargeting audience, and so the website demographics view shows the job functions, seniorities, industries and companies of the members who arrived. That view is the quickest check that a campaign reached the roles it was aimed at. The setup for LinkedIn consideration campaigns builds on the same audiences, carrying people who visited a pricing or case study page into the next message.
Those audiences fill slowly, which is the practical reason to install the tag early, before the first campaign spends, rather than on launch day. A tag installed months ahead gives the first retargeting campaign a pool to draw from. When the counts start to arrive, judge them as a rate against the visits that produced them; the conversion rate calculator does the arithmetic, and the account's own history is the only fair comparison.
A rostering software firm tags its demo and trial journeys (illustrative)
Byquane Rostering installs its tag once and finds a missing subdomain
Byquane Rostering, an invented company selling shift-rostering software to care homes and warehouses, plans its first LinkedIn campaigns for the following quarter. Its measurement plan names 2 conversions: a demo request, which lands on its own confirmation address, and a trial start, which happens inside the app without the address changing. The marketing lead takes the tag from the business's own ad account and the web developer adds it to the container with 1 all-pages trigger, wired to the consent banner so it fires on acceptance.
| Item | Setup |
|---|---|
| Demo request | URL conversion against the confirmation address copied from the browser, query string allowed |
| Trial start | Event conversion fired by a data layer event pushed when the account is created |
| Values | Demo request valued at £400, agreed with sales; trial start left without a value |
| Attribution | Click and impression settings chosen from the documented options and recorded with the date |
| Attachment | Demo request on the 2 lead campaigns; trial start on the 1 trial campaign |
| Consent test | Refused: no LinkedIn request; accepted: tag loads on the same page |
The end-to-end test passes for the demo request but the trial start never appears. The network tab shows why: the trial sign-up runs on a separate app subdomain whose pages carry no container at all. The developer adds the container there in the next release, the second test appears in Campaign Manager after the reporting delay, and both rows in the register are marked passed. Over the first 6 weeks of campaigns Campaign Manager reports 18 demo requests against 23 LinkedIn-sourced demos in the CRM, a steady gap the team records and decides not to chase with a server route yet.
Common mistakes
- Building conversions before deciding what counts. The account fills with actions nobody reports on, and the wrong ones get optimised.
- A tag from someone else's account. The visitors and conversions go to that account; the website looks the same either way.
- Page-specific triggers. They fail the day a template is added. Fire the base tag from one all-pages trigger.
- Testing only the home page. The tag has to load where the conversion happens, including subdomains and form providers' pages.
- Rules written from memory. Copy the real address from the browser after completing the action.
- Event conversions on a button click. They count attempts; fire on the confirmed outcome.
- Unrecorded attribution settings. Without a dated record, a change in the count cannot be told apart from a change in the rules.
- Conversions never attached to the new campaign. Attach them while building, and let the launch checklist catch the rest.
- Reaching for a server route to fix a broken rule. The server route inherits every mistake in the definitions; fix those first.
