LinkedIn qualified leads optimisation. Which CRM stage should you send back?

KEY TAKEAWAYS

Count usable events, not CRM totals. Only qualified leads shared within 30 days feed the optimisation.
Five per two weeks is the floor. At that average, a single fortnight usually shows between 1 and 10 events.
Optimise on the earliest strict stage that clears the floor. Report on SQL with a 180-day window.
The qualified leads goal bids automatically. Judge it on cost per SQL by lead cohort, not on cost per lead.

LinkedIn's qualified leads goal lets a lead generation campaign optimise towards the leads your CRM marks as good, not towards every form fill. Since August 2026 you can tell LinkedIn whether an event is a marketing qualified lead or a sales qualified lead. That raises a question the setup guides skip: which stage should drive delivery? The answer depends on two numbers from your own CRM, and this article shows how to work them out.

What changed in August and September 2026?

LinkedIn's API changelog lists two releases that matter here.

The August 2026 release added two conversion types, marketing qualified lead and sales qualified lead, next to the generic qualified lead type. The changelog says events sent on the new types feed the qualified leads optimisation goal in the same way.

The September 2026 release added a 180-day attribution window for lead conversion types. A 365-day option already existed. The same release lets you send hashed first and last names for matching.

The goal itself is older. Social Media Today reported its announcement on April 22, 2025, and the changelog shows it reached the API as an optimisation target in February 2026. Integrations pinned to an API version before August 2026 cannot send the two new types.

One inconsistency is worth knowing. LinkedIn's use-case page still says the conversion rule must use the generic qualified lead type to be eligible. The changelog says the new types count too. Check which conversions your ad set lets you select before you rebuild anything.

What does LinkedIn need from you?

The developer documentation sets three conditions:

Qualified lead data must be shared within 30 days to be included in ad set optimisation.
Optimisation needs a two-week learning phase.
LinkedIn recommends five or more qualified leads within two weeks to speed that phase up.

LinkedIn's own statement at launch, quoted by Social Media Today, adds that qualified leads can come from your other active lead generation campaigns.

The documentation does not say what the 30 days run from. SMK and Zephra both read it as 30 days from the lead being created. That is the stricter reading, so plan for it.

Three more limits come from the Conversions API schema and FAQ. An event's timestamp cannot be more than 90 days old when you send it. Data can take up to 72 hours to appear in reports. Only events LinkedIn can match to a member can be attributed, so send the lead ID from the form alongside a hashed email.

There is also a cost. The API describes this target as maximising qualified leads and spending the daily budget without an advertiser bid. Clix Marketing noted in February 2025 that choosing the goal moves you to automated bidding, while the standard leads goal offers maximum delivery, cost cap or manual bids. You give up bid control to get the signal.

Is five qualified leads every two weeks enough?

Sources disagree, and none of them publishes evidence.

LinkedIn's figure works out at about 11 a month by our calculation. Kiin's September 2026 Campaign Manager guide says that below about 50 qualified events a month the algorithm has "nothing to learn from". OptimizeLinkedInAds sets the bar at 100 or more a month and calls fewer than 30 not viable.

Real account volumes make the higher bars hard to meet. Kiin Labs studied 268 advertiser accounts and 20,277 demo leads between September 2025 and September 2026. The median account produced 6.6 demo leads per active month. The 75th percentile produced 17 and the 90th produced 44. Those are leads, not qualified leads.

Our calculations from those figures:

Account in Kiin's data	Demo leads a month	Share that must qualify to reach 5 per two weeks
Median	6.6	Not reachable, even at 100%
75th percentile	17	64%
90th percentile	44	25%

So a typical account running demo offers alone cannot meet LinkedIn's own floor, and a 50-a-month bar sits above the 90th percentile account even if every lead qualified. Content offers add volume, but Kiin's forms study reports that only about 3% of content leads are in-market when they download.

The floor is also noisy. Events arrive as random counts. If your true average is 5 per two weeks, a single fortnight will show between 1 and 10 at least 95% of the time by our calculation, and 2 or fewer about one time in eight. An average of 25 narrows the range to 16 to 35. This is the same sample logic we used to size creative tests.

What the evidence supports: five is the minimum for switching the goal on, not a promise that it will work. The agency thresholds are reasonable targets for a stable signal. They are not rules, and most accounts will sit between the two.

Why the best signal usually arrives too late

Several guides say to optimise on the deepest stage you can. GrowthSpree recommends the SQL or opportunity event. OptimizeLinkedInAds recommends weighting towards closed-won.

The 30-day limit works against that advice. Dreamdata's 2026 LinkedIn Ads benchmark, published in March 2026 from more than 3.5 million B2B customer journeys, puts the average journey at 272 days. PPC Land's report of the same study gives 92 days from MQL to SQL and 52 days from SQL to closed-won. Dreamdata's 2025 edition also described the MQL to SQL stage as taking over three months.

Those are averages across Dreamdata customers, and PPC Land notes that method changes make its year-on-year comparison imperfect. Your own lag will differ. But if your SQLs typically appear two or three months after the form fill, most of them land outside the 30 days and do not train the ad set.

Deep stages are still worth sending. They are the right events for reporting, and the 180-day and 365-day windows exist for that. The schema's defaults are 30 days after a click and 7 days after a view, which is too short for an SQL. This is the same split between what you optimise to and what you report that we described for Meta's attribution settings.

The usable-events formula

For each CRM stage, work out one number:

Usable events per two weeks = leads per month × share reaching the stage × share reaching it within 30 days of lead creation × 14 ÷ 30.4

Count leads from all active LinkedIn lead generation campaigns. If your match rate is poor, multiply by that too.

The third term is the one teams do not have to hand. Get it from a cohort. Take the leads created in one month at least four months ago, find those that reached the stage, and count how many did so within 30 days.

Which stage to send: a decision table

CRM stage	Typical problem	Use it as the optimisation event when	Otherwise
Lead (form fill)	Tells LinkedIn nothing about quality	You have fewer than 5 usable qualified events per two weeks	This is the standard leads goal. Stay on it
MQL	Loose definitions. A score that rises just because the form was filled sends LinkedIn back to form fillers	The definition includes fit criteria a form fill does not guarantee, such as company size, role and a work email, and it clears 5 usable events	Tighten the definition first
SQL	Strong signal, low volume, slow	Your cohort shows 5 or more usable SQLs per two weeks. This is most likely with demo requests that sales qualifies within days	Send it as its own conversion with a 180-day window and report on it
Opportunity or closed-won	Arrives months later	Rarely	Report only, with a 180-day or 365-day window

Keep each stage as a separate conversion rule. LinkedIn's documentation says most conversion types count only the first instance of a repeated conversion within the lookback window. A lead that becomes an MQL and then an SQL on the same rule would be counted once.

A worked example with invented numbers

This company is hypothetical. It sells B2B software and spends $9,000 a month on LinkedIn lead generation campaigns. It gets 45 leads a month, a cost per lead of $200.

Its CRM shows that 40% of leads become MQLs under a definition that requires a work email, a target job function and a company with more than 50 staff. That is 18 MQLs a month. Thirty percent of MQLs become SQLs, which is 5.4 a month. Cost per MQL is $500 and cost per SQL is about $1,667.

The cohort check shows 95% of MQLs are marked within 30 days of the lead, but only 35% of SQLs are.

Usable MQLs: 18 × 0.95 × 14 ÷ 30.4 = 7.9 per two weeks. That clears the floor.
Usable SQLs: 5.4 × 0.35 × 14 ÷ 30.4 = 0.9 per two weeks. That does not.

The company sends both stages on separate rules. It selects the MQL conversion for the qualified leads goal and sets the SQL rule to a 180-day window for reporting. It expects cost per lead to move, because the system is no longer buying the cheapest form fill. It judges the change on cost per SQL for leads created before and after the switch, once the later cohort has had the same time to mature.

A six-step method

Write the definition. Agree one MQL and one SQL definition with sales. LinkedIn does not compute qualification. It trusts your label.
Run the cohort check. Calculate usable events per two weeks for each stage.
Pick the stage. Choose the deepest stage with 5 or more usable events. If none qualifies, stay on the leads goal.
Build the rules. One conversion per stage. Set long windows on SQL and beyond. Send events as they happen, not in a monthly batch.
Leave it alone. Allow the two-week learning phase, then at least one more full lead-to-SQL cycle before judging.
Read cohorts. Compare cost per SQL by the month the lead was created. A before-and-after comparison is not an experiment, so repeat the read.

What to do below the floor

If no stage reaches five usable events, the goal has too little to learn from. Three options remain. Add qualifying questions to the form, so the lead event itself carries more meaning. Use your qualified lead conversions or forms as the source for a predictive audience, which LinkedIn's API accepts. And keep sending SQL events for reporting, so you know your real cost per SQL when volume grows. The arithmetic of lead quality against lead volume is the same one behind our piece on gating a guide.

Where the evidence stops

LinkedIn's headline claim, a decrease of up to 39% in cost per qualified lead, has no published sample or method. AdExchanger carried the same figure in February 2025 as an early result. The independent view is more modest. Laura Schiele of Jordan Digital Marketing wrote in Search Engine Land in March 2026 that the feature raised the share of qualified leads in her team's tests, and in MarTech that it is not yet as effective as the Meta and Google equivalents. Neither piece gives numbers. Treat the goal as a test with a defined read, not a setting to switch on and trust.

Ad platforms are not the only systems that now decide which brands a buyer sees. If you want to know how we measure brand presence in AI answers, the waitlist is open.

Questions

What is LinkedIn qualified leads optimization?

It is an optimisation goal inside the lead generation objective. You send LinkedIn the leads your CRM marks as qualified through the Conversions API, and delivery shifts towards members who resemble them. LinkedIn announced it in April 2025. The API has supported it as a target since February 2026.

How many qualified leads does LinkedIn need?

LinkedIn's developer documentation recommends five or more qualified leads within two weeks, which is about 11 a month by our calculation. Agencies quote 50 or 100 a month without published evidence. Five is a floor for switching the goal on. It is not proof of a stable signal.

Should I optimise LinkedIn campaigns on MQLs or SQLs?

Use whichever stage clears five usable events per two weeks inside the 30-day sharing limit. Dreamdata's 2026 benchmark puts the average gap from MQL to SQL at 92 days, so most SQLs arrive too late to count. Optimise on a strict MQL and report on SQLs.

What changed in LinkedIn's Conversions API in August and September 2026?

The August 2026 release added separate marketing qualified lead and sales qualified lead conversion types alongside the generic qualified lead type. The September 2026 release added a 180-day attribution window for lead conversion types and accepted hashed first and last names for matching.

Sources

Microsoft Learn, LinkedIn Recent Marketing API Changes learn.microsoft.com
Microsoft Learn, LinkedIn Conversions API Use Cases learn.microsoft.com
Microsoft Learn, LinkedIn Conversions API Schema learn.microsoft.com
Microsoft Learn, LinkedIn Conversions API learn.microsoft.com
Microsoft Learn, LinkedIn Conversions FAQ learn.microsoft.com
Microsoft Learn, Create and Manage LinkedIn Campaigns learn.microsoft.com
Microsoft Learn, LinkedIn Predictive Audiences learn.microsoft.com
Kiin Labs, The Cost of a Demo on LinkedIn, 2026 Benchmarks kiin.co
Kiin, LinkedIn Lead Gen Forms 2026 kiin.co
Kiin, LinkedIn Campaign Manager Tutorial 2026 kiin.co
PR Newswire, Dreamdata LinkedIn Ads Benchmarks Report 2026 prnewswire.com
PPC Land, LinkedIn ads hit 121% ROAS as B2B buyer journeys stretch to 272 days ppc.land
Dreamdata, The 2025 LinkedIn Ads Benchmarks Report highlights dreamdata.io
Social Media Today, LinkedIn Launches Qualified Leads Optimization for Custom Ad Targeting socialmediatoday.com
AdExchanger, LinkedIn Launches Its Own Conversion API adexchanger.com
Clix Marketing, New LinkedIn Optimization Goal for Qualified Leads clixmarketing.com
Search Engine Land, 5 B2B LinkedIn Ads tests to run in 2026 searchengineland.com
MarTech, 5 LinkedIn ad tactics for B2B marketers martech.org
Digital Applied, LinkedIn Wants Your MQL Definition. Do You Have One? digitalapplied.com
OptimizeLinkedInAds, LinkedIn Qualified Lead Optimization (QLA) optimizelinkedinads.com
OptimizeLinkedInAds, LinkedIn HubSpot Offline Conversions optimizelinkedinads.com
GrowthSpree, LinkedIn Ads QLA for B2B SaaS, CAPI Setup Guide growthspreeofficial.com
Zephra AI, LinkedIn Conversions API for B2B Qualified Leads zephraai.com
SMK, LinkedIn Boosts Lead Quality Targeting smk.co
HubSpot Knowledge Base, Create and sync ad conversion events with your LinkedIn Ads account knowledge.hubspot.com

Esh

Founder, Ethos Engine. Builds the Digital Authority Score and runs the placement marketplace. esh@ethosengine.io

NEXT STEP

See where your brand stands.

Your Digital Authority Score, who AI recommends for your target questions, and the placements that close each gap. Your first report is free.

Get your free report →