MQLs Do Not Work: Qualified Pipeline Instead of Lead Theatre

MQLs are the most discussed and most misunderstood metric in B2B marketing. Marketing celebrates MQL volume, sales watches close rate, neither talks to the other. In the end marketing explains the sales forecast and sales explains the bad leads. Same game for years.

The honest diagnosis: MQLs are a reporting construct that does not drive a business decision in most companies. Forrester has been recommending for years to replace MQLs with pipeline quality metrics, and most dashboards still hang on them.

What you take away:

• Why MQLs drive conflict instead of qualification.
• What a qualified pipeline definition delivers that an MQL cannot.
• How to align marketing and sales on a single definition.
• Which metrics carry the business decision instead of MQL volume.

The thesis

MQLs are broken because they measure activity, not pipeline quality. A GTM team needs a shared pipeline definition that marketing and sales both sign. The definition is the leadership task, the reporting is just its shadow.

What is an MQL and why does it not work?

An MQL is a Marketing Qualified Lead, a contact classified by marketing criteria (score, activity, firmographics) as ready for sales. It does not work because the criteria are usually defined by marketing alone and only loosely linked to what sales actually needs.

Concretely: a whitepaper download becomes an MQL, a cold call request does not. From a sales perspective the first is mostly noise, the second is often gold. As long as the definition is not shared, the game stays asymmetric.

Which problems are pipeline problems, not MQL problems?

Three problems show up everywhere. Marketing produces volume, sales does not work it (conversion). Sales ignores MQLs (trust). Both report different pipeline numbers in forecast (data). All three are pipeline problems, not lead problems.

How the handover itself gets clean is covered in my piece on the marketing-to-sales handover. This article is about the layer below: the definition every handover builds on.

First qualify, then accept

Jacco van der Kooij of Winning by Design puts the right sequence into a simple formula: first qualify, then accept. A lead becomes sales qualified first, sales accepted second. Sounds trivial. Most companies run it backwards: marketing hands over, sales formally accepts and qualifies afterwards. The MQL construct cements exactly that wrong order.

Van der Kooij goes further. Qualification by BANT, short for budget, authority, need and timeline, dates from the era of IBM mainframes. For high-velocity sales, meaning SaaS selling with short cycles and high cadence, it no longer works, he argues. His point: a SaaS prospect rarely has a budget problem, they have a time problem. If you run through a checklist instead of having a conversation, you qualify right past the buying process.

For DACH there is a detail van der Kooij himself flags: European contacts find the assertive US follow-up after every whitepaper download intrusive. Not every download is a lead. Ignore that and you burn exactly the accounts that would have bought later.

What does a qualified pipeline definition look like?

A qualified pipeline definition describes when an account becomes an opportunity, not when a contact becomes an MQL. Three parts: ICP fit (industry, size, region), buyer signal (trigger, need, budget), sales touchpoint (a human spoke to them, a meeting is booked). Only then is it pipeline.

This definition is signed by both sides. Marketing provides ICP fit and buyer signal. Sales provides the touchpoint. The SLA between them is response time, not volume.

Before you define the ICP fit, you need a clean ICP, meaning a sharp profile of your ideal customer. How to derive it from real customer data instead of wishful thinking is in my piece on lead research and ICP.

Must an MQL include the decision maker? Mark Roberge says no

Mark Roberge built HubSpot's sales organisation from zero to IPO as its first sales chief. In The Science of Scaling he makes a point many pipeline purists dislike: an MQL definition should not require contact with a decision maker. His reasoning: when a junior engages with your content, a decision maker's strategic direction usually sits behind it. The junior researches because the boss set the topic. Navigating up from junior to decision maker is sales craft, not an exclusion criterion for marketing.

That does not contradict my third component, it sharpens it. The sales touchpoint has to happen. But it may start with the junior.

More interesting is how Roberge runs the numbers on the interface. His example: a 6 million dollar ARR goal, a 50,000 dollar average deal, a 20 percent close rate on qualified leads. That makes 120 deals, so roughly 600 MQLs per year. Plus the SLA: sales contacts 100 percent of these leads within 24 hours, with at least 5 attempts over two weeks. A contract with numbers on both sides, not dashboard theatre.

For context: Roberge's numbers are US SaaS conventions, not laws of nature. The logic still travels. Define volume and quality together and response time becomes measurable. That is exactly what most MQL setups lack.

How do you get marketing and sales to align?

You get them aligned by forcing both to write in the same document. One page, three blocks (ICP, signal, touchpoint), three real examples per block from actual closed deals. Teams arguing from real data argue less. Teams arguing from gut feel always argue.

The approach is not new but rarely practiced. Gartner data on the B2B buying journey shows buying groups of 6 to 10 people on average. In that reality a "lead" is a fiction. An "account in a buying process" is the only useful unit.

Keep the definition alive, too. Test it every quarter against three newly won and three lost deals. Still fits, good. Does not fit, adjust and sign again. That is the same discipline I describe under GTM engineering: definitions are code, and code needs maintenance.

Which metrics actually steer the business?

Four metrics steer: pipeline coverage (pipeline to quota), pipeline velocity (days from first touch to close), conversion rate per stage, and source mix (where does closed pipeline come from). MQL volume is not on the list.

When marketing steers by these metrics, behaviour changes. Instead of more whitepaper downloads, marketing hunts buyer signals. Instead of more emails to MQLs, sales sends fewer emails to the right accounts. The reporting gets smaller and more honest.

What do you measure instead of MQLs?

Replies, meetings, opportunities. Those are the three counters that tell you whether your market responds. Add the four steering metrics from above: coverage, velocity, conversion per stage, source mix. In short: you measure revenue movement, not activity. Everything else is proof of keeping busy.

At team level that means: measure replies, not open rates. An open is a pixel, a reply is a human. Measure booked and held meetings, not sequences sent. Measure opportunities from defined sources, not leads in a database.

The yardstick behind it: every metric must carry a decision. If the number rises or falls and nobody changes their behaviour, it is not a steering metric, it is decoration. MQL volume falls into the second category almost every time.

And one practical test for the pipeline itself: plan coverage well above target. Not because your deals are bad, but because part of every pipeline is fiction. Price that in and your forecast gets honest.

Outro

✅ What works. A shared pipeline definition halves the internal arguments. Sometimes in weeks.

❌ What does not. Leaving MQL reporting in the dashboard "because we always did". That is inheritance, not steering.

⚠️ Warning. A pipeline definition without sales signature is just another marketing MQL. Then you changed nothing.

The deeper point: the MQL problem is not a metric problem, it is a leadership problem. Companies that do not write a shared pipeline definition live with the conflict. That is exactly where we started.

If you want to see what such a definition looks like, subscribe to the newsletter.

Frequently asked questions

Should I abolish MQLs entirely?

Not overnight. But your business decision should not lean on MQL volume. Use MQL as an internal signal, not as a steering metric.

What about lead scoring?

Lead scoring is fine, as long as sales co-defines the threshold and recalibrates it regularly. Without that joint calibration marketing scores in one direction while sales runs another.

Do I need new tools for this?

No. HubSpot, Salesforce, Pipedrive are enough. What you need is a shared definition and an SLA. That is process, not tooling.

How fast do I see the effect?

First effects in 4 to 6 weeks, mostly in per-stage conversion and lower internal friction. Volume effects take one sales cycle.

What about PLG, do I still need a pipeline definition?

Yes, even more urgently. Product Qualified Leads without a clear pipeline definition produce the same arguments with different vocabulary. Define when a usage signal becomes a pipeline signal.