Lead scoring: which questions actually predict a sale
Scoring only works if the scored questions predict buying: timeline, who decides, size of the job, budget band. Three levels per question (0 / 5 / 10) are enough, and disqualification belongs in a separate flag, never in negative points. After four to six weeks, cross the scores with closed jobs: if every score band closes at the same rate, the questions are wrong, not the weights.
Almost everyone who builds a quiz ends up attaching points to the answers. Almost nobody goes back to check whether those points have anything to do with revenue. That check is the difference between a score that organises your day and a decorative number on a dashboard.
What is a lead score?
Every answer adds points, and the total travels with the lead. It is not a grade for the customer — it is your call order. With twenty enquiries in one morning, the 85-point lead gets the first call and the 20-point one can wait until the afternoon or get an email instead.
Which questions actually predict a sale?
The ones about the ability and the will to buy now, not the ones about taste:
- Timeline — «when do you want to start?». The single best separator between a buyer and a browser, and an easy question to answer.
- Who decides — owner or tenant, private client or building committee, one signature or several. An enthusiastic tenant cannot sign off a renovation.
- Size of the job — square metres, number of rooms, units. A slider beats a list of ranges and gives you a number you can estimate from.
- Budget band — the exact figure is not needed; «under X / X to Y / over Y» is plenty.
- Starting point — new build, partial refit, urgent repair. It changes both the price and the type of customer.
- Whether they already have other quotes — a sign that they are comparing and that the timeline is real, not that they will haggle.
Notice what is missing: favourite colour, style preference, «how did you hear about us». Useful for marketing, but they do not predict a close, so they should not move the score.
How many points per answer?
Three levels per question is enough: 0 for the answer that tells you nothing, 5 for the middle, 10 for the one that points at a purchase. Keep the total near 100 and anyone on the team can read it at a glance. Tuning a weight from 7 to 8 is wasted effort — the precision lives in choosing the right questions, not in the weights.
Disqualifying is not «scoring low»
If someone fails a hard requirement — outside your service area, a tenant when you need the owner — that is not «few points», it is a no. Subtracting points blends two different things and produces the most awkward case of all: a lead that scores high on everything else, lands at the top of your list, and cannot legally hire you. Disqualification belongs in its own flag.
What about the lukewarm leads?
A mid-scoring lead is rarely a bad lead — it is usually the same customer a few months earlier. A figure often quoted is that around half of qualified contacts are not ready to buy yet; it circulates mostly in software vendors' publications, so treat it as an order of magnitude rather than a measurement. The practical conclusion holds either way: if you treat «not now» as «no», you are giving that half away.
Give them a different exit instead — a more informative result and a reminder a few weeks later, rather than the message you send a hot lead. Different result, different message, same quiz.
How do you know your scoring is right?
With one check that takes half an hour and is worth doing monthly:
- Wait until you have 30 to 50 leads with a known outcome — won or lost.
- Group them into three score bands: low, medium, high.
- Work out the close rate of each band.
- If the high band closes clearly more often than the low one, the scoring works — use it to order your day.
- If all three bands close at a similar rate, your questions predict nothing. Change the questions, not the weights.
It is the same principle as the funnel metrics that matter: without closing the loop between what you measure and what you invoice, every number looks fine.
How Quizanta does it
Any option can add points and, when needed, disqualify the lead through a separate flag. You can then create several result screens with score ranges: the 80-point lead sees a priority response and a different WhatsApp message from the 25-point one, without duplicating the quiz. And since a mishandled lead is expensive — we break it down in what an unqualified lead really costs — sorting the list properly often pays better than raising the ad budget.