Does at least 80% of this stage happen within 30 days of the lead?
Use your own lead-to-stage dates. The 30 days is LinkedIn’s rule; the 80% is our suggested working cut-off, not a platform threshold.
Guide · LinkedIn Ads signals
LinkedIn Conversions API: send qualified leads LinkedIn can learn from
The Conversions API doesn’t make LinkedIn smarter by itself. It gives LinkedIn a better definition of a good lead. That only helps if the event arrives within LinkedIn’s window, often enough per campaign, and means the same thing every week.
What does the LinkedIn Conversions API do for B2B lead generation?
Short answer
It’s LinkedIn’s server-to-server connection for sending conversions from your systems, such as a CRM, website backend or event list, to Campaign Manager. B2B teams use it to show LinkedIn which leads became qualified, so campaigns can be measured on them and, with qualified leads optimisation, delivered towards people more likely to become pipeline.
LinkedIn describes the Conversions API as a way to connect “both your online and offline data to LinkedIn so you can see how your campaigns influenced actions taken on your website, sales completed over the phone, or leads collected in-person at an event” (LinkedIn Help). For a B2B advertiser, that means the stages kept in your CRM or spreadsheet, such as sales-accepted, qualified or opportunity, can be credited to the LinkedIn campaigns that produced them.
It does two different jobs, and it helps to keep them apart:
- Reporting. Any conversion rule you stream events into shows up in Campaign Manager’s Conversions & Leads columns, with cost per conversion and, if you pass values, return on ad spend. You can create a rule for each CRM stage.
- Optimisation. Only events on a rule with the qualified lead conversion type are eligible for the qualified leads optimisation goal. That is the part that changes who LinkedIn shows your ads to (Microsoft Learn).
Events are matched to LinkedIn members using identifiers you send. The API accepts a SHA-256 hashed email, LinkedIn’s first-party click ID, first and last name, external IDs, and a lead field that references a Lead Gen Form response directly (Microsoft Learn). Only matched events can be used for attribution and optimisation, so LinkedIn advises sending more than one identifier where you have them.
If you also run the Insight Tag, LinkedIn deduplicates: when the same event arrives from both with the same event ID, it counts once. Microsoft Learn notes that the Insight Tag event is the one kept.
Platform rules
What does qualified leads optimisation require?
Short answer
Use a Lead Generation objective campaign with the qualified leads optimisation goal, and send a qualified-lead event through the Conversions API or CRM Sync, not both. LinkedIn says the data must arrive within 30 days and recommends five or more qualified leads within two weeks. Pick a stage that meets both conditions.
| Requirement | What LinkedIn says | What it means for a B2B account |
|---|---|---|
| Objective and goal | Choose the lead generation objective, then select the qualified leads goal | This works with Lead Gen Form campaigns. Website Conversions campaigns can still report qualified leads, but don’t use this goal. |
| Conversion type | Select the qualified lead conversion type for events to be eligible | A rule named “SQL” with the Lead type won’t train the goal. Check the type, not the name. |
| One data source | “We recommend you don’t use both Conversions API (CAPI) and CRM data connected through your Business Manager using CRM Sync qualified lead events” | Choose one path for qualified-lead events and switch the other off. |
| Timing | Qualified leads data must be shared within 30 days to be included | A stage that usually happens on day 35 can’t steer delivery. Send as soon as the stage is set, not in a monthly batch. |
| Learning | A two-week learning phase for peak performance | Don’t judge or edit the campaign inside the first fortnight. |
| Volume | “We recommend uploading five or more qualified leads within two weeks” | Count per optimised campaign. Three campaigns sharing eight qualified leads a fortnight will each learn slowly. |
| Reporting delay | Qualified lead metrics are delayed because they depend on your event source | Yesterday’s cost per qualified lead will usually look worse than it ends up. |
Insight Tag, Conversions API or CRM Sync: which do you need?
Short answer
Usually keep the Insight Tag and add one server-side route for qualified leads. The Insight Tag captures website activity in the browser. The Conversions API sends events from your servers or CRM, including offline outcomes. LinkedIn deduplicates events sent through both, so the combination gives better coverage without double counting, provided event IDs are consistent.
| Insight Tag | Conversions API (direct build) | Conversions API (partner) | CRM Sync | |
|---|---|---|---|---|
| Data source | Browser on your website | Your server, CRM or data warehouse | A tool that already holds your data | CRM connected through Business Manager |
| Offline stages | No | Yes | Yes | Yes |
| Lead Gen Form leads | No, forms stay on LinkedIn | Yes, via the lead reference or hashed email | Depends on the partner | Yes |
| Effort | Tag install | “Two to four weeks (on average)” plus maintenance, per LinkedIn | Faster, with possible partner fees | Configuration in Business Manager |
| Duplication risk | Deduplicated with the API by event ID | Don’t combine with CRM Sync for qualified leads | As direct build | Don’t combine with the API for qualified leads |
LinkedIn’s current partner list includes HubSpot, Salesforce Data 360, Zapier, Make, Segment, Hightouch, Tealium, Google Tag Manager and others (LinkedIn Help). Check the list before you choose, because it changes. If your stages live in a spreadsheet, a workflow tool on that list is often the shortest route.
Zephra sends CRM lead stages to LinkedIn through the Conversions API where configured, using lead status from your CRM or a spreadsheet, and the stage rules below apply just the same. See conversion tracking for how that fits with Google and Meta.
Decision
Which CRM stage should you send to LinkedIn?
Short answer
Send later stages for reporting, but only optimise to a stage that happens quickly and often. In most B2B accounts, opportunities and deals arrive weeks or months after the lead and in small numbers. A faster stage, such as sales-qualified or requirement confirmed, usually gives LinkedIn a usable signal. Validate it against later deals.
The 30/14/5 check turns LinkedIn’s two rules into three yes-or-no questions for each candidate stage. If any answer is no, keep that event for reporting and pick a faster stage for optimisation.
Will each optimised campaign get five or more in 14 days?
Monthly count per campaign × 14 ÷ 30. Below five, consolidate campaigns or choose an earlier stage.
Is the stage recorded the same way every week?
One owner, one definition, a service level for updating it. A stage sales sets “when they get round to it” fails this test.
The cross-platform version of this decision, covering Google and Meta windows as well, is in the signal gap.
Template
The qualified-lead event spec
Write one column per event before anyone builds anything. Most broken integrations fail on a question this table would have raised.
| Field | B2B SaaS | Industrial components | Consulting |
|---|---|---|---|
| Event name | SDR-qualified | Requirement confirmed | Consultation held |
| Trigger (exact field and value) | Lead status = “SQL” | RFQ sheet column “Spec OK” = Yes | Meeting outcome = “Held” |
| Who sets it, and by when | SDR, within 2 working days of the call | Application engineer, within 3 working days | Partner’s assistant, same day |
| Identifier sent | Lead Gen Form response reference plus hashed email | Hashed email plus first and last name | Hashed email plus click ID from the booking page |
| Match risk | Low with the form reference | High: the prefilled email may be a personal address while the sheet holds a work address | Medium |
| Timestamp source | Status change date in the CRM | Date column in the sheet, not the upload date | Meeting date |
| Value (optional) | Static value per stage | None | None |
| Send path (one only) | Conversions API via partner | Conversions API via workflow tool | CRM Sync |
| Conversion type | Qualified lead | Qualified lead | Qualified lead |
| Expected per campaign per month | 18 | 12 | 14 |
| Median days from lead to event | 6 | 8 | 5 |
Worked numbers
Two scenarios: a SaaS company and an industrial supplier
Both run the 30/14/5 check on their own stage data. Replace every figure with yours.
| Stage | Per month | Cost per stage | Median days after lead | 30/14/5 result |
|---|---|---|---|---|
| Lead Gen Form leads | 120 | US$150 | 0 | Fast and frequent, but it teaches LinkedIn to find form-fillers |
| SDR-qualified | 36 | US$500 | 6 | 31 of 36 (86%) within 30 days. 12 per campaign a month is 5.6 in 14 days: passes, just |
| Opportunity | 9 | US$2,000 | 35 | Fails the 30-day check and the volume check |
Decision: send SDR-qualified as the qualified-lead event.
Three campaigns pass only narrowly. Consolidating to two gives 18 qualified leads per campaign a month, or 8.4 in 14 days, which leaves room for a slow fortnight. Keep “opportunity” as a separate reporting conversion with a longer window, and judge budget on it once cohorts mature.
| Stage | Per month | Cost per stage | Median days after lead | 30/14/5 result |
|---|---|---|---|---|
| Lead Gen Form leads | 50 | ₹6,000 | 0 | Frequent, but many are students, suppliers and job seekers |
| Requirement confirmed | 12 | ₹25,000 | 8 | Passes: 12 a month is 5.6 in 14 days, most within 30 days |
| Quote sent | 4 | ₹75,000 | 25 | Fails on volume: about 1.9 in 14 days, and many land after day 30 |
Industrial · IndiaAct
Send “requirement confirmed”
- Why
- It is the earliest stage the sales engineer trusts, and it clears both LinkedIn rules on one campaign.
- Next
- Review quote rate by audience monthly. Don’t add a second campaign until requirement-confirmed volume reaches about 20 a month.
Industrial · IndiaWait
Hold edits for two weeks after switching
- Why
- LinkedIn’s learning phase is two weeks, and at 12 a month each week adds under three events.
- Next
- Compare cost per requirement confirmed in the first full month after learning against the month before.
How long should conversion windows be for long sales cycles?
Short answer
Start with LinkedIn’s default 90-day click and 90-day view window. For lead-type categories, including Qualified Lead and Sales Qualified Lead, you can extend to 180 or 365 days for reporting. A longer window doesn’t help optimisation, because qualified leads shared after 30 days aren’t used to steer delivery.
LinkedIn’s standard conversion windows are 1, 7, 30 or 90 days, and its recommended setting for most conversions is the Campaign Manager default of 90-day click and 90-day view. Some categories, including Qualified Lead, Marketing Qualified Lead, Sales Qualified Lead and Lead, can use a 180- or 365-day window (LinkedIn Help).
Two separate clocks apply, and they are easy to confuse:
- The attribution window decides whether a conversion is credited to a campaign the member interacted with. A 180-day window lets an opportunity created five months after the ad still count in reporting.
- The optimisation cut-off is fixed: qualified lead data shared after 30 days isn’t used for optimisation, whatever the attribution window. Separately, the API rejects events whose timestamp is more than 90 days old (Microsoft Learn).
A practical set-up for a long-cycle B2B account is one qualified-lead rule on the fast stage for optimisation, plus reporting rules for opportunity and closed-won with longer windows where the category allows. Mind the 90-day timestamp limit: send each stage soon after it is set, not months later in a clean-up.
The attribution model matters too. “Each ad set” credits every ad set with an interaction, so totals added across campaigns overcount unique conversions. Read unique counts at campaign level.
Failure modes that quietly break optimisation
- Running the Conversions API and CRM Sync together for qualified leads, against LinkedIn’s recommendation.
- Marking every lead “qualified”. If the qualified rate is 90%, the event carries almost no information beyond “submitted a form”.
- Changing the qualification definition mid-learning. LinkedIn learns a moving target, and the before-and-after comparison becomes meaningless.
- Monthly batch uploads. A lead qualified on day 3 but uploaded on day 32 misses the optimisation cut-off.
- Low match rates. Lead Gen Forms prefill the member’s profile email, which may differ from the work email sales later records. Send the Lead Gen Form response reference where you can, and watch for a Low Match Rate status on the rule.
- Judging inside the two-week learning phase. The right decision in week one is almost always to wait.
- No permission for data sharing. LinkedIn says you must obtain any necessary permissions from your customers before sharing data with it.
How do you check it’s working?
Short answer
Compare lead cohorts from your CRM before and after the switch, once the two-week learning phase has passed. Look at qualified rate and cost per qualified lead first, then check 60 to 90 days later that the newer cohort also produces more opportunities for the same spend. Rising reported conversions alone prove little.
- Record a baseline. For the four weeks before switching, note leads, qualified rate and cost per qualified lead from your CRM, not from Campaign Manager.
- Wait out learning. Ignore the first two weeks after the goal changes, and hold targeting, bid and creative edits.
- Compare cohorts. Leads created in the four weeks after learning against the baseline, on qualified rate and cost per qualified lead. Expect cost per lead to rise; that alone isn’t failure.
- Validate 60–90 days later. Check whether the newer cohort also produces more opportunities per unit of spend. If qualified leads rose and opportunities didn’t, the stage definition doesn’t predict value.
Better reporting is not new demand
Connecting the Conversions API usually lifts reported conversions, because events that were invisible now match. That is recovered measurement. Only the cohort comparison in your CRM tells you whether campaigns started finding better leads.
In Zephra, a recommendation made while a new signal settles is labelled Wait rather than Act; see how it works.
Consent and data handling in India, the UAE and the US
LinkedIn states that Conversions API data is processed on LinkedIn’s servers in the US, that email addresses are hashed with SHA-256, and that data is deleted after 180 days (LinkedIn Help). Procurement and legal teams will ask about all three. A hashed email is still personal data under most privacy laws.
- India. Your lead-form notice and consent should cover sharing lead data with advertising platforms under the DPDP Act 2023.
- UAE. The UAE PDPL applies to most companies; entities in DIFC or ADGM follow those free zones’ own data protection rules.
- US. State privacy laws differ on sharing data for advertising. Partner routes through a CRM are common here; confirm your CRM’s own settings for hashing and consent flags.
This isn’t legal advice. See lead generation compliance for the detail.
What to measure next
- Share of qualified-lead events sent within 30 days of the lead, and within 2 days of the stage being set.
- Qualified leads per optimised campaign per fortnight.
- Match rate status on each conversion rule.
- Opportunities per unit of spend for cohorts before and after the switch.
For how LinkedIn fits alongside Google and Meta on pipeline, see LinkedIn Ads for B2B and Google Ads vs LinkedIn Ads for B2B. To raise the qualified rate at the form itself, see Lead Gen Form lead quality.
Questions, answered
LinkedIn Conversions API FAQ
01Does the Conversions API work with Lead Gen Forms?
Yes. A conversion event can reference a Lead Gen Form response through the API’s lead field, alongside a hashed email. Qualified leads optimisation itself runs on Lead Generation objective campaigns that use Lead Gen Forms.
02Can I send qualified leads from a spreadsheet?
Yes, through a direct build or a partner. LinkedIn’s partner list includes workflow tools such as Zapier and Make that can read a sheet. Record the stage date in its own column so the event timestamp reflects when the lead qualified, not when you uploaded it.
03How long is the learning phase for qualified leads optimisation?
LinkedIn says optimisation requires a two-week learning phase, and recommends sharing five or more qualified leads within those two weeks to speed it up. Avoid other campaign edits during that time.
04What if we get fewer than five qualified leads in two weeks?
Consolidate campaigns so one campaign gets the volume, move to an earlier stage that sales still trusts, or keep the standard lead goal and review qualified rates manually in your CRM until volume grows.
05Is a hashed email personal data?
Under most privacy laws, yes. Hashing protects the address in transit but still identifies a person once matched. Collect consent that covers sharing with advertising platforms, and follow LinkedIn’s requirement to obtain the necessary permissions.
06Can we use CRM Sync and the Conversions API at the same time?
LinkedIn recommends not using both for qualified-lead events, so pick one route for them. The Insight Tag and the Conversions API can run together, because LinkedIn deduplicates events that share an event ID.
07Does this replace pipeline attribution reporting?
No. Campaign Manager shows conversions LinkedIn could match and attribute within its windows. Your CRM remains the record of which leads became pipeline, and that is where channel comparisons belong.
Your account
Not sure which stage LinkedIn can learn from?
The free ad account assessment checks LinkedIn, Google and Meta for wasted spend, structure, constraints and measurement gaps. No CRM needed, no campaign changes.