Product-Led Lead Scoring Models: Score Leads by Usage Data

Quick answer: A product-led lead scoring model ranks prospects by combining traditional signals (company size, role, email engagement) with product usage data (features used, frequency, team invites). For PLG companies, product usage typically carries 60-70% of the total score because actual behavior predicts conversion better than demographics.

• Traditional lead scoring - Demographics + email opens. Misses the main signal: product usage.
• Product-led scoring - Adds feature adoption, usage frequency, team size, power user behaviors to the model.
• Implementation challenge - Product data lives in your app, not HubSpot. Getting it into HubSpot requires either a data warehouse + reverse ETL ($500+/mo + engineering) or a direct sync tool like Zoody ($149/mo, no engineering).
• Best practice - Weight product signals at 60-70% of the total score. A freemium user who invites 5 teammates and uses your core feature daily is a stronger lead than a VP who downloaded one whitepaper.

What is a Lead Scoring Model?

A lead scoring model is a systematic approach to ranking prospects based on how likely they are to convert. You assign point values to specific attributes and behaviors, add them up, and use the total score to decide who sales should talk to first.

Traditional models use demographic data (company size, industry, role, location) and engagement signals (email opens, content downloads, website visits, webinar attendance). A VP at a 500-person company who attended your webinar gets more points than an intern at a 10-person startup who clicked one link.

The problem: traditional models miss the most important signal for PLG companies. Actual product usage.

A freemium user who invites their team, connects an integration, and uses your product daily is more likely to convert than a VP who downloaded a PDF. But traditional lead scoring treats the VP as the better lead because of their title and company size.

Product-led lead scoring fixes this by incorporating how prospects interact with your product, not just your marketing content. You track feature adoption, usage frequency, team growth, and power user behaviors, then use those signals to score leads alongside (or instead of) demographic data.

Traditional vs. Product-Led Lead Scoring

Traditional lead scoring optimizes for intent signals outside the product: form fills, ad clicks, demo requests. It works when you have a sales-led motion where prospects research before they buy.

Product-led scoring optimizes for intent signals inside the product. You watch what users do when they have hands-on access. This works better for PLG because the product is the primary evaluation tool.

Aspect	Traditional Scoring	Product-Led Scoring
Primary signals	Demographics, engagement	Product usage, feature adoption
Data source	Marketing automation, website	Product database, event tracking
Score weight	Job title, company size carry most weight	Usage depth and frequency carry most weight
Best for	Sales-led, long evaluation cycles	Product-led, hands-on trials
Sales readiness	MQL based on form fills	PQL based on usage milestones

Why Product Usage Data Matters for Qualification

Product usage predicts conversion better than job titles because it shows actual fit, not theoretical fit.

A user who logs in daily, invites teammates, and uses advanced features has verified that your product solves their problem. They've already adopted it. Sales conversations become expansion discussions, not cold pitches.

A VP with the right title who hasn't touched the product is just a name in your CRM. You have no idea if they care, if the product fits, or if they have budget.

PLG companies that score leads purely on demographics waste sales time on cold outreach to titles while warm, active users sit in the queue unnoticed. Product-led growth works because it lets prospects self-qualify through usage, then routes the qualified ones to sales.

How to Build a Product-Led Lead Scoring Model

Start with your ideal customer profile, but define it using product usage patterns in addition to firmographics. Look at your last 50 customers and ask: what did they do in the product before they bought?

Step 1: Identify your ICP using both firmographic and product usage patterns. Company size matters, but so does whether they invited teammates, connected an integration, or used your core workflow five times in the first week. Build two profiles: the company you want to sell to, and the usage pattern that predicts conversion.

Step 2: Map the buyer journey including product trial/freemium behaviors. Most PLG journeys include: sign up, activate (complete first task), adopt (regular usage), expand (invite team or upgrade tier), convert. Each stage has product signals you can score.

Step 3: Assign point values to traditional signals. Keep the basics: company size (+10 for 50-500 employees, +5 for 10-50), role (+15 for decision maker titles, +5 for influencer titles), industry fit (+10 if target industry). Add engagement: +5 for email open, +10 for website visit, +15 for content download.

Step 4: Assign point values to product usage signals. Feature adoption: +20 for integration setup, +15 for team invite, +10 for advanced feature use. Frequency: +15 for 5+ sessions in 7 days, +10 for daily active streak of 3+ days. Depth: +10 for creating 10+ records, +15 for custom configuration, +20 for API usage.

Step 5: Weight product signals appropriately. This is where most teams get it wrong. If you give 80% of the score weight to demographics and 20% to product usage, you're still optimizing for titles, not behavior. For PLG companies, flip it: product usage should carry 60-70% of the total possible score. A power user at a small company should outscore a VP who signed up but never logged in.

Step 6: Set threshold scores for MQL, PQL, and SQL stages. Example: MQL = 30 points (basic fit + minimal engagement), PQL = 60 points (strong fit + meaningful product usage), SQL = 80 points (strong fit + power user behaviors + buying signals like pricing page visit or team expansion).

Step 7: Test, measure, and iterate. Track conversion rates by score range. If your 60-80 point leads convert at 15% but your 80-100 point leads only convert at 12%, your high score threshold is mis-calibrated. Adjust weights quarterly based on actual closed/won data.

Framework: Weighting Product Signals vs. Traditional Signals

Most successful PLG scoring models use a 60/40 or 70/30 split: product usage carries the majority of the score, traditional signals fill in gaps.

A typical breakdown:

• Product usage: 60-70 points max (feature adoption, frequency, depth, collaboration)
• Demographics: 15-20 points max (company size, role, industry fit)
• Engagement: 10-15 points max (email opens, content downloads, webinar attendance)

This means a freemium user who hits all your product usage milestones can score 70 points before you even know their job title. A VP who downloaded a whitepaper but never logged into the product caps out at 35 points.

The key: build your scoring model around the hypothesis that people who use your product are more likely to buy it than people who read about it.

Recommended Scoring Thresholds and Stages

Your thresholds depend on your sales capacity and close rates, but here's a starting framework based on what works for PLG companies running HubSpot as their CRM:

Stage	Score Range	What It Means	Sales Action
Raw lead	0-29	Signed up, minimal activity	Automated onboarding emails only
MQL	30-49	Basic fit + some engagement	Monitor, nurture via email
PQL	50-79	Good fit + meaningful usage	Sales touches (email, LinkedIn)
Hot PQL	80+	Strong fit + power user signals	Immediate outreach, prioritize

Adjust these ranges based on your data. If your PQLs are converting at 25% but your hot PQLs only convert at 20%, you're mis-weighting something at the top end.

Product Usage Scoring Criteria and Examples

The signals that predict conversion vary by product type, but these categories apply to most PLG companies.

Feature adoption signals show buying intent when users engage with features that imply scale, team usage, or integration with their existing stack. A user who connects your Slack integration or HubSpot sync is signaling that they're embedding your tool into their workflow. A user who stays in the free tier features might just be experimenting.

High-intent features to track:

• Integration setup: +20 points (they're committing to your tool as part of their stack)
• Team invite: +15 per additional user (buying decisions are made by teams, not individuals)
• Admin or settings configuration: +10 (they're treating this like a long-term tool, not a trial)
• Advanced feature use: +15 (they've graduated past basic capabilities)
• Workflow or automation setup: +20 (they're building process around your product)

Usage frequency and consistency separate tire-kickers from real prospects. Daily active users convert at 3-5x the rate of weekly users in most PLG products.

Frequency signals:

• Daily active for 5+ consecutive days: +20 points
• 10+ sessions in a 7-day window: +15 points
• Weekly active for 4+ consecutive weeks: +10 points
• Single session, then dormant: -10 points (negative scoring matters too)

Usage depth shows how much value they're extracting. Someone who creates 100 records or 10 projects has a much higher switching cost than someone who created 3 records and stopped.

Depth signals:

• 50+ records/projects created: +15 points
• 10+ custom fields or configurations: +10 points
• Data imported from another tool: +20 points (they're migrating, not experimenting)
• 5+ different feature areas used: +10 points

Power user behaviors predict high lifetime value and faster conversion. These users understand your product and are already treating it like a paid tool.

Power user signals:

• API key generated: +25 points (developers involved, they're building on your platform)
• Custom reporting or dashboard created: +15 points
• Zapier/automation connected: +20 points
• Attended a live training or asked a technical question in chat: +10 points

Collaborative signals show team buy-in. A single user can churn. A team of 5 users across 3 departments has organizational momentum.

Collaboration signals:

• 3+ active team members: +20 points
• 5+ active team members: +30 points
• Users from 2+ departments (based on email domains or roles): +15 points
• Shared workspace or project created: +10 points

Time-based signals show sustained interest. A user who's been active for 30 days is a better lead than a user who went hard for 3 days then disappeared.

Time signals:

• Account age 14+ days with continued activity: +10 points
• Account age 30+ days with continued activity: +15 points
• Usage streak of 10+ days: +20 points
• Account created 7+ days ago but no login in last 5 days: -15 points (decay score)

High-Intent Product Signals to Track

Not all product usage is equal. Prioritize signals that show the user is embedding your product into their workflow, not just experimenting.

Top 5 signals that predict conversion across most PLG products:

1. Integration or API connection - They're making your tool part of their stack.
2. Team invite or multi-user activity - Buying decisions are collective.
3. Data import from another tool - They're migrating, not trying.
4. Daily usage for 7+ days - Habit formation predicts retention.
5. Usage of a paid-tier-only feature in trial - They've tested the thing they'd pay for.

Track these as boolean properties in HubSpot (integration_connected: true/false, team_size: number, days_active_last_7: number) and score off them.

Example Product-Led Scoring Matrix

Here's a real scoring matrix for a hypothetical project management tool. Total possible score: 100 points. PQL threshold: 60 points.

Product usage (70 points max):

• Integration connected (Slack, HubSpot, etc.): +20
• Team members invited: +5 per user (max +25)
• Projects created: +2 per project (max +10)
• Daily active 5+ days in last 7: +15
• Custom workflow or automation created: +10
• Account age 14+ days with continued usage: +10

Demographics (20 points max):

• Company size 50-500 employees: +10
• Company size 500+ employees: +15
• Decision maker title (VP, Director, Head of): +10
• Manager or team lead title: +5

Engagement (10 points max):

• Pricing page visit: +5
• Help doc or template accessed: +3
• Email open (onboarding series): +2

A user who invites 3 teammates (+15), connects Slack (+20), creates 5 projects (+10), and uses the tool daily for a week (+15) scores 60 points before you know their job title. That's your PQL threshold. Sales gets notified.

A VP at a 200-person company (+10 + +10 = 20) who signed up but never came back scores 20. They stay in nurture.

Implementing Product-Led Scoring in HubSpot

The hard part: your product usage data lives in your app's database, not in HubSpot. HubSpot's scoring tool can only score properties that exist on the contact record.

The Data Sync Challenge for RevOps Teams

Most RevOps teams hit this wall: we want to score by product usage, but we can't get the data into HubSpot without engineering help.

Traditional approach requires:

1. A data warehouse (Snowflake, BigQuery, Redshift) to centralize product event data
2. dbt or a transformation layer to aggregate events into user-level metrics
3. A reverse ETL tool (Hightouch, Census) to sync warehouse tables into HubSpot custom properties
4. Engineering time to set up and maintain all three layers

Total cost: $500-$1,500/mo for the tools, 20-40 hours for initial setup, ongoing engineering maintenance when something breaks. Most teams under $10M ARR don't have the resources.

The alternative some teams try: HubSpot's Operations Hub custom code workflows. You can write Node.js code that calls your product database or API, then updates HubSpot properties. This works for small-scale projects but hits rate limits fast, requires ongoing code maintenance, and still needs an engineer to write and deploy it.

What you actually need: product events flowing into HubSpot in real time as custom properties, without setting up a data stack.

Syncing Product Data to HubSpot with Zoody

Zoody syncs product usage data into HubSpot without requiring a data warehouse or engineering work. You send product events from your app (user invited teammate, user created project, user connected integration), and Zoody writes them as properties and timeline events on the HubSpot contact and company records in real time.

Setup:

1. Install Zoody from the HubSpot marketplace, authorize the connection.
2. Add Zoody's JavaScript snippet to your app, or send events via API (any language, standard REST calls).
3. Define which events and properties to track (feature usage, team size, last active date, usage frequency).
4. Zoody maps those events to HubSpot custom properties automatically.

Example: you send zoody.track('integration_connected', { integration_type: 'slack' }) when a user connects Slack. Zoody creates a integration_connected boolean property on the contact, sets it to true, and adds a timeline event. You can now score off that property in HubSpot.

Once the data is in HubSpot, you have two options for scoring:

Option 1: HubSpot's native Predictive Lead Scoring tool (Professional/Enterprise only). This uses HubSpot's built-in scoring model, which can incorporate custom properties. You define which properties to include, HubSpot assigns weights automatically based on historical conversion data. Downside: less control over exact point values.

Option 2: Custom calculated properties or workflows. Create a "PQL Score" number property, then either:

• Use a calculated property (formula-based, real-time) to sum up points based on property values. Example formula: (integration_connected * 20) + (team_size * 5) + (days_active_last_7 * 3). This updates instantly whenever a property changes.
• Use a workflow that increments the score property when specific events happen. More flexible, but workflows lag behind real-time events and have execution limits.

Most PLG teams use calculated properties for simplicity and real-time updates. The setup is covered in detail here, including how to map product events to HubSpot properties and build the scoring formula.

Once the score exists on the contact record, you can:

• Segment lists by score range (PQL score 60-79, hot PQL score 80+)
• Create workflows that notify sales when a contact crosses the PQL threshold
• Add score to contact views and report on score distribution
• Track how score correlates with closed/won deals

Example workflow: When PQL score changes to 60 or above, create a task for the account owner, send a Slack notification to the sales channel, and move the lifecycle stage to "Product Qualified Lead". This is the same pattern used to automate PLG sales handoff.

Zoody costs $149/mo (flat rate, unlimited contacts). Compare that to Hightouch ($350+/mo) + warehouse ($200+/mo) + engineering time ($5,000+ to set up). For most teams under $20M ARR, direct product-to-HubSpot sync makes more sense than building a data stack.

Optimizing and Scaling Your Lead Scoring Model

Launch your model, then measure conversion rates by score range every month. If your 60-80 point leads are converting at 20% but your 40-60 point leads are converting at 18%, your threshold is too high or your weighting is off.

Monitor conversion rates by score range to validate your model. Pull a report: how many leads in each score bucket (0-29, 30-49, 50-79, 80+) converted to closed/won in the last 90 days? The conversion rate should increase as score increases. If it doesn't, your model is broken.

A/B test different weighting approaches. Run two scoring models for 60 days: one that weights product usage at 60%, one at 70%. Compare which bucket of leads converts better. Use the data to tune your weights.

Update scoring criteria as your product and ICP evolves. The features that predicted conversion in year one might not predict it in year three. Review your model quarterly. Look at the last 50 closed/won deals: what did they do in the product before they bought? Adjust your scoring criteria to match current patterns.

Segment scoring models by use case, company size, or industry when appropriate. Enterprise customers might need higher team size and admin usage scores. SMB customers might score higher on self-serve adoption and quick time-to-value. If your conversion patterns vary significantly across segments, build separate models.

Train sales on what scores mean and how to prioritize follow-up. A score is useless if sales doesn't trust it. Walk the team through the model, explain why product usage matters, and show them the conversion data. Make it clear that a high-scoring freemium user is a better use of time than a low-scoring VP.

Use score decay to downgrade inactive users over time. A user who scored 70 points three months ago but hasn't logged in since is not a hot lead. Add negative scoring for inactivity: -5 points for no login in last 14 days, -10 points for no login in last 30 days. This keeps your PQL pipeline current.

Common pitfalls to avoid:

• Over-complicating the model with 50+ scoring rules. Start simple (10-15 rules), add more only if conversion data justifies it.
• Ignoring negative scoring. Inactivity, unsubscribes, and churned users should lose points.
• Setting thresholds too high. If only 2% of your leads hit "hot PQL", you're starving sales. Calibrate thresholds so 10-20% of active users qualify.
• Failing to iterate. The first version of your model will be wrong. Plan to adjust it every quarter based on actual conversion data.

Understanding the difference between lead scoring and PQL scoring helps you decide when to use each approach and how to layer them in HubSpot.

FAQ

How to build a lead scoring model?

Start by defining your ideal customer profile using both demographic data (company size, role, industry) and product usage patterns (features used, frequency, team size). Assign point values to each attribute, with product usage carrying 60-70% of the total score for PLG companies. Set threshold scores for MQL (30-49 points), PQL (50-79 points), and hot PQL (80+ points). Track conversion rates by score range and adjust weights quarterly based on closed/won data.

What is an example of lead scoring with product usage data?

A project management tool scores users as follows: integration connected (+20), team members invited (+5 each,