LinkedIn Wants Your MQL Definition. Do You Have One?

LinkedIn’s Marketing API version 202608 added two granular conversion types to the Conversions API — MARKETING_QUALIFIED_LEAD and SALES_QUALIFIED_LEAD — alongside the existing generic QUALIFIED_LEAD, which has fed the qualified-leads optimization goal since that goal shipped in February 2026. Events streamed on rules with the new types feed the same deeper-funnel campaign optimization.

That is the entire news item, and it is deliberately small. The reason it deserves a full framework rather than a changelog bullet is what the schema says next: both new types describe a lead that “has been filtered by the advertiser.” LinkedIn ingests your label. It does not compute it. Which means the real work — the work that decides whether these enum values ever carry signal — happens in your CRM, in a field most teams have never formally defined.

This guide covers what actually changed in v202608, the six-month chronology that most coverage will miss, where the MQL/SQL state has to live in HubSpot and Salesforce, a definition worksheet for the marketing-to-sales handoff, and the plumbing that moves the label from a CRM field to a LinkedIn conversion event. Everything here is sourced from LinkedIn’s own API documentation and the two CRM vendors’ own field references, each linked inline where it is used.

The change entry itself is one sentence. Starting with the 202608 version, LinkedIn’s /conversions endpoint supports MARKETING_QUALIFIED_LEAD and SALES_QUALIFIED_LEAD as granular conversion types in addition to the existing QUALIFIED_LEAD. Conversion events streamed on rules with the new types are used in the qualified-leads optimization goal for deeper-funnel campaign optimization, the same way the generic type already was.

The detail that matters operationally sits on the schema page, not the changelog: the granular types are “supported only for API versions 202608 and later.” That is a hard gate. LinkedIn’s versioning policy keeps each version supported for a minimum of one year before sunset — and the clock is genuinely running: v202508 shipped in August 2025 and was sunset on August 17, 2026, the same day v202608 landed. If your integration or your ads partner pins anything below 202608, these enum values simply do not exist for you. Check that before any CRM work.

The same v202608 release also took the Matched Audiences API to general availability, added account-level tracking parameters to the Dynamic UTM API, and imposed a 1,000-segment cap per sponsored account on DMP segments — a cleanup problem we cover separately in our guide to the DMP segment cap. Alongside the release, LinkedIn published a new Conversions API use-cases document covering qualified-leads optimization, retargeting audiences, predictive-audience source data, and conversion lift testing. Read together, the direction is clear: qualified-lead signal is a program LinkedIn is building out, not a one-off field.

Most coverage will treat this as an isolated changelog bullet. The actual product story spans two releases. In February 2026, API version 202602 introduced MAX_QUALIFIED_LEAD as a supported optimization target for the /adCampaigns and /adBudgetPricing endpoints under the Lead Generation objective — requiring either a conversion rule of type QUALIFIED_LEAD or CRM data shared through Business Manager. That entry is where advertiser-defined lead quality, rather than raw form fills, becomes a documented optimization target for Lead Generation campaigns.

August’s change refines that existing mechanism. Six months of the optimization goal running on a single undifferentiated “qualified” signal evidently made the case for granularity: a marketing-decided MQL and a sales-decided SQL are different events, decided by different teams, at different funnel depths. Splitting the enum lets a campaign optimize toward the stage that actually matters for its objective — and lets reporting distinguish the two instead of collapsing them into one bucket.

The build-out extends sideways, too. LinkedIn’s Company Intelligence API added a paidQualifiedLeads field in version 202603 — a count of Conversions-API-sourced qualified leads attributed to a target company, aimed at account-based marketing prioritization. Qualified-lead signal is being wired into campaign optimization, reporting metrics, and ABM surfaces at once. That is worth registering before you decide the CRM work is optional: every quarter, more of LinkedIn’s B2B surface area assumes you can tell it which leads were actually good.

Read the Conversions API schema’s own enum definitions closely, because they are the most important sentences in this entire release. Both new types are defined identically except for one word: the user filled out a lead generation form, and has been filtered by the advertiser to be a marketing — or sales — qualified lead. “Filtered by the advertiser” is doing all the work. LinkedIn is not scoring your leads, inferring qualification from engagement, or applying some platform-side quality model. It accepts whatever you send and optimizes toward it.

That cuts both ways. It is empowering — your definition of quality, not a platform’s guess, steers delivery. And it is unforgiving — if your “MQL” field is set inconsistently by three different automations and hand-edited by whoever noticed the lead first, that noise is now your optimization signal. An ad delivery system trained on a label is only as good as the process that produces the label. Garbage definitions in, garbage optimization out, with budget attached.

This is why the durable part of this story is CRM work, not LinkedIn work. The question “what, precisely, makes a lead marketing-qualified here?” has an answer at most companies only in the folk-knowledge sense — marketing has one version, sales has another, and the CRM field that supposedly records it has no enforced rules about who sets it or when. LinkedIn just gave that ambiguity a direct cost.

The two dominant CRMs take opposite defaults, and the contrast is instructive. HubSpot ships an opinionated, ordered lifecycle-stage property with eight default values in sequence: Subscriber, Lead, Marketing Qualified Lead, Sales Qualified Lead, Opportunity, Customer, Evangelist, Other. The MQL/SQL boundary is pre-baked into the product. Salesforce ships the opposite: its standard Lead object’s Lead Status field is a picklist “set by an administrator” — example values Open, Contacted, Qualified — with no native MQL or SQL concept anywhere in the standard object. If a Salesforce org has an MQL stage, an admin built it.

Two details from the vendors’ own documentation matter for anyone wiring these fields to LinkedIn. First, HubSpot’s lifecycle stage is a single property holding one value at a time, and its default behavior only allows automatic progression forward through the stage list — moving a record backward requires manually clearing the property first. One field, one value, forward-only: that is the shape a qualification gate should have, enforced by the product. Second, Salesforce deliberately separates Lead Status (where the lead is in the process) from Rating (Hot, Warm, Cold — how good the lead looks). Conflating “how good” with “how far along” is one of the most common CRM-hygiene mistakes, and it is exactly the ambiguity that corrupts an MQL signal.

One scope note on the Salesforce row: the absence claim is about the out-of-the-box standard Lead object, per Salesforce’s own Lead Fields reference. Plenty of real orgs have added custom MQL/SQL picklist values or fields — which is precisely the point. Where HubSpot hands you a definition-shaped container, Salesforce hands you a blank, and either way the definition itself is still yours to write. The vendors ship the field. Nobody ships the meaning.

An MQL/SQL definition is not a paragraph in a strategy deck. It is six operational answers, each living in a specific place in your CRM, each with a specific failure mode when left undefined. Work through them in order — the first three define the label, the last three make it auditable. If you already run a lead-scoring workflow that feeds a clean MQL threshold, question one is done; most teams find the gaps start at question two.

Questions two and three are where most definitions die, because they are organizational rather than technical. The mechanics of how a lead gets routed once it is qualified and the SLA that actually enforces the marketing-to-sales handoff are established disciplines — the change is that LinkedIn’s optimization goal now consumes their output directly. A handoff that exists only as goodwill between two team leads is now a data quality problem in your ad account.

Two paths carry a qualification signal into LinkedIn’s qualified-leads optimization goal: the Conversions API — a direct build or a third-party partner integration streaming events — and CRM Sync, where LinkedIn creates conversions itself from CRM data connected through Business Manager. LinkedIn’s guidance is to use one source per ad account, not both, to avoid double-counting the same qualification. CRM Sync trades integration work for latency — its processing is measured in days, not minutes — while a CAPI event fires as fast as your trigger does.

The API imposes real timing constraints. The event timestamp, conversionHappenedAt, must fall within the past 90 days at submission — you cannot backfill last year’s MQLs to bootstrap the system. In practice the useful window is tighter still: LinkedIn’s help guidance ties optimization eligibility to recently shared data and recommends seeding the learning phase with a steady flow of qualified leads in the first weeks. The operating conclusion is the same either way — the handoff from CRM field to LinkedIn event needs to be near-real-time or batched frequently, not a quarterly export.

Match quality is the other half. A streamed event carries user identifiers — hashed email (SHA256_EMAIL), LinkedIn’s first-party ads tracking UUID (LINKEDIN_FIRST_PARTY_ADS_TRACKING_UUID), and several others — plus optional userInfo fields (first and last name together, company, title, country) that improve match rates. For leads that came through LinkedIn Lead Gen Forms there is a stronger option: a leadGenFormResponseUrn ties the conversion straight back to the specific form submission. And for agencies, the conversion rule’s ownershipType supports OWNED and SHARED — a SHARED rule under one Business Manager serves multiple client ad accounts without rebuilding the plumbing per account.

One attribution nuance worth hedging precisely: LinkedIn supports attribution windows up to 365 days for a short list of conversion types, and the generic QUALIFIED_LEAD is on that list — an explicit accommodation of long B2B sales cycles. The schema does not say whether the two new granular types inherit that long window, so verify against the current documentation before you assume it. Defaults are 30 days post-click and 7 days view-through.

Where to start depends on which CRM you run and how your LinkedIn integration is pinned. Four common scenarios:

If the qualification decision itself is the bottleneck — leads sitting unscored, or scored by gut feel — that is an automation problem before it is a LinkedIn problem. We have covered building an AI lead-scoring agent wired into the CRM for exactly this step; if you automate it, note that EU AI Act risk-tiering touches automated lead scoring for teams operating in Europe. And if the whole field-and-handoff layer needs building from scratch, that is the core of our CRM automation service — the definition workshop, the field architecture, and the event plumbing as one engagement.

The forward-looking read: advertiser-defined outcome signals are becoming the optimization currency of B2B ad platforms, and LinkedIn is building its version out methodically — an optimization goal built on the existing generic type in February, granular MQL/SQL splits in August, and qualified-lead counts surfacing in ABM reporting via the Company Intelligence API. Teams that treat the CRM definition layer as durable infrastructure will plug into each new surface in days. Teams that treat each platform ask as a one-off integration will rebuild the same plumbing every quarter — a pattern we see constantly in paid media engagements.