Why lead scoring exists
Without scoring, your reps treat every lead the same. They spend an hour qualifying a tire-kicker while a hot lead waits 48 hours and goes cold. Scoring just routes attention where it pays.
The two signal categories
Every lead score combines fit (do they look like our customer?) and intent (are they showing buying signals?).
Fit signals (demographic / firmographic)
- Job title / seniority
- Company size (employees, revenue)
- Industry / vertical
- Geography (target markets vs not)
- Tech stack (uses competitor or complement)
Intent signals (behavioral)
- Visited pricing page
- Downloaded a comparison guide
- Watched 50%+ of a demo video
- Replied to an email or SMS
- Requested a callback
- Multiple visits in 7 days
- Visited from a high-intent source (Google search vs random social)
Build a scoring model in 30 minutes
Open a spreadsheet. List every signal. Assign each a weight from 1-25. Total possible = 100. Example for a B2B SaaS:
| Signal | Weight |
| Title contains "VP", "Director", "Head of" | +15 |
| Title contains "Manager" | +8 |
| Title contains "intern", "student" | −10 |
| Company 50-500 employees | +12 |
| Company <10 employees | −5 |
| Industry = our top vertical | +10 |
| Visited pricing page | +15 |
| Watched demo video | +12 |
| Replied to email/SMS | +20 |
| Visited 3+ times in 7 days | +8 |
| Email is gmail/yahoo/hotmail | −5 |
| Phone is on DNC list | −25 (effectively disqualify) |
Sum the score for every lead. Bucket them.
Routing logic by score bucket
- 80+ (hot) — call within 5 minutes. Senior rep. Bypass any qualification queue.
- 50-79 (warm) — call within 24 hours. SDR sequences.
- 30-49 (lukewarm) — into a 5-touch nurture sequence (email + SMS + WA). No calls.
- 0-29 (cold) — into the long-tail nurture or just don't contact.
- <0 (disqualified) — drop. Don't even nurture.
Disqualification rules (the underrated half of scoring)
Auto-disqualify is more valuable than auto-prioritize. Examples:
- Email domain is competitor (no demo for them)
- Country isn't in our serviceable list
- Phone is on national DNC
- Company too small to ever afford the product
- Filled out form with obvious junk ("aaa", "test", "asdf")
Disqualifying 30% of inbound up front frees reps to spend 3x more time on the remaining 70%.
Iterating with conversion data
Your initial weights are guesses. After 100 closed-won deals, run the numbers:
- Pull every closed-won lead's source signals
- Compare which signals are over-represented in won deals vs lost deals
- Increase weights for over-represented signals
- Decrease weights for irrelevant ones
- Repeat quarterly
Tools (you don't need a fancy one)
- Google Sheets / Airtable — perfectly fine for <10K leads/month
- HubSpot built-in scoring — comes with the platform
- Salesforce Einstein scoring — ML-based, needs 1,000+ closed deals to be useful
- Madkudu / Breadcrumbs — dedicated scoring platforms for high-volume teams
Start in a spreadsheet. Move to a tool when manual updates become painful (~5K leads/month).
Channel routing by score
The hottest leads deserve real-time human contact. Lukewarm leads should be touched by automation. Match channel to score:
- 80+: Phone call within 5 min
- 60-79: SMS within 1 hour, then call within 24 hours if reply
- 40-59: WhatsApp message + email sequence
- 20-39: Email-only nurture for 90 days
Where the data comes from
Lead scoring is only as good as the data you have on each lead. Phone, email, WhatsApp activity status, country, company — collect what you need. Verified country-targeted leads →
The trap to avoid
Don't over-engineer. A 10-signal model that you actually maintain beats a 50-signal "perfect" model that drifts out of date in 3 months. Start small, iterate based on what closes.