The short version
When pipeline is thin, the reflex is to add volume. But more leads into an inconsistent qualification process just produces more inconsistency, faster: more meetings that go nowhere, more reps improvising their own definition of “qualified,” more forecasts nobody trusts.
Fix the definition first. One framework, written down, wired into CRM stage exits, and tested against your last 20 won and 20 lost deals. Then feed it volume. Quality of process multiplies quantity; zero times any number is still zero.
The won-and-lost autopsy, step by step
Export 20 won and 20 lost-qualified deals into one sheet. Add columns for facts that were knowable at the first meeting: company size band, who attended, whether a budget owner was named, whether a dated trigger existed, how the lead arrived, and how many stakeholders were on the call. Ignore anything learned later; you are looking for what predicts, not what explains afterwards.
Colour won rows green and lost rows red, then sort by each column. Most teams find two or three columns where the colours separate clearly and several where they do not. Those separating columns become your criteria. The non-separating ones, often favourites like job title or industry, get demoted from requirements to notes.
Writing criteria that can be checked
A criterion is usable if two reps would give the same answer looking at the same record. “Strong need” fails. “Prospect named a specific process that is failing and who owns it” passes. “Budget exists” fails; “Budget holder named, and the prospect confirmed funding source or timing” passes. Rewrite until each line is a yes or no with evidence in a note field.
Cap the list at five bullets. Every additional criterion halves compliance because reps start to treat the list as paperwork. If you want a sixth, remove one. A short list applied by everyone beats a thorough list applied by nobody, and you can lengthen it later once adherence is stable.
Wiring the definition into the CRM
In HubSpot or Salesforce, create one required field per criterion on the opportunity, and make the stage change from Discovery to Qualified fail validation unless all five are populated. Add a single-line evidence field for each. Expect pushback in the first fortnight; reps will fill fields with rubbish unless a manager samples them weekly.
Run a Monday report of opportunities created in the last seven days with any criterion empty or evidenced by words like “TBC” or “good”. Name the rep in the list, privately. Within three weeks the rubbish disappears, and the stage conversion rates per rep start to converge, which is the signal that you are applying a definition rather than a mood.
When to turn the volume back up
My rule is two consecutive months in which the gap between reps on stage-two conversion is under ten points and forecast accuracy improves. Until then, extra leads simply spread the inconsistency across a bigger base. When you do scale, add volume in 25% steps and re-check conversion after each step.
Also review what you stopped booking. If a rep says a lead that failed the criteria would have closed, treat it as a hypothesis, log it, and test it on the next ten. Criteria should change when evidence says so, not when someone is frustrated.
Worked example (illustrative)
A hypothetical team of four reps books 40 meetings a month and tightens its definition over three months without adding any new lead sources.
What changed: Fewer meetings produced more opportunities because unsuitable leads stopped being booked, and reps stopped inventing their own standards. Time saved on 12 non-starters per month went into follow-up on the opportunities that mattered.
Do it: the steps
Tick them off as you go. Your progress stays on this device.
Traps to avoid
- ×
Treating the meeting count as the success metric. A meeting that should never have been booked is a cost, not an output.
- ×
Letting each manager tweak the definition for their own team, which recreates the improvisation you were trying to remove. Change the definition centrally and version it.
- ×
Adding qualification criteria after the meeting is booked. The definition must apply before the calendar invite goes out, or it only measures disappointment.
What to track
- Lead-to-qualified-opportunity rate
- qualified opportunities divided by meetings held; track by source and by rep, and watch the spread.
- Stage-two conversion spread
- highest rep rate minus lowest rep rate, ideally under 10 points once stable.
- Forecast accuracy
- absolute difference between forecast and closed revenue divided by forecast, reviewed quarterly.
Benchmarks are typical planning ranges, not guarantees. Calibrate to your own data.
Questions people ask
Should marketing and sales share one qualification definition?
They should share one handover definition, yes, but it does not need to be identical. Marketing qualifies on fit and engagement signals; sales qualifies on need, buyer and timing. The handover threshold is agreed jointly, written down, and reviewed monthly using the accepted-versus-rejected lead report. Disputes should be resolved with data from closed deals.
What if there are too few won deals to analyse?
Use what you have and say so. Ten wins and ten losses still reveal obvious patterns, though conclusions are weaker. Supplement with interviews of two or three customers about why they bought, and borrow a standard framework as a starting hypothesis. Revisit the definition after every ten new closed deals rather than waiting for a perfect sample.