MQL vs SQL (Marketing Qualified Lead vs Sales Qualified Lead)

GLOSSARY

Definition

An MQL is a lead that marketing deems ready for sales based on engagement signals: they downloaded three whitepapers, attended a webinar, and visited the pricing page. An SQL is a lead that sales has accepted and is actively working. The gap between the two is where most marketing-sales friction lives.

Why it matters

The MQL/SQL framework comes from the SiriusDecisions (now Forrester) demand waterfall model from 2006. It made sense when marketing's job was to fill the top of a funnel and sales caught whatever fell out. The problem is that the handoff criteria are usually negotiated once, written into a service-level agreement, and never revisited. Marketing says they delivered 500 MQLs. Sales says 480 were junk. Both are probably right.

How it works

MQL and SQL are funnel stages, not people. A marketing qualified lead has shown enough interest that marketing hands it to sales. A sales qualified lead has been vetted by a rep and is worth active pursuit. The boundary is usually a score threshold plus an activity rule, for example a score over 50 and a demo request. The definitions vary per company, which is fine as long as both teams agree and the pipeline honors the handoff.

Practical uses

The MQL to SQL conversion rate is the health metric between marketing and sales. If 30% of MQLs become SQLs, marketing passes quality leads. If the rate sits at 5%, the scoring model is too loose or the leads are wrong. Teams also use the split to forecast: pipeline starts with SQL count, not MQL count. Reporting the two stages separately stops the classic game where marketing celebrates volume that sales never accepted.

How to choose

The stage definitions should be written down and reviewed quarterly. Start with the easiest rule that predicts a demo, then tighten by looking at which behaviors precede closed deals. Scoring models with explicit points for job title, company size, and product activity beat black-box models because they can be audited and tuned. If the model cannot be explained to the sales team, it will not survive first contact with them.

The numbers

Handoff benchmarks worth knowing: typical B2B MQL-to-SQL conversion runs 13-20%, and SQL-to-close around 20-30%. If your MQL volume doubled but SQL count stayed flat, marketing qualified more people that sales does not want - a targeting problem no scoring tweak will fix. The fastest diagnostic is a 20-minute audit of ten rejected leads with whoever rejected them.

Common mistakes

The classic failure is marketing optimizing MQL count while sales complains about quality, because the two metrics pull in opposite directions. The fix is to reward the conversion rate, not the volume. The second mistake is letting the definitions drift until the weekly report shows a handoff that no longer means anything. The third is ignoring the time dimension; a lead that scores as an MQL after three months of silence is not the same lead.

What changed with AI

AI lead scoring reads intent signals that rules miss: which pages a contact visits, what the conversation contained, how engagement changed over time. The models can also draft the initial sales outreach and summarize the record before the first call. The caveat mirrors everything else about agents. The model is only as good as the fields feeding it, so stage hygiene matters more once AI is involved, not less.