What Lead Scoring Is and Why It Matters
Lead scoring is a methodology that assigns numerical values to each lead based on how likely they are to become a paying customer. Every lead gets a score based on two dimensions: how well they fit your ideal customer profile (fit score) and how engaged they are with your brand (engagement score). The combined score determines whether a lead should be routed to sales immediately, nurtured by marketing, or deprioritized entirely.
The business case for lead scoring is straightforward. According to research from MarketingSherpa, organizations that use lead scoring see a 77% increase in lead generation ROI. Eloqua found that companies with lead scoring experience a 30% increase in close rates and a 18% decrease in revenue leakage from lost deals. The reason is simple: when sales reps focus on the highest-quality leads first, they close more deals in less time.
But here is the problem: most lead scoring models fail. Sirius Decisions (now Forrester) found that 68% of B2B companies have a lead scoring model, but only 22% describe it as effective. The rest are either too complex to maintain, too disconnected from sales reality, or simply ignored by the reps who are supposed to use them.
Why Most Lead Scoring Models Fail
Before building your model, understand why others fail so you can avoid the same mistakes.
Too Complex from Day One
The number one failure mode is overengineering. Marketing teams build models with 50 or more scoring criteria, weighted across multiple categories with intricate rules and exceptions. The model is theoretically elegant but practically useless. Nobody understands why a lead scored a 73 versus an 81. When sales reps cannot explain the score, they do not trust it. When they do not trust it, they ignore it. Start with 10-15 criteria maximum and add complexity only when you have data proving more criteria improves accuracy.
No Sales Buy-In
If your sales team was not involved in building the lead scoring model, they will not use it. Full stop. Too many lead scoring initiatives are marketing-driven projects imposed on sales without their input. The result is predictable: sales ignores the scores and continues working leads based on gut feeling. The fix is to involve your top 2-3 reps in the model design process. Ask them what characteristics their best customers share. Ask them what behaviors signal a lead is ready to buy. Build the model around their lived experience, then validate it with data.
Never Updated
Markets change. Buyer behavior evolves. Your product and positioning shift. A lead scoring model built in Q1 of last year may be completely wrong by Q4. Yet most teams treat lead scoring as a one-time project rather than an ongoing process. The best lead scoring models are recalibrated quarterly against closed-won and closed-lost data. If the model says a lead should score highly but they consistently fail to convert, something needs adjustment.
The Two Types of Lead Scoring
Effective lead scoring requires two separate scores that combine into an overall assessment: fit scoring and engagement scoring.
Fit Scoring: Who They Are
Fit scoring evaluates whether a lead matches your ideal customer profile based on their attributes - things like company size, industry, job title, and technology stack. Fit scoring answers the question: "Is this the kind of company and person that typically buys from us?" A VP of Sales at a 200-person SaaS company might have a high fit score because that matches your ICP, while a marketing intern at a nonprofit would score low.
Engagement Scoring: What They Do
Engagement scoring evaluates how much a lead has interacted with your brand based on their behavior - website visits, email opens, content downloads, and demo requests. Engagement scoring answers the question: "Is this person actively interested in what we sell?" A lead who visited your pricing page three times, downloaded a case study, and attended a webinar is clearly more engaged than one who opened a single email.
Why You Need Both
Neither score alone is sufficient. A perfect-fit lead with zero engagement is not ready for sales outreach. A highly engaged lead who does not fit your ICP will waste your reps' time. The magic happens when you combine both dimensions. A lead who matches your ICP and is actively engaging with your content is a genuine sales-ready opportunity.
Building Your Fit Score: Firmographic Criteria
Here are the firmographic criteria that matter most for B2B lead scoring, along with suggested point values. These are starting points - adjust them based on your actual closed-won data.
Industry (Up to 25 Points)
Identify the 3-5 industries where you close the most deals and assign maximum points. For example: SaaS and software (25 points), financial services (20 points), professional services (15 points), healthcare tech (15 points), all other industries (5 points). If you have industries where you have never closed a deal, assign 0 or even negative points.
Company Size by Employee Count (Up to 20 Points)
Your ideal company size depends on your product and pricing. A typical B2B SaaS model might look like: 51-200 employees (20 points), 201-500 employees (20 points), 501-1000 employees (15 points), 11-50 employees (10 points), 1000+ employees (10 points), 1-10 employees (0 points). The key is mapping this to where your product delivers the most value and where your win rates are highest.
Revenue (Up to 15 Points)
Revenue serves as a proxy for both budget availability and company maturity. A practical scoring range: $10M-$50M ARR (15 points), $50M-$200M ARR (15 points), $5M-$10M ARR (10 points), $200M+ ARR (10 points), Under $5M ARR (5 points). Revenue data can be harder to obtain, so you may need enrichment tools to fill in gaps.
Funding Stage (Up to 10 Points)
For companies selling to startups and growth-stage companies, funding stage matters: Series B or C (10 points), Series A (8 points), Series D+ (5 points), Bootstrapped/Profitable (5 points), Pre-seed/Seed (2 points). Recently funded companies are more likely to be in buying mode because they have budget to invest in growth.
Technology Stack (Up to 15 Points)
Technographic data tells you whether a company uses tools that indicate fit. If you sell a Salesforce integration, a company that uses Salesforce scores higher than one that uses a different CRM. Example: Uses Salesforce CRM (15 points), uses HubSpot CRM (12 points), uses complementary tools in your integration ecosystem (10 points), uses competitor product (5 points, they might switch), uses no relevant technology (0 points). Tools like BuiltWith, Wappalyzer, or Clay can provide technographic data for enrichment.
Job Title and Seniority (Up to 15 Points)
Not all leads are created equal. A VP-level contact at your target company is worth more than an individual contributor. Score based on both the functional role (is this person in the department that buys your product?) and seniority. Example: VP/Director in target function (15 points), C-level in target function (12 points - often not the day-to-day buyer), Manager in target function (10 points), VP/Director in adjacent function (8 points), Individual contributor in target function (5 points), completely unrelated role (0 points).
Building Your Engagement Score: Behavioral Signals
Engagement scoring tracks what a lead does, not who they are. Here are the behavioral signals to track with suggested point values.
Website Visits (Up to 15 Points)
Not all page visits are equal. Weight high-intent pages more heavily. Pricing page visit (10 points), case study or customer story page (5 points), product feature page (3 points), blog post (1 point), careers page (-5 points, they are likely a job seeker). Multiple visits to the same high-intent page within a 7-day window should be scored once to avoid score inflation from refreshes.
Email Engagement (Up to 15 Points)
Email engagement signals interest but varies by type. Email open (1 point, opens are unreliable due to privacy features), email click (3 points), reply to sales email (10 points), unsubscribe (-10 points). Cap the total email engagement score to avoid inflating scores for people who simply open everything. A common cap is 15 points total from email activity.
Content Downloads (Up to 20 Points)
Gated content downloads are strong engagement signals because the lead exchanged their information for your content. ROI calculator or assessment tool (15 points), product comparison guide (12 points), case study PDF (10 points), industry report (8 points), general ebook or whitepaper (5 points). Weight content closer to a purchase decision more heavily than top-of-funnel educational content.
Demo and Meeting Requests (Up to 30 Points)
These are the strongest engagement signals short of an actual purchase. Demo request form submission (30 points), contact sales form (25 points), free trial signup (20 points), webinar registration (8 points), webinar attendance (12 points). A single demo request should move any reasonably-fit lead into the sales queue immediately.
Score Decay
Engagement scores should decay over time. A lead who downloaded a whitepaper six months ago is not as engaged as one who downloaded it yesterday. Implement score decay by reducing engagement points by 10-20% per month for activities older than 30 days. Some platforms like HubSpot and Marketo have built-in decay functionality. If yours does not, build a monthly workflow that reduces engagement scores for leads who have not had recent activity.
Combining Fit and Engagement: The Scoring Matrix
Once you have both scores, combine them using a 2x2 matrix that creates four categories.
High Fit + High Engagement = MQL (Marketing Qualified Lead)
These are your best leads. They match your ICP and are actively engaged. Route them to sales immediately with a target follow-up time of under 5 minutes for inbound or within 24 hours for enriched outbound leads. Suggested threshold: Fit score above 60 (out of 100) AND engagement score above 40 (out of 100).
High Fit + Low Engagement = Nurture
These leads match your ICP but have not shown enough interest yet. They are the right company and person, they just need more touchpoints. Add them to marketing nurture campaigns with targeted content relevant to their industry and role. Consider outbound outreach to create engagement if the fit is strong enough. Suggested threshold: Fit score above 60, engagement score below 40.
Low Fit + High Engagement = Monitor
These leads are engaged but do not match your ICP well. They might be students, consultants, competitors, or companies outside your target market. Do not ignore them entirely - sometimes they convert or refer you to companies that do fit. But do not prioritize them for sales outreach. Suggested threshold: Fit score below 60, engagement score above 40.
Low Fit + Low Engagement = Deprioritize
These leads neither fit your ICP nor show meaningful engagement. Do not invest time or resources. Keep them in your general marketing database in case something changes, but do not route them to sales or include them in targeted nurture. Suggested threshold: Both scores below 40.
Implementing Lead Scoring in HubSpot
HubSpot offers built-in lead scoring through their HubSpot Score property. Here is how to implement the model we have described. Navigate to Settings, then Properties, then search for HubSpot Score. You can create scoring rules based on contact properties (for fit scoring) and contact activities (for engagement scoring). Create positive attributes for each criterion: set Industry equals [target industry] to add the corresponding points. Set Job Title contains VP or Director to add title points. For engagement, set Contact has visited specific URL to add points for pricing page visits. Set Contact has submitted specific form to add points for demo requests.
HubSpot also supports negative scoring, which is important. Set Email hard bounced to subtract 50 points. Set Job Title contains Student or Intern to subtract 10 points. Create a workflow that triggers when HubSpot Score exceeds your MQL threshold. This workflow should assign the lead to the appropriate sales rep and create a task with a follow-up deadline.
Implementing Lead Scoring in Salesforce
Salesforce does not include native lead scoring, but there are several approaches. The most common is using Salesforce Einstein Lead Scoring, which is available on Enterprise edition and above. Einstein uses machine learning to analyze your historical conversion data and automatically score leads based on patterns in closed-won deals. It is easier to set up than manual scoring but less transparent in how scores are generated.
For manual lead scoring in Salesforce, create a custom number field called Lead Score on the Lead object. Then build a Process Builder flow (or Salesforce Flow) that calculates the score based on field values. For example: if Industry equals Technology, add 25. If Annual Revenue is greater than $10M, add 15. If Lead Source equals Demo Request, add 30. Use a scheduled flow to recalculate scores daily, and create a lead assignment rule that routes leads above your MQL threshold to the appropriate queue.
Testing and Iterating Your Lead Scoring Model
Your first lead scoring model will be wrong. That is expected and fine. The goal is to be directionally correct and improve over time. Here is how to test and iterate.
In the first 30 days after launching your model, run it in shadow mode. Calculate scores for all leads but do not change any routing or processes. Instead, compare your model's predictions against what actually happens. Did the high-scoring leads actually convert at higher rates? Were there conversions that your model missed?
Quarterly recalibration is essential. Pull your last quarter's closed-won deals and analyze their scores at the time they were created. What was the average fit score? The average engagement score? What criteria contributed most to conversions? Do the same analysis for closed-lost deals. Look for criteria that do not differentiate between won and lost - those criteria are noise and should be removed or reweighted.
Track three key metrics: MQL-to-SQL conversion rate (should be 20-30% or higher), average time from MQL to first sales touch (should be under 24 hours), and sales acceptance rate (the percentage of MQLs that sales agrees are worth pursuing, should be above 50%). If your sales acceptance rate is below 50%, your model is generating too many false positives and needs tighter criteria.
Predictive Lead Scoring with AI
Traditional lead scoring relies on humans defining the criteria and weights. Predictive lead scoring uses machine learning to analyze your historical data and identify the patterns that predict conversion. The model discovers which combinations of attributes and behaviors lead to closed deals, often finding signals that humans miss.
Tools like 6sense, Madkudu, Infer (now part of Ignite), and HubSpot's predictive scoring feature can analyze thousands of data points across your CRM, marketing automation, and website analytics to generate scores. The advantage is accuracy - predictive models typically outperform manual models by 20-30% in identifying conversion-ready leads.
The disadvantage is the black box problem. When a predictive model gives a lead a score of 87, nobody knows exactly why. This makes it harder for sales teams to trust and act on the scores. The best approach is to use predictive scoring as a complement to manual scoring, not a replacement. Let the AI surface high-potential leads, but use your manual model to explain why they are high-potential.
You also need sufficient data volume for predictive scoring to work. Most AI scoring tools need at least 1,000 closed-won deals and 10,000 total leads in your database to build a statistically meaningful model. If you are a startup with 50 closed deals, stick with manual scoring until you have the data volume.
Lead Scoring Template: A Starting Point
Here is a complete lead scoring template you can implement today. Fit Score (max 100 points): Industry match (0-25), Company size (0-20), Revenue (0-15), Tech stack (0-15), Job title and seniority (0-15), Funding stage (0-10). Engagement Score (max 100 points): Demo or trial request (30), Pricing page visit (10), Content download (5-15 based on type), Email engagement (1-3 per action, capped at 15), Webinar registration (8) or attendance (12), Website visits (1-10 based on page). Score decay: reduce engagement points by 15% monthly for activities older than 30 days.
Thresholds: MQL (route to sales) - Fit above 60 AND Engagement above 40. Nurture (marketing campaigns) - Fit above 60 AND Engagement below 40. Monitor - Fit below 60 AND Engagement above 40. Deprioritize - Fit below 40 AND Engagement below 40.
Making Lead Scoring Work Long-Term
Lead scoring is not a project. It is a process. The companies that get the most value from lead scoring treat it as a living system that evolves with their business. Schedule quarterly model reviews with both marketing and sales. Recalibrate against actual conversion data. Add new scoring criteria as your ICP evolves. Remove criteria that do not predict conversion. And most importantly, listen to your sales team. If they say the scores do not match reality, they are probably right.
The ultimate goal of lead scoring is alignment. When marketing and sales agree on what a qualified lead looks like and the data backs up that definition, you eliminate the most common source of friction between the two teams. Pipeline velocity increases, conversion rates improve, and both teams spend their time on the work that actually drives revenue.
Need help building or optimizing your lead scoring model? GTME specializes in RevOps infrastructure, including lead scoring implementations in HubSpot and Salesforce. We also provide the data enrichment layer that makes fit scoring possible, using waterfall enrichment to fill in firmographic, technographic, and demographic data for every lead in your CRM. Learn more at gtmeagency.com/services.