Lead Scoring for a Ten-Person Sales Team

Trelio Systems sells a warehouse management tool out of Bengaluru with ten people on the sales floor. Their CRM had lead scoring switched on. Nobody used it. When we asked the two best reps how they picked who to call, both described the same mental shortcut, and neither of them mentioned the number on the screen. One said "I call anyone from a company with more than one warehouse who filled the form during work hours." That sentence was a better scoring model than the 43-rule contraption their previous consultant had installed, and it took nine words. Good lead scoring for small business usually looks like that: short, obvious in hindsight, and derived from what the team already knows.

Why most small-team lead scoring dies

Three causes, in roughly this order of frequency.

First, complexity nobody owns. A model with 40 rules needs somebody to check whether those rules still hold. In a ten-person team, that person does not exist, so the model ages into fiction within two quarters. Second, scoring behaviour that no longer happens. Half the rules in a typical inherited model reference an email nurture sequence that was switched off last year, so leads accumulate points for events that cannot occur. Third, and most damaging, a score with no consequence. If a lead scoring 92 gets handled exactly like a lead scoring 31, then the score is decoration, and reps correctly ignore it.

The two-axis lead scoring model

Score fit and intent separately. Do not add them into one number, because a single number hides the difference between "perfect customer, just browsing" and "wrong customer, very keen", and those two need completely different treatment.

Lead scoring on two axes stays legible because each axis means one thing. Fit is about the account: size, sector, geography, whether they have the problem you solve, whether they can pay your minimum. Intent is about the behaviour: what they did, how recently, and how much effort it took them. Score each out of five. A ten-person team can hold that in their heads, which is the entire design goal.

Finding your real lead scoring signals

Here is the part most teams skip, and it is the only part that makes the model yours. Pull your last 40 closed-won deals and your last 40 closed-lost, and look for what separated them at the point of first contact. Not at the point of close, which is contaminated by everything your reps did afterwards.

Trelio's exercise produced five rules. Multiple warehouse locations, submitted during business hours, came from a pricing or integration page, gave a company email, and asked about a specific integration in the message field. That last one turned out to be their strongest predictor by a distance, and it had never been in the old model because nobody had read the message field systematically.

Turning lead scoring into a routing rule

Lead scoring exists to answer one question: who gets a human, how fast. Write the consequence table before you write the rules.

Speed in the top quadrant is the single highest-return change most small teams can make. The difference between calling a keen, well-matched lead in ten minutes and calling them the next morning is not marginal. Whoever calls first usually sets the terms of the whole conversation.

Run lead scoring by hand for four weeks

Before any automation, do the lead scoring for every inbound enquiry manually on a shared sheet. Two columns, fit and intent, filled in by whoever triages that day. It takes about fifteen seconds per lead and it does three things no software can do at this stage.

• It surfaces the rules that are ambiguous in practice, because two people will score the same lead differently and then have to argue about it.
• It builds trust, since the reps who defined the rules will believe the output.
• It tells you the volume in each quadrant, which determines whether automation is even worth building. If you get six leads a day, a sheet may be the permanent answer.

Trelio ran the manual sheet for five weeks and found that their top quadrant held about four leads a day. Four. All the tooling anxiety had been about a problem that fit comfortably on a whiteboard, and their real bottleneck was that nobody was answering the phone after 6pm.

Automating without overbuilding

Once the rules are stable, encode them where the leads already land. Most CRMs will do simple attribute scoring natively. For anything beyond that, a small automation workflow that reads the form submission, applies the eight rules, sets two fields and posts the hot ones into a sales channel is usually a day of work.

Resist three temptations. Do not add score decay in the first version, because it makes the number hard to explain and reps stop trusting numbers they cannot explain. Do not buy intent data before your own first-party signals are working. And do not connect a model to automatic emails that fire without a human ever seeing the lead, at least not until the false-positive rate is known.

Keeping lead scoring honest

Put a recurring 45-minute review in the calendar every quarter with one agenda item: of the deals we won last quarter, how many were scored good-fit high-intent at first contact? If the answer drops below roughly half, your market has moved and the rules need editing. That is the whole maintenance ritual.

An honest limitation worth stating. Lead scoring does not create demand and it will not rescue a weak pipeline. If you get twelve leads a month, scoring them is theatre, and your time is better spent on the top of the funnel. Lead scoring starts to pay when volume exceeds the team's capacity to call everyone properly, and not a day before.

The best model your team will ever use is the one they can recite from memory on a call. Keep it to eight rules, tie it to a promise about response time, and let the wins and losses edit it every quarter. Done that way, lead scoring for a small business stops being a CRM feature nobody opens and becomes the thing that decides who your ten people call next.