Lead Scoring: Predicting Who Will Actually Buy

Your sales team has 200 leads in the pipeline.

They have time to call 20.

Which 20 should they call?

Lead Scoring is the practice of predicting which leads are most likely to buy so your sales team focuses on them.

How Lead Scoring Works

You assign points to behaviors and characteristics.

Higher score = more likely to buy.

Example scoring system:

Company size (static data):

Company size 1-10: 0 points
Company size 11-50: 5 points
Company size 51-200: 10 points
Company size 201+: 15 points

Page views (behavioral):

1-2 page views: 0 points
3-5 page views: 5 points
6-10 page views: 10 points
11+ page views: 15 points

Pricing page visit (behavioral):

Did not visit: 0 points
Visited once: 10 points
Visited 2+ times: 20 points

Email engagement (behavioral):

No opens: 0 points
Opened 1-2 times: 5 points
Opened 3+ times: 10 points

Time since signup (temporal):

Signed up 30+ days ago: 0 points
Signed up 14-29 days ago: 5 points
Signed up 7-13 days ago: 10 points
Signed up 0-6 days ago: 15 points (fresh leads are most likely to be in buying mindset)

A lead gets scored:

Company size 50: 10 points
7 page views: 10 points
Visited pricing page twice: 20 points
Opened emails 2 times: 5 points
Signed up 10 days ago: 10 points
Total score: 55 points

If your threshold for “sales-ready” is 50+ points, this lead gets called.

Two Types of Lead Scoring

Type 1: Demographic/Firmographic Scoring

Scores based on who they are:

Company size
Industry
Geography
Revenue size
Role/title

This answers: “Are they the type of customer who buys our product?”

Example: You sell to healthcare companies. A healthcare company gets higher score than a fashion company.

Type 2: Behavioral Scoring

Scores based on what they do:

Pages visited
Time on site
Email opens/clicks
Pricing page visits
Demo requests
Content downloads

This answers: “Are they actively interested in buying right now?”

Example: Someone who visited your pricing page and clicked a pricing comparison email is more likely to buy than someone who just read a blog post.

The best lead scoring uses both.

Building Your Lead Scoring Model

Step 1: Define “converted” customers

Look at your best customers. What did they do before they became customers?

How many pages did they view before converting?
How long were they active before they bought?
What pages did they visit?
Did they click certain emails?

Example findings:

90% of customers visited the pricing page
80% visited the ROI calculator
75% were active for at least 7 days before purchasing
Average score of customers: 65+ points

Step 2: Look at non-converted leads

What did people who didn’t convert do?

Example findings:

50% never visited pricing page
30% visited pricing page but bounced immediately
60% were active for less than 3 days
Average score of non-converts: 20 points

Step 3: Set a threshold

If converted customers score 65+ and non-converts score 20, your threshold might be 50.

Anyone above 50 is “sales-ready.” Below 50 needs more nurturing.

Step 4: Track and iterate

Once you implement, track:

Of the people you called (score 50+), how many converted?
Of the people you didn’t call (score <50), how many converted on their own?

Adjust thresholds based on reality.

If you’re calling people below 50 and they convert a lot, lower the threshold.

If you’re calling people above 50 and they don’t convert, raise the threshold or change the scoring model.

Common Lead Scoring Mistakes

Mistake 1: Giving too much weight to demographic data

A fortune 500 company visits your site but bounces immediately. They score high on size, low on behavior.

Your sales team calls them. They’re not interested.

Meanwhile, a 50-person company is deeply engaged (visited 15 pages, opened 10 emails) but scores low because they’re small.

Lesson: Behavioral signals are stronger than demographic signals. Weight them heavier.

Mistake 2: Not accounting for time decay

Someone signed up 6 months ago and hasn’t been active. They score the same as someone who signed up yesterday.

But the person from 6 months ago has likely moved on or lost interest.

Better approach: Recent activity scores higher than old activity.

Mistake 3: Using the wrong conversion definition

You might define “converted” as “became a customer” (signed a contract).

But maybe you should define it as “scheduled a demo” or “got to proposal stage.”

If your sales cycle is long, you want to predict “who will schedule a demo next week” not “who will buy in 6 months.”

Mistake 4: Ignoring negative signals

Some behaviors indicate someone will NOT convert.

Visited pricing page once and left (didn’t come back)
Unsubscribed from emails
Requested demo but didn’t show up

These should reduce score, not stay neutral.

Negative Scoring

Add negative points for bad signals:

Negative behaviors:

Unsubscribed from emails: -20 points
Marked email as spam: -30 points
Visited pricing page but never clicked anything: -10 points
Bounced from demo: -15 points
Requested feature you don’t offer: -5 points

A lead could score high on positive signals but be brought down by negative signals.

Example:

Positive signals: 60 points
Unsubscribed from emails: -20 points
Net score: 40 points (below threshold, don’t call)

The Distinction: Lead Scoring vs. Lead Grading

Lead Scoring predicts: “Are they ready to buy now?”

Lead Grading determines: “Are they a good fit for us?”

Lead grading happens first (is this even the right customer type?) and lead scoring happens second (are they ready now?).

Example:

Lead A:

Fortune 500 company (great lead grade)
But they visited 2 pages and left 3 months ago (low lead score)
Action: Nurture with content, not sales call

Lead B:

50-person startup (okay lead grade)
Visited 15 pages, opened 8 emails, visited pricing 3 times (high lead score)
Action: Sales call immediately

Without separating these, you’d prioritize A (bigger company) when you should prioritize B (higher buying intent).

Automating Lead Scoring

In practice, you use a tool (CRM, marketing automation, or dedicated lead scoring platform) to automatically assign scores.

Every time someone visits a page, opens an email, or fills out a form, their score updates.

Your CRM then automatically alerts your sales team: “This lead just hit 50 points. Call them.”

The automation removes manual work. Leads don’t fall through the cracks because you forgot about them.

The Real Question Lead Scoring Answers

Lead scoring isn’t perfectly predictive. Some people will surprise you.

The real question is: “Where should we focus our limited sales time?”

If your sales team can call 20 leads per week and you have 200 leads, you need some way to prioritize.

Lead scoring answers that. It says: “These 20 are most likely to buy. Start here.”

This lets your small sales team focus on high-probability opportunities instead of calling randomly.

The Takeaway

Lead scoring predicts buying intent based on behavior and demographics.

Build your model on your actual customers: What did they do before they converted?

Use behavioral signals more heavily than demographic signals.

Implement negative scoring for bad signals.

Track whether your scoring is actually predictive. Adjust constantly.

We help you build and refine your lead scoring model so your sales team focuses on leads most likely to buy.