How to Build an Intent-Scoring Engine for 20% Higher Conversion

A practical starting framework:

• 0–30: Do not contact. Nurture passively with content.
• 31–55: Low intent. Monitor for signal changes, do not reach out yet.
• 56–79: Medium intent. Add to a lower-priority sequence with longer delays.
• 80+: High intent. Prioritize immediately, contact within 24–48 hours of score crossing threshold.

Your thresholds will shift as you calibrate against actual conversion data. Run the model for 30 days before changing the weights based on gut — you need data, not intuition, to optimize a scoring system.

Building the Data Layer

The scoring model is only as good as the data feeding it. Here is how to construct the data layer from cheapest to most sophisticated:

Level 1 — Manual + Spreadsheet (day one, zero budget). Your CRM or a Google Sheet. Manually log signals for each prospect. Score by hand. This is not scalable, but it forces you to understand which signals actually matter before you automate.

Level 2 — Clay + Apollo/LinkedIn. Clay is the highest-leverage tool for intent scoring at the growth stage. It pulls firmographic and technographic data, enriches contact records with 50+ data sources, and lets you build scoring logic with no-code formulas. Apollo provides contact data and some behavioral signals. At this level you can build a functioning automated scoring pipeline for under $300/month.

Level 3 — Bombora + HubSpot/Salesforce. Bombora tracks B2B intent signals across thousands of websites and surfaces companies showing elevated research activity on topics relevant to your offering. This is third-party behavioral intent data at scale. Expensive ($1,500–$3,000/month) but high signal. Worthwhile once you are past the early traction stage and have validated your ICP.

Level 4 — Custom infrastructure. Proprietary scrapers monitoring job boards, LinkedIn, company websites, and news sources. Enrichment APIs (Clearbit, PeopleDataLabs) for firmographic depth. A scoring engine built in Python or Node that runs on a cron job and pushes high-intent leads directly to your sales team’s queue. This is what mature demand generation operations look like.

Most operators should be operating at Level 2 within their first 90 days. Level 3 comes when you are ready to scale outreach volume significantly and need predictive signal coverage.

The Decay Function

Intent signals expire. A prospect who visited your pricing page 6 months ago is not a high-intent lead today. Build decay into your model. Every signal should have a half-life:

• Pricing page visit: high decay, 50% weight loss after 14 days
• Funding event: medium decay, relevant for 6–9 months
• New executive hire: medium decay, relevant for 3–6 months
• Email open: high decay, largely irrelevant after 30 days
• Tech stack adoption: low decay, remains relevant until the tool is replaced

This prevents your score from being dominated by historical data and keeps high-intent leads surfacing based on current behavior.

Calibration and Iteration

After 60 days of running your scoring model, pull the data and run this analysis:

• Of all leads who scored 80+, what percentage converted to a sales conversation?
• Of all leads who scored 50–79, what percentage converted?
• Which individual signals are most correlated with conversion in your data?

Use these answers to reweight your model. The signals with the highest correlation to actual conversion should have the highest point values. Remove signals that are consistently present in low-converting accounts — they are adding noise, not signal.

A well-calibrated intent scoring engine improves outreach-to-meeting conversion by 15–30% compared to unscored outreach to the same contact list. The math compounds: the same number of outreach messages, directed at higher-intent targets, produces materially more pipeline.

What to Do When a Lead Scores High

Speed matters. When a prospect crosses your high-intent threshold, contact them within 24 hours. Research shows that lead response time is one of the highest predictors of conversion — not because urgency is inherently persuasive, but because a high-intent buyer is often evaluating multiple options simultaneously. The first credible outreach they receive has a disproportionate advantage.

Your outreach message should reference the specific signal that triggered the alert. Not generically — specifically. “I noticed you recently posted about [specific problem]” performs far better than a generic opening. The score tells you who to contact; the signal tells you what to say.

Ready to put your intent engine to work? Read the personalized outreach at scale guide to build the sequence architecture on top of your scoring layer.