What's the difference between an MQL, SQL, and SAL?

An MQL, a SAL, and an SQL are three different handoffs, not three grades of the same lead. A marketing qualified lead (MQL) is a lead marketing believes is worth a sales conversation. A sales accepted lead (SAL) is that same lead after a salesperson has looked at it and agreed to work it. A sales qualified lead (SQL) is that lead after the salesperson has actually spoken to the buyer and confirmed there is a real, qualified opportunity. The order is MQL → SAL → SQL, and each arrow is a decision someone made, not a score threshold the lead crossed on its own.

That distinction is the whole point of having three terms instead of one. If you only track MQLs, you learn how much marketing produced and nothing about whether sales agreed. If you only track SQLs, you learn what survived but not where the rest died.

Where the three stages come from — and why SAL looks backwards

The sequence trips people up because "accepted" sounds weaker than "qualified," so most people assume SQL comes first. It doesn't, and the reason is historical. The model comes from SiriusDecisions' Demand Waterfall, the B2B funnel framework that became the industry default before SiriusDecisions was acquired by Forrester. In that model the stages describe who is making the call at each step, not how good the lead is:

• MQL — marketing's judgement. Marketing says: this one meets our bar, take it.

• SAL — sales' judgement about marketing's judgement. Sales says: agreed, this is worth my time, I'll work it.

• SQL — sales' judgement about the buyer. Sales says: I've talked to them, there's a real opportunity here.

Read that way the order is obvious. Acceptance has to happen before qualification, because nobody can qualify a lead they never picked up.

What each stage is actually for

MQL answers: is marketing producing the right kind of demand?

An MQL should be defined against a written ICP, not against activity. A lead that downloaded three things but works at a 12-person company you'd never sell to is not an MQL. This is where most funnels go wrong first, and it is why defining your ICP as checkable rules has to come before you define an MQL at all. If "qualified" means "filled in a form," the stage is measuring form fills and calling them demand.

SAL answers: do marketing and sales agree on what qualified means?

SAL is the stage almost everyone skips, and skipping it is expensive, because it is the only place the disagreement becomes visible. The gap between MQLs created and MQLs accepted is the single most honest number in a B2B funnel. A high rejection rate is not a sales attitude problem — it means two teams are working from two different definitions, and the MQL definition is the thing to fix. Without a SAL stage that argument happens in meetings instead of in the data.

SQL answers: is this a real opportunity?

An SQL has been through a human conversation. Budget, authority, need and timing have been probed to whatever depth your sale requires. This is the stage your forecast should be built from, and it is the input to building a revenue plan from at-bats and close rates rather than from hope.

How to define the stages so your reporting actually works

Four rules, in order. Each one is cheap and each one is routinely skipped.

1. Write one ICP definition and make every stage inherit it. All three stages should mean "fits the ICP, and has progressed this far." If marketing scores against firmographics and sales qualifies on gut feel, the stages are not measuring the same population and the conversion rates between them are meaningless.

2. Make the MQL verdict a live data point, not a monthly cleanup job. Score each inbound lead against the ICP the moment it arrives and route on the verdict. Auto-qualifying inbound leads against your ICP in HubSpot is the mechanism; the point is that "qualified" should be a field the CRM maintains rather than an opinion someone types in later.

3. Require an explicit accept-or-reject on every MQL, with a reason code. No silent stalling. A lead sits in "accepted" or "rejected — wrong size / wrong role / no need / bad timing," and the reason codes are the feedback loop that tells marketing what to stop producing.

4. Close the loop back to the channel that produced each stage. Stage counts on their own tell you where leads died, not which spend caused it. Closed-loop reporting joins each stage back to its source, which is what lets you compare channels on cost per qualified lead rather than cost per lead. A channel that produces cheap MQLs and no SALs is not a cheap channel.

When the MQL model stops being the right one

There is a real argument, running right now, that the individual-lead handoff is the wrong unit for B2B altogether — because nobody buys B2B software alone. Forrester has been pushing B2B organisations to replace MQL handoffs with buying groups: tracking the set of people at an account who are collectively evaluating, rather than routing whichever one of them filled in a form.

The results Forrester has published are worth taking seriously. In its account of Palo Alto Networks' move to buying groups, opportunities with multiple people attached were eight times more likely to advance in the pipeline than opportunities with a single contact, and focusing on buying groups produced a 17% higher closed-won rate in the months that followed.

Two honest conclusions sit next to each other here.

First: the stages are still worth defining. If your current funnel has no SAL stage and an MQL definition that means "downloaded something," you do not have an MQL problem you can solve by adopting buying groups. You have a definitions problem, and it will follow you into any model you move to. The discipline of writing the stages down, agreeing them with sales, and closing the loop to the channel is the work either way. 

Second: if your deals genuinely involve five or six people, a single-lead handoff is throwing away most of what you know. Once the stages above are real and your reporting is honest, moving the unit of measurement from the person to the account is the next step, not a replacement for it.

The failure mode to avoid is adopting the newer model as a way to avoid fixing the older one.

Want to see where your funnel definitions are actually leaking?

Get your free growth audit →

Frequently asked questions

Does SAL come before or after SQL?

SAL comes first. The order is MQL → SAL → SQL. It reads backwards because "accepted" sounds weaker than "qualified," but the stages describe who is making the decision, not how strong the lead is: marketing qualifies it, sales accepts it, then sales qualifies it after talking to the buyer.

What's the difference between an MQL and an SQL?

An MQL is marketing's judgement that a lead is worth a sales conversation, usually based on ICP fit plus behaviour. An SQL is sales' judgement, after an actual conversation with the buyer, that there is a real and qualified opportunity. One is a prediction; the other is a confirmation.

Do we really need a SAL stage?

If marketing and sales ever disagree about lead quality, yes. SAL is the only stage that records whether sales accepted what marketing sent, and the gap between MQLs created and MQLs accepted is the clearest signal that the two teams are working from different definitions. Without it, that disagreement lives in meetings instead of in your reporting.

Should an MQL be defined by lead score or by ICP fit?

By ICP fit first, with behaviour as a secondary signal. A high score from someone at a company you would never sell to is not a qualified lead. Score ICP fit and engagement separately so you can tell "right company, not ready" apart from "wrong company, very active" — they need completely different follow-up.

Are MQLs being replaced by buying groups?

Some organisations are moving that way, and Forrester has published client results supporting it — in its Palo Alto Networks account, opportunities with multiple contacts attached were eight times more likely to advance and buying-group focus produced a 17% higher closed-won rate. But a buying-group model does not fix a vague MQL definition or a missing handoff agreement. Get the stage definitions and closed-loop reporting right first; the unit of measurement is the next decision after that, not instead of it.