From MQL to DQL: Lead Qualification as a State, Not a Status
Written by LeadScale
Every few months another article announces that the MQL is dead. The marketing qualified lead, the argument goes, is a relic of a form-fill era that no longer matches how anyone buys. There is something to the complaint, but burying the MQL does not leave anything useful in its place, and defending it ignores a real problem. This piece takes a different route. The issue is less that qualification is worthless and more that we have been recording it in a way that does not match how buying actually unfolds. It is worth exploring what a more honest model of lead qualification might look like, and LeadScale has a hypothesis about that, a concept it calls the dynamically qualified lead, or DQL, which the second half of this article sets out plainly as a hypothesis rather than a finished answer.
A marketing qualified lead is a lead that has crossed some threshold, a score, a set of behaviours, a demo request, and been passed to sales as ready for follow-up. That is a reasonable thing to want. The assumption underneath it is the part worth examining: that qualification is a status, fixed at the moment it is granted and stable afterwards. A lead is qualified, and then it stays qualified, sitting in a queue with a stamp on it.
This article looks at why that assumption has weakened, what the evidence actually shows about how buyers behave, and where the idea of qualification might be heading. The verified facts carry the argument. The forward-looking part is offered as a direction of travel, not as a shipped mechanism.
What a Marketing Qualified Lead Assumes
Start with what the MQL gets right, because it is not nothing. Faced with more inbound than a sales team can work, some way of sorting is better than none, and a threshold that separates the more promising leads from the rest is a sensible tool. Where it falls down is the assumption that the result, once granted, stays true.
An MQL is a snapshot. It says that at a particular moment, and on the strength of particular signals, one contact looked ready. Each part of that is contingent, from the moment it was granted to the single contact it rests on. The moment passes, the signals age, and the one contact is rarely the whole story. The model treats a snapshot as if it were a standing fact, and most of the friction operators feel with the MQL comes from that gap between a stamp granted once and a situation that keeps changing. A lead that qualified in March, on a whitepaper download by a mid-level analyst, is filed as qualified in June even if the project has since been shelved, the analyst has moved teams, and the budget has gone elsewhere. Nothing in the model notices.
Why a One-Time Stamp No Longer Fits the Buyer
The evidence on how B2B buying works has moved a long way from the tidy funnel the MQL was built for. Gartner’s research on the B2B buying journey describes buying not as a sequence of stages but as a loop. Buyers move across a set of buying jobs, problem identification, solution exploration, requirements building, supplier selection, validation, and building consensus, and they revisit each of those jobs at least once rather than completing them in order. The journey does not end where a form-fill sits; it circles back on itself.
Two further findings sharpen the mismatch. Gartner reports that buyers spend only around 17% of their total purchasing time meeting with potential vendors, and that time is split across all the vendors they are considering, so any single supplier sees a sliver of the process. And Forrester, in the research behind its B2B Revenue Waterfall, notes that more than 80% of B2B purchases involve groups of three or more people, which is why the Waterfall shifts the focus from the individual lead to the buying group. Forrester sells that framework, so read the framing with that in mind, but the buying-group finding is a survey result rather than a sales claim.
Set those together. A qualification granted once, on one person, at one moment, is being asked to stand in for a process that loops, that is mostly invisible to any one vendor, and that is run by a group rather than an individual. The static markers of the MQL and SQL were not wrong when buying was more linear. They fit less well now, and the intent signals that would let a model keep up with a moving buyer are exactly what a one-time stamp throws away.
"Qualified" Has Already Fractured
There is a quieter problem that the “is the MQL dead” debate tends to skip. The word “qualified” has already stopped meaning one thing.
The marketing-analytics firm marqeu, looking at MQL-to-opportunity conversion, found that the reported rate for the same funnel stage ranges anywhere from 13% to 45%, and that the spread is not measurement noise. It comes from the fact that “MQL” is defined differently across scoring models. A team that calls any high-intent form-fill an MQL ends up with a small, high-quality pool that converts north of 35%. A team that calls any lead crossing a behavioural score an MQL ends up with a larger pool converting closer to 13 to 15%. The industry median sits somewhere around 15 to 21%, but that average hides the fact that the same three letters describe several incompatible things. marqeu sells scoring and waterfall consulting, so treat its specific numbers as vendor-published rather than audited; the useful part is the range and the reason for it.
That is the real state of play. Different organisations already run different qualification standards and call them all the MQL. When a trade publication such as Demand Gen Report argues, in an opinion piece this year, that the MQL is dead because the industry has optimised for volume and qualifies poorly, it is describing a symptom of this fragmentation, not discovering a new one. The term has quietly become a label that can mean almost anything, which is a strange foundation for a metric that sales and marketing are supposed to agree on.
The DQL Hypothesis: Qualification as a State
Here the article moves from what the evidence shows to what LeadScale thinks it might imply, and it is worth being clear about the change in footing. What follows is a hypothesis, first set out by LeadScale’s Robin Caller in a May 2026 piece, and it is offered as a direction of travel rather than a mechanism anyone can buy or has proven to work.
The hypothesis starts by questioning the permanence. If the MQL and SQL are static markers, a lead is qualified because it hit a score and then it waits, what if qualification were treated instead as a state that fluctuates? In Robin’s framing, qualification would not be a trophy a lead wins once but a state of being that rises and falls. If an account’s intent surges on a Friday, its qualification would rise with it; if a stakeholder’s behaviour shifts on Monday, the read would recalibrate. The name he gives this is the dynamically qualified lead, or DQL.
The idea leans on a composable architecture, where data moves freely between the data warehouse and the activation layer, so that the qualification could in principle be recalculated at the edge of each interaction rather than fixed at a form-fill. Robin pairs it with what he calls differential activation, designing the response to a buying group in real time from the signals as they arrive, rather than following a journey mapped six months earlier, and with a learning loop, using AI less to write copy than to work out which actions actually moved an account and to adjust accordingly. The retrospective nature of scoring, which ranks a lead after activity has accumulated, is part of what the dynamic frame is reacting against.
None of this is a shipped product or a proven outcome, and it would be wrong to present it as one. No evidence in this article shows a DQL running in production and beating an MQL, because none exists yet. What the hypothesis has going for it is that it follows logically from the verified facts above: if buying is a loop run by a group and mostly invisible, a qualification that can move with it is at least the right shape for the problem, whatever the practical difficulty of building it.
The Sharpest Edge: A State Can Fall
If there is one part of the hypothesis that separates it from the general drift toward more intent data and more scoring, it is this. Most intent and scoring systems only ever move in one direction. They add points, raise scores, and escalate; they are built to find reasons to chase harder. A qualification state, by contrast, could fall as well as rise.
That matters because a falling state carries a different instruction. Robin’s own framing is blunt about it: often the best next action is to decline, detarget, or suppress, to stop trying to convert a person or a buying group because doing so is no longer economic or valid. A static MQL has no way to express that. Once a lead is stamped and queued, the model has no mechanism for it to expire; it can only ever be worked or ignored, not actively stood down. A dynamic state, at least in principle, could recognise when a buying group has gone cold or when a record has stopped being valid, and treat that as a signal to stop spending rather than a gap to fill with more follow-up.
This is where the idea connects to Zero Waste, the principle that a demand engine should not keep spending on demand that is no longer worth pursuing. Knowing when to stop is as much a part of not wasting effort as knowing whom to pursue, and a qualification that can only rise has no way to tell you. Whether the connection to data truth and validating at source can be built into a working system is an open question. As a principle, a state that can fall is the part of the hypothesis that seems most clearly to point at something the current models cannot do.
What an Operator Can Take From This Now
It would be easy to read all this as a promise to wait for, and that is not the intention. The DQL is a hypothesis, and building anything like it is a serious undertaking that no one in this article claims to have finished. The useful part is available now, and it is a change of mental model rather than a change of tooling.
Treat qualification as a moving read rather than a stamp. A lead that qualified a month ago is a question, not a settled fact, and it is worth re-checking rather than assuming the status holds. Watch for falling states, not just rising ones, and be willing to stop working an account that has gone quiet rather than escalating out of habit. Keep the definition of qualified honest within your own team, since the fragmentation marqeu describes starts at home, and a shared, current definition is worth more than a sophisticated model reading an outdated one. This is also the ground on which sales and marketing stop arguing over lead quality: a quality that is re-read rather than asserted once is harder to dispute. None of these habits needs a new platform. They need a willingness to treat last month’s qualification as a claim to be tested rather than a fact to be inherited, which is most of what the DQL idea is really asking for.
Most of that discipline lives at the point of capture and re-validation, where an engine such as the LeadScale Engine checks and, in a more dynamic model, could re-check a record against compliance, truth and fit rather than trusting a one-time stamp. Whether the full DQL arrives as Robin describes it or in some other form, the direction is worth watching, and the change in how you think about qualification is worth making now.
Frequently Asked Questions
A dynamically qualified lead is a concept, coined by LeadScale, for treating lead qualification as a state that is continuously recalculated rather than a status stamped once. In the idea, a lead’s qualification rises and falls as an account’s intent and behaviour change, rather than being fixed at the moment it crosses a scoring threshold. It is a hypothesis and a direction of travel, not a shipped or proven mechanism, and no evidence yet shows a DQL running in production and outperforming a traditional marketing qualified lead. The useful part today is the mental model it offers.
No, and the “MQL is dead” framing is less useful than it sounds. The marketing qualified lead is still a workable way to sort more inbound than a team can handle. The real problem is that it records qualification as a one-time status, which sits awkwardly against research showing that B2B buying loops rather than runs in a line, that buyers spend little of their time with any one vendor, and that most purchases involve a group rather than an individual. The MQL is not dead so much as mismatched to how buying now works.
An MQL is a status: a lead crosses a threshold once and is marked qualified, and that mark stays put until someone works or discards the lead. A DQL, in LeadScale’s hypothesis, is a state: the qualification is recalculated as signals change, so it can rise when intent grows and fall when it fades. The practical difference is that a status can only be added to a queue, while a state can also decline, which would carry the instruction to stop rather than to chase. The DQL is a proposed idea, not an available product.
Several reasons overlap. The buyer keeps moving after the moment the lead was qualified, so a stamp granted then is often out of date by the time sales acts. The individual who filled in the form is usually one of several people in a buying group, and their readiness is not the group’s. And “qualified” is defined inconsistently across teams and tools, so one company’s MQL is not comparable to another’s, with reported conversion for the same stage ranging from roughly 13% to 45% depending on the definition. The lead was often never as qualified, in a durable sense, as the stamp implied.
In the way most systems are built today, no: a scored or stamped lead only moves up or holds, because scoring and intent models are designed to escalate. In the DQL hypothesis, yes: a qualification state could fall as an account goes quiet or a record stops being valid, and a falling state would be a signal to stop spending on that lead rather than to chase it harder. That ability to expire is one of the clearest things a dynamic state could do that a static marketing qualified lead cannot, though it remains a proposed capability rather than a proven one.
No. The dynamically qualified lead is a hypothesis and a direction of travel, not a product on sale or a mechanism shown to work in production. Nothing in this article claims a DQL exists, has been deployed, or beats the MQL, because no such evidence exists yet. What is available now is the mental model: treat qualification as a moving read, watch for states that fall as well as rise, and keep your own definition of qualified current. Those habits are useful whether or not a full DQL is ever built as described.
