Lead Scoring: How to Rank Leads by Buying Readiness

Your sales team has 60 open leads and time to seriously work maybe 15. Which 15? Most teams answer by gut feel, or by whoever came in most recently. Both miss money. The CFO who downloaded a pricing sheet sits in the same queue as a student writing a thesis, and a rep burns an afternoon on the student.

Lead scoring fixes the queue. You assign points based on who the lead is and what they do, then sort. The ones near the top get a call today. The ones near the bottom keep getting emails until they warm up or fall off. Done well, it means your best reps spend their hours on the people most likely to buy.

This is not a CRM feature you switch on and forget. It is a model you build, watch, and correct. Below is how to build one that reflects how your buyers actually behave, not a generic template that scores everyone into the same lukewarm middle.

Two questions every score has to answer

Buying readiness has two parts, and a good model keeps them separate.

Fit. Is this the kind of company and person you can actually sell to? A perfect-fit account that is not ready yet is worth more long-term than a poor-fit account that is ready today. Fit comes from explicit data: industry, company size, role, region, budget signals.

Intent. Is this person showing buying behavior right now? Intent comes from what they do: pages visited, content downloaded, emails opened, demo requested, pricing page viewed three times this week.

Score them on one axis and you get noise. A lead who opens every email might just be a curious competitor. A lead who matches your ideal profile perfectly might have landed on your site by accident and left. You want both signals high at once. That overlap, good fit plus active intent, is what "ready to buy" really means.

Keeping the two scores apart also tells you what to do, not just who to call. High fit and low intent means nurture and wait. High intent and low fit means qualify hard before a rep spends real time, because eagerness from the wrong-size buyer often ends in a deal that never closes or churns in three months.

Build your fit score from real customer data

Start with the accounts you already closed and kept. Pull your last 30 to 50 good customers and look for what they share. That pattern is your ideal customer profile, and it is the backbone of the fit score.

For a B2B service business, the attributes that usually matter:

Company size (headcount or revenue band)
Industry or vertical
Role and seniority of the contact
Region, if you only sell in certain markets
Tech or tooling signals, if relevant (they run a platform your product plugs into)

Assign points for matches and, this part gets skipped too often, negative points for disqualifiers. A free email domain on a "company" inquiry, a job title of "student" or "consultant" when you sell to in-house teams, a country you do not serve: subtract. Negative scoring stops obviously bad leads from floating up just because they were active.

A simple version might look like this. Numbers are illustrative; yours come from your own data.

Attribute	Condition	Points
Company size	50 to 500 employees	+20
Company size	Under 10 employees	-10
Role	Director or above	+15
Industry	Target vertical	+15
Email	Free webmail address	-15
Region	Outside served markets	-25

Fit data comes from forms, enrichment tools, and your CRM. Ask for the fields you actually score on. A form that collects job title and company size feeds a far better model than one that grabs only an email, which is part of why the lead magnets you use to collect data, and what you ask in exchange for it, deserve as much thought as the score itself.

Build your intent score from behavior

Intent is the moving part. It rises and falls week to week, so it needs a different logic than fit.

Weight actions by how close they sit to a buying decision. Reading a top-of-funnel blog post is weak intent. Viewing the pricing page, returning to the site three times in a week, or requesting a demo is strong intent. The closer the action is to "I am evaluating a purchase," the more points it earns.

A rough hierarchy, weak to strong:

Opened a marketing email: +1
Read a blog article: +2
Downloaded a guide or whitepaper: +5
Visited the pricing or product page: +10
Returned to the site 3+ times in 7 days: +10
Requested a demo or a quote: +25

Two refinements separate a useful intent score from a vanity counter.

Decay. Behavior gets stale. Someone who downloaded a guide eight months ago and went dark is not hot today. Decay the intent score over time, for example losing a percentage each week of inactivity, so the number reflects current interest, not a lifetime tally. Without decay, your highest scores drift toward your oldest, most checked-out contacts.

Frequency and recency beat volume. Three visits this week say more than thirty visits spread across last year. A lead returning again and again in a short window is in an active buying cycle. Reward the burst.

Intent scoring is where slow follow-up quietly kills deals. A lead who hits your pricing page is telling you the window is open now, and response time decides whether you walk through it or a competitor does. The score is only worth building if a hot signal triggers fast action.

Set thresholds and route the lead

A score is just a number until it changes what happens next. The point of the model is to trigger routing.

Set a threshold where a lead is "hot" enough to hand to sales. Above it, the lead becomes a sales-qualified lead and gets a call, fast. Below it, the lead stays in marketing for nurturing. This is the practical line between a marketing-qualified lead and a sales-ready one, and getting that handoff right is most of the value, which is why the MQL to SQL distinction is worth pinning down before you set numbers.

Routing rules to define up front:

The SQL threshold. Combined score, or both fit and intent crossing their own minimums. Many teams require fit above a floor before any intent score counts, so eager bad-fit leads never auto-route to a rep.
What happens at the threshold. A task in the CRM, a Slack alert, a round-robin assignment. The trigger should be automatic and immediate.
The nurture path for below-threshold leads. Good-fit, low-intent leads go into a sequence that keeps them warm until behavior lifts their score.

Resist one temptation: do not set the threshold so low that everything qualifies. If 80% of leads clear the bar, you have not scored anything, you have just relabeled the whole list. The threshold should match how many leads your team can genuinely work well in a week.

Test, then trust the model

Your first scoring model is a hypothesis. It will be wrong in places, and the only way to find out is to compare scores against outcomes.

After a quarter, run the numbers. Of the leads you scored "hot," what share became opportunities? What share closed? Then look at the deals that closed from leads you scored low: those are your model's blind spots, and they tell you which signal you under-weighted. Maybe a particular content download predicts purchase far better than its current points suggest. Maybe company size matters less than you assumed and role matters more.

Watch for these failure patterns:

Score inflation. Everyone creeps toward "hot" because you keep adding positive points and never subtract. Audit the negatives.
Activity bias. The model rewards email-openers and bot traffic over real buyers. Bot and low-quality clicks can inflate intent, so it helps to keep your scoring tied to genuine engagement and to watch where the traffic comes from.
Stale weights. Your market shifted, your scoring did not. Revisit weights at least twice a year.

Lead scoring also exposes a quality problem you might be blaming on the sales team. If your "hot" leads keep failing to close, the issue may sit upstream in your channels and how you generate leads, not in the score. The model is a mirror. It will show you where the funnel actually leaks.

When to keep it simple, and when to go further

Do not over-engineer this on day one. If you get fewer than 50 leads a month, a spreadsheet with five fit attributes and three intent actions will already beat sorting by hand. Build the elaborate model when volume makes manual triage impossible, not before.

The advanced path, predictive scoring, lets a machine learning model find patterns in your closed-won and closed-lost history and weight signals for you. It can catch correlations a human would miss. It also needs a meaningful volume of clean, outcome-labeled data to work, so most teams earn their way to it after a manual model has been running and proving its worth for a year or so.

Whatever the level, the discipline is the same: score on fit and intent, route on a real threshold, and check the score against closed revenue. A model nobody validates is just a number that makes the dashboard look busy.

FAQ

What is the difference between lead scoring and lead grading?

Grading usually refers to fit only (A, B, C, D by how well a lead matches your ideal profile), while scoring often refers to intent points from behavior. Many teams use both: a letter grade for fit and a number for intent. The combination is more useful than either alone.

How many points should "hot" be?

There is no universal number, because points are relative to your own scale. Set the threshold by working backward from capacity: if your team can work 20 leads a week well, tune the threshold so roughly that many clear it. Then adjust as you see which scores actually convert.

Do I need expensive software to start?

No. A CRM with basic automation, or even a well-built spreadsheet, is enough to test your first model. Dedicated marketing automation platforms make scoring and routing smoother once volume grows, but the model matters more than the tool. Prove the logic first, then buy software to scale it.

How often should I update the scoring model?

Review it at least twice a year, and after any major change to your offer, market, or lead sources. The trigger for an off-cycle review is simple: when reps start telling you the "hot" leads are not converting, the model is out of date.

Should negative scoring be part of it?

Yes, and skipping it is the most common mistake. Without subtractions, poor-fit but active leads (students, job seekers, competitors, wrong regions) climb the list and waste sales time. Negative points keep the ranking honest.

Can lead scoring work for a long, complex sales cycle?

It works especially well there, because long cycles produce many small signals over months. Decay matters more in long cycles: a buyer's intent six months ago tells you little about today. Score recent behavior heavily and let old activity fade.

Where to start this week

Lead scoring earns its keep when your best reps stop guessing and start working a ranked list. You do not need a platform or a data scientist to begin. You need to know who your good customers look like, which behaviors precede a purchase, and a threshold that respects your team's capacity.

A short checklist to get a first model live:

Pull your last 30 to 50 good customers and list the fit attributes they share
Assign fit points, including negative points for clear disqualifiers
Weight 4 to 6 behaviors by how close each sits to a buying decision
Add decay so old activity fades
Set one threshold that routes hot leads to sales automatically
Check scores against closed deals after a quarter and re-weight

If you would rather not build this from a blank page, that is a fair place to ask for help. Send us your lead data and current funnel, and we will map a scoring model to how your buyers actually behave, then show you where your funnel is leaking before you spend another month guessing which leads to call.