MQL to SQL: Why Your Conversion Rate Is Lying to You

A high MQL-to-SQL conversion rate is supposed to mean your marketing team is sending sales qualified, high-intent leads that are ready to work. In practice, it usually means one of two things: your definition of an MQL is so loose that almost anything qualifies, or your sales team is accepting leads to avoid a conversation with marketing about quality. Neither is good. Both are common. And both will silently destroy your pipeline while your dashboards look healthy.

The MQL-to-SQL conversion rate is the single most politically loaded metric in B2B revenue operations. Marketing owns the numerator. Sales owns the denominator. And the two teams have structurally opposed incentives when it comes to where the line gets drawn. Marketing wants the conversion rate to be high so their pipeline contribution looks strong. Sales wants flexibility to reject leads that do not meet their criteria — but not if it triggers a conversation with the CMO. The result is a number that everybody reports and nobody trusts.

What a High Conversion Rate Actually Signals

A conversion rate above 60% is a warning sign in most B2B environments, not a success signal. It suggests either that your MQL threshold is extremely low — you are qualifying almost everything — or that your sales team is accepting leads indiscriminately to avoid friction. In either case, the pipeline you are generating from those conversions will underperform.

Think about what happens downstream. If your sales team accepts 70% of MQLs as SQLs, but only 15% of those SQLs ever advance past a first call, your real conversion efficiency is far lower than the headline number suggests. You have filled your pipeline with leads that are not buying-ready. Reps spend time working them. Win rates fall. Sales cycle length extends. The cost per closed deal goes up. And when the CRO asks why the pipeline is not converting, both teams point at each other.

A conversion rate is only meaningful if the leads that convert are ones that can actually close. The rate tells you about the handover process. The downstream metrics — SQL-to-opportunity rate, opportunity win rate, deal velocity from MQL-sourced leads — tell you whether that handover is producing revenue. Most organisations track the former and ignore the latter.

The Blame Dynamic Between Sales and Marketing

The blame cycle is predictable. Marketing generates MQLs. Some convert to SQLs. Sales works the SQLs. Pipeline is thin. Wins are slow. Marketing says sales is not following up on the leads they provide. Sales says marketing is sending garbage. Both are partly right.

What keeps the cycle running is the absence of shared accountability. Marketing is measured on MQL volume and MQL-to-SQL conversion rate. Sales is measured on quota attainment. There is no metric that sits across both teams and measures the quality of the handover at its actual output: revenue from marketing-sourced leads. Until you measure that — and hold both teams accountable to it — the blame cycle continues regardless of how many alignment sessions you run.

The second thing that keeps it running is ambiguous qualification criteria. Most MQL definitions are built on behavioural signals: form fills, email opens, page views, webinar attendance. These signals indicate engagement. They do not indicate fit, intent, or purchasing authority. A marketing manager at a 20-person company who downloaded your whitepaper is engaged. Whether they are a real buyer is a different question that behavioural data cannot answer.

How to Define a True MQL

A true MQL definition has two components: fit and intent. Fit is firmographic — does this company look like your ideal customer profile? Company size, industry, tech stack, growth stage, geography. Intent is behavioural — has this contact demonstrated signals that suggest they are actively evaluating solutions, not just passively consuming content?

Fit Criteria

Pull your last 24 months of closed-won deals. Look at the firmographic profile of the accounts — not just the ones that converted from MQLs, all of them. Identify the attributes that are consistently present in your best customers: headcount range, revenue band, industry vertical, geography, technology stack. These are your positive fit signals. Now look at closed-lost deals and identify the firmographic attributes that are consistently present there. These are your negative fit signals — and they are just as important as the positive ones.

A lead that matches three positive fit criteria and no negative ones should score higher than a lead that matches two positive criteria and one major negative one — even if the second lead has more behavioural engagement. A highly engaged contact at a company that has never bought software like yours, has no budget cycle alignment, and operates in a sector you consistently lose in is not a good MQL. They are an interested non-buyer.

Intent Criteria

Not all behavioural signals are equal. Downloading a thought-leadership piece is a weak signal. Visiting your pricing page twice in a week is a stronger signal. Requesting a demo is a strong signal. Attending a live product webinar — not a recording — is a moderate signal. Build a scoring model that weights these signals proportionally and sets a threshold that has been tested against downstream outcomes.

The key test: take 50 leads that crossed your current MQL threshold in the last six months. Track what happened to each of them. How many became SQLs? Of those, how many advanced past a first meeting? How many generated a proposal? How many closed? If your current threshold is passing leads that consistently stall at the first meeting, your threshold is too low. Raise it until the downstream conversion rates improve.

The Service Level Agreement Problem

Even a well-defined MQL is useless without a functioning SLA. An SLA in the MQL-to-SQL context specifies two things: how quickly sales must respond to an MQL, and what criteria sales will use to accept or reject it. Without both, the handover breaks down in predictable ways.

Response speed matters more than most teams acknowledge. Research consistently shows that lead response rates drop sharply after 5 minutes and fall off a cliff after an hour. A prospect who submitted a demo request on a Tuesday morning and received a first call Thursday afternoon is not the same prospect they were on Tuesday. Their interest has cooled. They may have spoken to a competitor. The context is gone. Treating a 48-hour response as acceptable is treating your marketing spend as mostly wasted.

On the rejection side, the SLA needs to define what happens when sales does not accept an MQL. If the only options are "accept" or "silent rejection," you will never improve your qualification criteria. You need a structured rejection reason — "wrong company size," "wrong title," "no budget signal," "duplicate account," — with data that rolls back to marketing. Without that data, marketing cannot improve targeting. They are flying blind and optimising for engagement metrics that have no downstream accountability.

Building the SLA

An effective MQL-to-SQL SLA covers: maximum response time by lead tier (high-fit leads within 2 hours; standard leads within 24 hours), acceptance criteria that mirror your MQL definition, a required set of rejection codes with definitions, and a joint review cadence — at minimum monthly — where both teams review conversion data and agree on threshold adjustments.

The review cadence is the part that almost never happens. Teams agree on criteria, implement them, and then stop looking at whether they are working. Six months later, the criteria are stale, the market has shifted, and nobody has noticed because nobody is looking at the data that would tell them. Build the review into the operating rhythm or the SLA degrades to irrelevance.

Building a Conversion Rate That Means Something

The MQL-to-SQL rate as a standalone metric is nearly worthless. The metrics that make it meaningful are: SQL-to-opportunity rate (what share of accepted leads actually progress to a real deal), opportunity-to-close rate from MQL-sourced pipeline, average sales cycle from MQL to close, and revenue attributed to MQL-sourced leads as a percentage of total revenue.

When you measure all of these together, the conversion rate becomes interpretable. A 55% MQL-to-SQL rate looks different depending on whether the downstream pipeline generates $2M a quarter or $400K. If it generates $400K, your conversion rate is inflated — you are accepting leads that do not close. If it generates $2M, your conversion process is working regardless of whether the top-of-funnel number matches the industry benchmark.

Stop optimising for the handover rate. Start optimising for the revenue output of the leads that cross the line. That reframe changes the incentives for both teams. Marketing has to care about whether their leads close, not just whether sales accepts them. Sales has to care about whether their rejection criteria are valid, not just whether they can push back on leads they do not want to work.

A 70% MQL-to-SQL conversion rate built on a weak definition is not a funnel. It is a pipeline full of noise with a good-looking headline number attached to it.

The fix is unglamorous. It requires sitting in a room with marketing and sales leadership, pulling actual data on what happened to the last 100 MQLs, and agreeing on what the numbers are telling you. That conversation is uncomfortable. It will surface bad news on both sides. But it is the only conversation that moves the metric from a political scoreboard to an operational tool.