Verified against ChatGPT · 2026-08-13
Build a lead qualification scorecard from your last 10 closed-won and closed-lost deals
Derives a practical lead qualification scorecard from patterns in your own actual past deals rather than a generic BANT template, weighting the criteria that historically separated your real wins from your real losses.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Build me a lead qualification scorecard — but base it on my actual past deals, not a generic BANT or MEDDIC template that ignores what's actually predicted wins and losses for my business specifically. CLOSED-WON DEAL PATTERNS Almost all closed-won deals had a champion who was personally frustrated with the current process, not just assigned to evaluate options, and involved 3 or fewer stakeholders. CLOSED-LOST DEAL PATTERNS Lost deals frequently involved a champion evaluating on behalf of a committee with 5+ stakeholders, and initial contact came from a generic content download, not a specific pain point. WHAT I CAN REALISTICALLY LEARN ABOUT A LEAD EARLY (FIRST CALL OR FORM) First call always covers current tooling, team size, and who initiated the evaluation; budget is rarely disclosed that early. DEAL SIZE / SALES MOTION Average deal ~$18k ACV, sales cycle 4-8 weeks, one rep handles the full cycle. From the won and lost patterns given, identify the 4-6 factors that actually differed between deals that closed and deals that didn't — not a generic checklist of "budget, authority, need, timeline" applied without evidence, but factors you can genuinely infer separated the two groups based on what I told you. For each factor, state explicitly what evidence from the patterns supports including it — if a factor seems intuitively important but isn't actually supported by the won/lost data I gave you, leave it out or flag it as unconfirmed rather than including it because it's a standard qualification category. Weight each factor by how strongly it appears to differentiate outcomes, not equally by default. For each factor, specify how it can actually be assessed at the point I have it available (early call or intake form) — a factor that can only be known after weeks of a sales cycle isn't useful as an early qualification criterion, even if it's genuinely predictive later. Build a simple scoring rubric (e.g., 0-2 points per factor) and a threshold for what separates a lead worth pursuing now from one to nurture or disqualify. Do not invent additional deal patterns or assume information about my business I haven't given you. Output: a table (Factor, Evidence From My Deals, Weight, How To Assess Early, Score Range) followed by the pursue/nurture/disqualify thresholds.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The instruction to derive factors from the actual won/lost patterns given, rather than defaulting to BANT or MEDDIC, matters because those frameworks are generic industry checklists that a model will reach for immediately if not explicitly redirected — they sound authoritative and complete, but a factor like "budget" being on a standard checklist says nothing about whether it actually predicted outcomes in this specific business's deal history, which is the only evidence that should determine what's on this particular scorecard. Requiring each factor to cite the specific evidence from the given patterns, and explicitly permitting the model to flag an intuitively-standard factor as unconfirmed rather than including it by default, is the check against a well-documented failure mode: language models asked to build a "qualification framework" tend to backfill in familiar categories (authority, timeline, competitive situation) that feel complete even when the actual data provided doesn't support them, because those categories are heavily represented in training data about sales generally. The requirement that every factor be assessable at the point it would actually be used — first call or intake form, not deep into the sales cycle — closes a subtler gap: a factor can be genuinely predictive of deal outcome while being useless for early qualification if it's only knowable after weeks of discovery, and a model not told to check for this will happily include lagging indicators as if they were usable filtering criteria on lead one. Weighting factors unequally rather than defaulting to a flat checklist reflects that real qualification signals aren't equally strong, and forcing the model to differentiate weight based on the given evidence produces a rubric that actually discriminates between good and bad leads instead of one where every box checked looks equally reassuring regardless of its real predictive value.
What you get back
Factor: Champion is personally frustrated with current process (vs. assigned to evaluate). Evidence: present in nearly all closed-won deals, largely absent in closed-lost. Weight: high (2 points). How to assess early: ask directly on the first call what specifically prompted them to look for a solution now, and listen for a personal pain statement vs. a generic mandate. Score range: 0 (assigned to evaluate, no personal stake) to 2 (unprompted, specific frustration stated).
Verified against
ChatGPT GPT-5.1 · 2026-08-13
Changelog
- 2026-08-13 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — Google Ads management, built and maintained for you — that's Scult's day job.
EXPLORE GOOGLE ADS MANAGEMENT
