The MQL is NOT Dead!

“The MQL is dead” has been a recurring claim in B2B marketing for the better part of a decade. Drift published an entire book making the case at least eight years ago, and the headlines have continued to appear ever since. This year, the Marketing Qualified Lead officially turns 20. As Dale W. Harrison puts it, “next year, it won't have to keep using a fake ID to get into the bars.”

So why is it still here?

That was the starting point for Episode 1 of B2B Effectiveness: Evidence-Based Marketing Ideas for B2B Practitioners, in which Liam Moroney sat down with Dale W. Harrison, a data-driven B2B marketing strategist known for his evidence-first views on lead scoring, demand generation and the metrics the industry routinely takes for granted.

Why the “MQL is dead” myth won't die

The reason the MQL keeps coming back is fairly straightforward: the underlying job it represents isn't going anywhere.

Sales needs a supply of potential customers to work. That was true before anyone had invented the MQL, and it will remain true regardless of what the next generation of marketing terminology looks like. Without marketing generating opportunities for sales to pursue, sales has to create that pipeline itself, typically through cold outreach and individual prospecting rather than through campaigns that can reach thousands of people at once.

As long as a business has both a sales function and a marketing function, there needs to be some kind of handoff between them. At some point, marketing has to decide that a contact is worth a salesperson's time. Whether that handoff is called an MQL, an MQA, an MQO or Bob is largely beside the point. The terminology can change; the underlying organizational need does not.

The MQL itself, however, is much younger than the job it was created to formalize.

Marketing has been passing leads to sales for as long as the two functions have existed. Glengarry Glen Ross hardly invented the idea of sales leads, but it captures just how old the underlying relationship is. The MQL, as a defined concept, emerged roughly 20 years ago through the SiriusDecisions demand waterfall model, and the framework has continued to influence B2B marketing through Forrester's ownership of the brand.

So if the function isn't going away, why does the MQL feel so broken?

Because the problem was never really the existence of a marketing-qualified lead. It was the assumption that we knew what “qualified” meant.

The problem was never the M or the L. It was the Q.

Marketing generating leads isn't particularly controversial. The difficult question is deciding which of those leads deserve attention from sales.

Harrison describes an example from a RevOps team at a CRM company roughly seven or eight years ago. The team had become frustrated with its elaborate, points-based lead scoring model, so they began assigning random scores to a subset of contacts using a random number generator. Those contacts were then mixed in with leads scored by the company's actual model and passed to sales.

The result was revealing. There was no meaningful difference in close rates between the randomly scored leads and those selected by the supposedly sophisticated scoring system. If that result holds, the scoring model isn't providing much predictive value at all.

The underlying problem is an assumption that has shaped much of lead scoring for the past 20 years: that engagement data can be treated as evidence of buying intent.

Intent is an internal state. A person can be interested in something without clicking, downloading or visiting a website, and they can perform all of those actions without having any intention of buying. What marketers can observe is behavior. Calling that behavior “intent data” doesn't change what it is.

The weakness becomes obvious when you look at how conventional scoring works. If downloading an eBook earns a certain number of points, downloading two can earn twice as many. Ten downloads can earn ten times as many.

But nobody seriously believes that someone who downloads ten eBooks is ten times more likely to buy than someone who downloads one. At some point, additional activity stops providing much new information. A scoring system that continues to add points regardless of that diminishing value creates an impression of increasing certainty that the underlying evidence simply doesn't justify.

Why more data doesn't mean more information

This is where much of the industry's recent enthusiasm for AI-driven scoring runs into the same problem.

An AI model can process vastly more signals than a manually designed scoring system. That can be useful, but volume alone doesn't turn weak evidence into strong evidence. Ten observations of the same behavior don't necessarily tell you ten different things.

Dale uses the distinction between data and information to make this point. Data is the container in which information arrives. You can have one useful piece of information represented by a single data point, or repeated across hundreds of records. The number of observations doesn't tell you how much genuinely new information they contain.

This matters for approaches such as Marketing Qualified Accounts as well. An account-level model may aggregate activity from multiple people at the same company, but if the underlying signals don't reliably distinguish buyers from non-buyers, multiplying those signals across a larger group doesn't solve the problem. It simply gives the same scoring logic more data to process.

Long-running Gartner benchmarks provide a useful backdrop to the argument. MQL-to-sales conversion remains low, and even opportunities that have made it much further through the qualification process still don't convert at anything close to certainty. That is consistent with a much more modest interpretation of what marketing data can tell us: it can help establish whether an organization looks like a plausible customer, but it has far less power to tell us whether that organization is ready to buy today.

The 95-5 problem

This is where the familiar “95-5 rule” becomes relevant.

At any given moment, only a small proportion of potential buyers are actively in-market. If 5% of a relevant audience happens to be considering a purchase today, the remaining 95% aren't necessarily bad prospects. They simply aren't buying at this particular moment.

The difficulty is identifying which five are in-market without confusing observable engagement with actual buying intent.

A person can research a category extensively and still have no immediate intention to purchase. Another can be ready to buy after a conversation with a colleague and leave almost no detectable digital trail. The information available to a marketer is therefore incomplete by definition.

That doesn't mean prediction is impossible. It means the promise of knowing with certainty who is ready to buy is much harder to defend.

What sales actually needs: thinking in bets, not certainties

This is the shift Harrison argues is more useful. An MQL was never really a guarantee. It was always a bet, even if the language surrounding it sometimes made it sound like a declaration of fact.

Sales understands this instinctively. A salesperson knows that most conversations will not become deals. The job is to allocate time and effort where the probability of success is sufficiently attractive.

A better qualification system should work in much the same way. Rather than claiming to identify the people who are definitely ready to buy, it should improve the odds of finding worthwhile opportunities.

Harrison points to three questions that can help move those odds:

• Are they likely to buy from this category at all? This is where ICP fit and some forms of engagement data can be useful.
• Are they likely to buy from a company like ours? Market position, reputation and the type of organizations a business has successfully sold to all matter here.
• Are we likely to be able to sell to a buyer like them? Sales organizations develop their own areas of strength, experience and specialization, so the fit between buyer and seller matters in both directions.

None of these questions tells a salesperson when a deal will close. What they can do is improve the quality of the bet being made.

That is a much more realistic role for marketing qualification: reducing uncertainty rather than pretending to eliminate it.

The MQL isn't dead. It just needs new tires.

A car with four flat tires isn't necessarily beyond repair. The useful response is to fix the tires and get it moving again.

The same logic applies to the MQL. The handoff between marketing and sales remains necessary, so replacing the acronym doesn't solve the underlying problem. Calling something an MQA or an intent-qualified lead while continuing to rely on the same points-based logic simply moves the terminology around without changing the mechanism.

The more meaningful change is to rethink what qualification is supposed to accomplish. If engagement data cannot reliably tell us when someone is ready to buy, it shouldn't be treated as if it can. The useful question is whether the available evidence can improve our estimate of where sales effort is most likely to pay off.

That leads to a more interesting question for the next episode. If engagement and firmographic data can't tell us exactly when someone is ready to buy, what other factors can move the odds, and which of them can we gather enough reliable evidence on to use?

Harrison and Moroney pick up that question in Episode 2 of the series.