setscale

Lead qualification framework for DM setting teams

August 25, 2026·15 min read
Woman around thirty with a short blonde bob in a terracotta blouse, writing a short qualification checklist on paper cards at a wooden counter beside a large window in a plant filled neighbourhood cafe office, warm morning light, teacup and notebook next to her

A lead qualification framework is the written rule that decides which inbox conversations become calls on a closer's calendar, and which ones do not. In a DM inbox it has to do that in three or four messages, before the thread cools and before anyone has a discovery call to lean on.

This is written for the person who runs several setters or several client accounts: agency owner, head of sales, ops manager. The problem at that scale is not knowing what a good lead looks like. It is that six setters each know it slightly differently, so the same lead gets booked on Tuesday and dismissed on Thursday, and nobody can tell whether the calendar is full of the right people or just full.

TL;DR

  • A qualification framework is only real when two setters, on the same thread, produce the same verdict. Anything else is a shared vocabulary, not a framework.
  • Five criteria are enough in a DM: fit, stated problem, ability to pay, decision access, timing. Score each one 0 or 1 against a written test, not a feeling.
  • Book at four out of five, with fit and ability to pay mandatory. Recycle at two or three. Disqualify below that or on any hard disqualifier.
  • The framework fails at the handoff, not in the inbox. A qualified lead that arrives at the closer without evidence quotes is an unqualified lead with extra steps.
  • Audit the disqualified pile, not just the booked one. False negatives are invisible in every report you send.

Table of contents

Why lead qualification breaks in a DM inbox

Classic qualification was designed for a call. You had thirty minutes, a person who had already accepted a meeting, and permission to ask direct questions about budget and authority. Almost none of that survives contact with a direct message.

Three constraints change the exercise. The first is length: a lead who wrote four words will not answer a five part questionnaire. The second is tone, because a DM reads as a conversation, and a conversation that turns into a form gets abandoned in silence rather than refused out loud. The third is time. Meta's messaging policies define a standard window of 24 hours after a person's last message during which you can reply freely, documented in the Messenger Platform policy overview. Outside that window your options narrow, so a framework that needs six exchanges to reach a verdict will regularly run out of room before it reaches one.

The consequence is not that qualification is impossible in the inbox. It is that qualification has to be designed as a sequence of small commitments rather than an interview, and that the criteria have to be few enough to hold in a setter's head at message two. That is what makes response time and qualification the same problem in practice: the faster you answer, the more of the window you still own to establish anything at all.

The four things every framework tries to establish

Every qualification method in circulation is a different arrangement of the same four questions. Seeing that plainly is what lets you stop shopping for acronyms and start writing your own tests.

Method What the letters cover What it optimises for Where it struggles in a DM
BANT Budget, authority, need, timing Speed of disqualification on cost Asking about budget at message two reads as a price interrogation
CHAMP Challenges, authority, money, prioritisation Leading with the problem instead of the wallet Still needs an explicit money question before the handoff
MEDDIC Metrics, economic buyer, decision criteria, decision process, pain, champion Complex deals with several stakeholders Far too heavy for an inbox, built for enterprise cycles
ANUM Authority, need, urgency, money Threads where the wrong person answers first Starting on authority sounds dismissive to a warm lead

The four things underneath are always the same: is this the right kind of person, do they have a problem you solve, can they act, and is there a reason to act now. Pick the arrangement that suits your offer, then throw the acronym away and write the tests. Your setters will never apply a letter, they will apply a sentence.

A lead qualification framework in five criteria

The version below splits the classic four into five, because in a DM the difference between a person who has a problem and a person who has said what it is matters more than anything else. A stated problem is evidence. An inferred problem is a guess a closer will inherit.

# Criterion The test that establishes it What counts as a yes The false positive to watch
1 Fit Does the person match the written profile for this account, by segment, size and role Their own words or public profile confirm it A lookalike from an adjacent market that never converts
2 Stated problem Have they named a problem your offer addresses They wrote it, in their words, in the thread You named the problem and they said yes
3 Ability to pay Can they work within the price band, tested as a band and not a number They confirmed the band without flinching or asked how payment works Enthusiasm read as ability
4 Decision access Are they the decider, or can they bring the decider to the call They say who decides and commit to including them Nobody asked, so it defaults to yes
5 Timing Is there a reason to act in the next thirty days They name an event, a deadline, a season or a cost that is running A polite "soon" with nothing behind it

Score each criterion 0 or 1. Book at four out of five, with criteria 1 and 3 mandatory. Recycle to nurture at two or three, and let the follow up sequence do its work instead of the calendar. Disqualify at zero or one, and disqualify immediately on any hard disqualifier regardless of score: outside the served territory, a service you do not sell, a competitor's employee, a previous non payment, a compliance exclusion.

The thresholds matter less than the fact that they are written down. A team that books at three out of five consistently is easier to fix than a team where the bar moves with the mood of the week.

Writing criteria two setters score the same way

A criterion is finished when a new setter and a two year veteran, reading the same thread separately, tick the same box. Until then it is an opinion with a number next to it.

Three rules get you there. Write the test as an observable event, not as a judgement: "they named a deadline in the next thirty days" survives two readers, "they seem motivated" does not. Attach one real example of a yes and one of a near miss to each criterion, taken from your own threads, because examples resolve edge cases faster than definitions. And require evidence for every yes, in the form of the lead's own sentence pasted into the record.

Then test it. Take twenty archived threads, have two setters score them independently, and compare. Anywhere they disagree is not a training problem, it is a wording problem, and the fix belongs in the criterion rather than in a coaching call. Run the same exercise as part of the thirty day ramp for every new hire, and you get a calibration measurement instead of a vague sense that someone is picking it up.

Where qualification happens in the thread

Sequence matters as much as content, because each criterion has a natural moment where asking it costs nothing and an unnatural one where it costs the thread.

  • First reply. Establish fit and open the problem. One question, never two. The reply that opens the problem is also the one that keeps the free form window alive.
  • Second exchange. Let them state the problem in their words, then reflect it back in yours. This is where a stated problem becomes evidence rather than a guess.
  • Third exchange. Ability to pay, framed as a band attached to what they described, and timing in the same message. Two questions land together here because they share a logic: what it costs and when it makes sense.
  • Before proposing a slot. Decision access, framed as an inclusion rather than a test: who else should be on the call so nobody has to repeat this.

Four touches, at most one question each except the third. If a criterion is still open when you are ready to propose a slot, propose the slot anyway and mark the criterion as unknown rather than assuming a yes. An honest gap travels well to a closer. A silent assumption does not, which is exactly the failure the setter and closer split is supposed to prevent.

The handoff contract with your closers

The framework produces its value at the handoff or nowhere. What the closer receives should be short enough to read in forty seconds and complete enough that they never open the thread to reconstruct it.

Six fields carry it: the five scores with an evidence quote for each yes, the unknowns named as unknowns, the lead's own sentence about the problem, the source of the lead (organic, referral or a click to message ad), the timezone and stated availability, and the hard disqualifiers checked. Nothing else, because everything else is where the summary starts to editorialise.

The contract runs in both directions. Closers accept or reject within a fixed delay, four working hours is a reasonable default, and a rejection comes with a reason picked from a closed list: wrong fit, no real problem, cannot pay, no decision access, no timing. A free text rejection teaches you nothing next month. A coded one becomes the mix you review, and it is what tells you whether the setters are loose, the closers are picky, or the ad set is bringing the wrong people entirely.

Running one grid across several client accounts

At agency scale the temptation is a bespoke grid per client. It looks attentive and it quietly destroys your ability to compare anything, because ten definitions of a qualified lead produce ten reports that cannot be read side by side.

Keep one spine, the five criteria and the scoring rule, identical everywhere. Vary only what is genuinely client specific: the fit profile, the price band, the hard disqualifier list. Write those three per account, on one page, signed off by the client in writing, and version them with a date. When a client disputes a booked call, you are then arguing about a document they approved rather than about taste.

Two habits keep it alive across a portfolio. Re-approve the fit profile and the band at every quarterly review, since a client who raised prices in March invalidated your criterion 3 without telling you. And sample threads across accounts weekly rather than per account, which is how you catch a setter who is drifting on their busiest client while looking fine everywhere else. The multi account playbook covers the rest of that operating rhythm.

SetScale is being built for exactly this shape, teams and agencies running several accounts and several closers. It is not open yet, so join the waitlist to hear when it is.

Three ways a qualification grid rots

Grids do not fail loudly. They erode, and the reports keep looking normal while they do.

Criteria drift at the end of the month. When the booked call target is behind, criterion 3 gets read generously. The tell is a jump in booked calls in the last week of the month with a drop in show rate the following month, visible only if you keep both numbers side by side in the client report.

The thread turns into an interrogation. Setters under pressure ask all five things at once. Reply rates fall, and the leads who do answer are the compliant ones rather than the qualified ones. The tell is a rising rate of threads that die after your second message.

Qualifying on interest instead of ability. The most enthusiastic replies feel like the best leads, so criterion 3 gets scored on tone. The tell is a healthy show rate combined with closers reporting the same sentence after every call: they loved it, they cannot do it right now.

Proving the grid works, in four numbers

Four numbers, reviewed monthly, tell you whether the framework is doing its job. None of them require a new tool.

  1. Booked to show rate, split by score. Leads scored five out of five should show more often than leads scored four. If they do not, your criteria are not measuring anything real.
  2. Closer rejection rate and its reason mix. The rate matters less than the mix. All rejections landing on one reason points at a single broken criterion.
  3. Qualified to closed rate. The only number that connects the grid to money, and the one to read alongside the true cost of each route to booked calls.
  4. The false negative sample. Pull twenty disqualified threads a month at random and re score them. Anything above one clear miss in twenty means the grid is too tight, and unlike the other three, nothing in your reporting will ever surface it on its own.

Read these together, not one at a time. A grid that looks strict on numbers 1 and 2 and produces misses on number 4 is not a strict grid, it is a narrow one, and the difference is entirely in what it costs you.

Where SetScale fits

SetScale is an AI setter built for teams and agencies rather than for a single inbox: several accounts, several closers, a per seat view of what happened. The qualification layer is the part of that shape this article describes, applied consistently across accounts rather than rewritten per setter.

The product is not open yet, so there is nothing to try today and no pricing to quote. A white label direction is part of the plan for agencies who resell the service under their own brand, and a seven day free trial is planned for launch. If a framework applied identically across every account you run is the thing you are missing, join the waitlist and you will hear when it opens. If you want the surrounding architecture first, the setting infrastructure overview is the place to start, and the closer calendar problem is what all of this is ultimately for.

FAQ

Is BANT still usable for DM qualification? The four things it covers are still the right things. The order is wrong for an inbox, because budget first reads as a price check on someone who has not yet said what they need. Keep the coverage, lead with the problem.

How many questions can a setter ask before it feels like an interrogation? One per message for the first two exchanges, two in the third when they share a logic. The limit is not a number of questions, it is whether each one visibly follows from what the lead just wrote.

Should a setter mention price in the DM? Mention a band, never a quote. A band tests ability to pay without turning the thread into a negotiation the closer then has to reopen from a worse position.

What if the client's ideal profile is vague? Write the version you observe, send it to them, and ask them to correct it. A profile they corrected is a profile they own, and it converts a recurring argument about lead quality into a document.

Do you need a human setter to qualify, or can automation do it? Automation is reliable on the mechanical parts: fit against a written profile, hard disqualifiers, the evidence record. The stated problem and the decision access still benefit from a human reading, which is why the split matters more than the tooling, and why the WhatsApp setup should route rather than decide.

Conclusion

A lead qualification framework is not a document you write once, it is a definition your team can apply the same way on a Thursday afternoon. Five criteria, a written test for each, evidence for every yes, a coded rejection path back from the closers, and a monthly look at the pile you threw away.

Start with the calibration exercise: twenty archived threads, two setters, independent scores. The disagreements will tell you which criterion is actually just an opinion, and fixing that one sentence is worth more than any new acronym.

If you run several accounts and want the qualification rule to be the same in all of them without rewriting it per setter, join the waitlist.