How to Score and Prioritize Sales Leads

Two failure modes account for almost all of it.

Measuring engagement only

Downloads, email opens, page views — all easy to capture, which is why they dominate. But a student researching a project and a buyer with budget both open emails. Engagement without fit ranks curiosity as highly as intent.

Too many factors

A model with fourteen weighted inputs cannot be explained, and a score nobody can explain is a score nobody trusts. When a salesperson disagrees with the ranking they will ignore it, and an ignored model is worse than none because it created work.

Three factors is enough for a small business, and three you actually use beats fourteen you do not.

Fit is how closely a lead resembles the customers you have served well and profitably. It is the most predictive factor and the least used, because working it out requires knowing your own book.

Start from your existing customers, not from a persona document. List your best ten — best meaning profitable and pleasant, not just largest — and find what they share:

• Size. Revenue band or headcount. Most businesses have a range where they are genuinely competitive and a range where they are not.
• Industry or job type. Where your experience compounds.
• Geography, if travel or licensing constrains you.
• The trigger. What was happening when they bought.

Not all engagement is equal. Weight by how much effort the action took them:

• High — replied to an email, booked or attended a meeting, asked about price, sent a specification, introduced a colleague.
• Medium — visited a pricing page more than once, downloaded something requiring a form, returned after a gap.
• Low — opened an email, single page view, social follow.

Email opens deserve almost no weight. Image proxies and privacy features make them unreliable, and a great many "opens" are not humans. Replies are the signal worth having, because replying costs something.

Recency matters as much as volume. Someone highly engaged four months ago and silent since is colder than someone who replied yesterday.

The hardest to capture and the most decisive. A perfect-fit, highly engaged lead with no reason to act this quarter will not act this quarter.

Timing signals worth recording when you hear them:

• An existing contract or supplier relationship ending.
• A regulatory or audit deadline.
• Something visibly broken — a failure, an incident, a bad audit.
• New funding, a new hire in a relevant role, or a stated project start.

Most of these only surface in conversation, which means the person who hears them has to record them. A field for "why now" on every lead is worth more than most automated scoring inputs.

A workable starting weighting for a small business is roughly fit 50%, engagement 30%, timing 20% — then adjust it against outcomes after a quarter.

Fit carries the most because a bad-fit lead that closes is often worse than one that does not: it consumes delivery capacity, produces a mediocre outcome, and generates no referral.

Work five, not fifty

Ranking the whole database produces a list nobody works. Take the top five each morning and work those. Five is small enough to actually complete, and completion is what makes the routine survive past the first busy week.

Each of the five needs one line of context — what they did, what to say. A ranked list without context still requires the salesperson to do the research, which is where the routine breaks down. Building that list daily is covered in how to build a daily sales call list.

As useful as positive scoring and almost never done. Some signals should actively push a lead down:

• Competitor or student email domains. Research, not buying.
• Outside your service area, where you cannot deliver.
• Wrong size band — too small to afford you, or large enough to need something you do not offer.
• Repeatedly engaged, never progressed. After enough cycles this is a pattern, not a prospect.
• Price-first contact. Not disqualifying, but predictive of a transaction rather than a relationship.

A contractor scoring three leads, using fit 50 / engagement 30 / timing 20. The point is not the arithmetic — it is that the ranking inverts what instinct would have chosen.

Lead A — the loud one

Downloaded two things, opened every email, visited pricing four times. Sole trader, well below the size band where you are competitive, no stated deadline.

Fit low, engagement high, timing low. Instinct says call first because they are visibly interested. Scoring says they are researching, and their budget will not survive contact with your quote.

Lead B — the quiet one

One email, replied within an hour with two specific questions. Right industry, right size, and mentioned their current supplier's contract ends in March.

Fit high, engagement moderate but high-effort, timing high. One reply is worth more than twenty opens because replying costs something, and the March date is the strongest signal on the page.

Lead C — the familiar one

Inquired three times over two years, always engages, has never bought. Right size, right industry.

Fit high, engagement high, but negative scoring applies. Repeated engagement with no progression is a pattern rather than a prospect. Keep them on a low-effort nurture and stop spending call time.

A scoring model is a hypothesis. Test it quarterly by asking one question: did the deals that closed actually score highly?

• If your best customers scored in the middle, your fit criteria are wrong.
• If high scorers never close, engagement is over-weighted relative to fit.
• If everything scores similarly, the model is not discriminating and needs sharper criteria rather than more of them.

A model reviewed twice is better than a sophisticated one built once and never checked.

Keep the review cheap enough that it actually happens. Twenty minutes with the last quarter's closed-won deals, checking where each one sat in the ranking at the time you first contacted them, is enough to tell you whether the weighting is roughly right or badly wrong.

Defining fit is the real work and it is a one-off — a couple of hours with your customer list and margin data.

The daily part should be ten minutes. If it is longer, the list is too long or it lacks the context line, and either way it will stop happening within a fortnight.

Scoring implies a top five, which implies forty-five leads you are not calling today. What happens to them decides whether scoring helps or simply narrows your funnel.

Nurture, on a schedule

Good-fit leads with poor timing are the most valuable group in the pipeline and the most neglected. Their timing changes — contracts end, people get promoted, something breaks — and the business they remember when it does is the one that stayed in contact without pestering.

A monthly touch that is genuinely useful beats a weekly one that is not. Send the thing you would send a peer.

Re-score, do not re-rank manually

Scores go stale. A lead that scored low in January because nothing was forcing a decision may score high in April. If scoring only happens once at entry, you have built a filter rather than a prioritization system, and everything that arrives at a bad moment is lost permanently.

Disqualify honestly

Some leads should leave the pipeline. Wrong size, wrong geography, needs something you do not do. Marking them disqualified with a reason is more useful than leaving them to drift, because the reasons aggregate — if a third of your leads are disqualified for being too small, that is a marketing targeting problem showing up as a sales problem.

Defining what a good customer looks like is judgment and it is yours. Applying it every morning across every open lead, checking recent activity, and writing the context line is not — it is assembly, and it has to happen before 8am to be useful.

Which is precisely why it stops happening in the weeks the pipeline is busiest.