Lead Qualification: Criteria, Scoring, Routing and the Feedback Loop That Keeps It Honest

Lead qualification is the process of deciding whether an enquiry is a good enough fit, and ready enough, to deserve a salesperson's time. Good qualification has three possible outcomes: pursue now, nurture for later, or decline. Most weak processes have only the first and a vague pile for everything else.

It sits between marketing capturing the lead and sales owning the opportunity, which is why disputes about it are really disputes about definitions. The sharper the written criteria, the fewer arguments. A criterion that two people can apply to the same record and reach the same answer is a working criterion.

Qualification is also a courtesy to the lead. Someone who needs a quick answer on a small job is poorly served by a three-call discovery process, and a prospect who is early in research is poorly served by a hard pitch. A good process matches the effort to the situation, which is often the difference between a lead that warms up and one that quietly goes elsewhere.

BANT asks about Budget, Authority, Need and Timeline. HubSpot's guide to it cautions against treating budget as a single dollar figure and suggests looking at how a prospect spends and who sits on the buying committee. MEDDIC is another widely used checklist; practitioners commonly expand it as metrics, economic buyer, decision criteria, decision process, identified pain and champion, though versions differ and this is a practitioner convention, not a published standard. Many teams adapt such frameworks or combine pieces of them.

Neither is a law. BANT is quick, suits shorter sales with a clear buyer and can feel blunt when a purchase involves several stakeholders. MEDDIC-style questions suit longer, committee-led sales and are heavier to run on every inbound form. A service business with a same-week quote cycle often needs three questions, not a framework: what do you need, where, and by when.

Whichever you pick, turn the letters into specific questions that suit your offer and record the answers in fields, not in free-text notes. A framework that lives only in a rep's head cannot be reported on, and the answers are what you need when you later examine why some leads closed and others did not.

Use scoring to sort volume and people to judge exceptions. A score assigns numbers to records based on attributes and behaviour. HubSpot's lead scoring tool, for example, can build a fit score from properties like region, company size and industry, and an engagement score from interactions with campaigns. That is a good way to put the obvious fits at the top of a queue.

Scores have blind spots. They reward activity, so a student downloading everything can outscore a buyer who submits one form and calls. They also carry whatever assumptions you wrote in. For high-value enquiries, a human should read the message. A practical split is automatic routing for clear matches and clear non-matches, and a short human review for the middle.

Test a score before you trust it. Take last year's closed and lost records, apply the proposed scoring, and see whether closed deals scored higher. If they do not, the score is describing something other than buying likelihood. Revisit it when your offer, pricing or audience changes, since the signals that mattered last year may not be the ones that matter now.

Routing sends a qualified lead to the right person with a clock attached. Decide the rule (territory, product line, account ownership, or round robin among available reps), the response expectation, and what happens when the first person does not act. A rule nobody can explain in a sentence usually has exceptions that nobody maintains.

Keep routing separate from qualification. Qualification says whether a lead is worth pursuing; routing says who pursues it. Mixing them produces rules like "send to Dana unless the score is under 40 and it is Tuesday", which are impossible to audit. When a lead fails to match any rule, route it to a named owner, not a shared inbox.

Think about time zones, holidays and absences. A rule that routes to one person fails when that person is on leave, so define a fallback and review the routing report for leads that sat unassigned. Where a lead matches an existing customer or open deal, send it to the owner of that relationship instead of creating a parallel conversation.

A lead that sales declines without a reason is lost information. Use a short, mandatory list of reasons: not a fit for the offer, wrong location or size, no budget, no response after attempts, duplicate, student or job seeker, existing customer, and spam. HubSpot places Lead Status sub-stages inside its Sales Qualified Lead lifecycle stage, which is one place such a field can live; the point is that the field exists and is required, wherever it is stored.

Reasons become useful in aggregate. If the top reason is "wrong location" and most come from one source, the fix is in the form or the ad targeting, not in sales. If many are "no budget", the criteria or the offer pages may be attracting the wrong audience. Each reason points at a different owner, so a reason with no owner is a field nobody will read.

Keep the list short enough to choose from in a few seconds. Fifteen reasons produce random picks, and an Other option with a free-text box becomes the default. If a reason keeps appearing under Other, promote it to the list. Review the list twice a year and merge the options nobody uses.

Once a month, put a sample of accepted leads and rejected leads side by side with their eventual outcomes: closed, lost, still open. Ask which signals separated them. If leads that sales rejected later became customers elsewhere, the criteria are too strict. If accepted leads rarely progress, they are too loose.

Change one criterion at a time and write down the date, so the next review can tell what changed the pattern. Adjust scores sparingly; a model rebuilt every fortnight cannot be compared with itself. The loop works only if sales and marketing look at the same records together, because each side sees only half the evidence.

Give the loop a single page: the criteria, the date each was last changed, the reason, and the result seen after. Over a year this becomes a small history of what your market responded to, and it prevents the same argument from being held every quarter by people who were not present for the last one.