Marketing Demand Generation

Set up lead scoring based on the leads that convert

How do you check that your lead scores predict which leads convert?

The Scoring Analyst agent measures how many leads above your qualification threshold convert. The Scoring Analyst agent finds the top scoring gaps, where a lead's score does not match its outcome. The Model Tuner agent changes the scoring criteria and their weights. After 60 days, the Model Tuner agent compares accuracy before and after the change.

Published Updated

What you get
Lead scoring model · Saltmarrow Studio
Confirmed after the 60-day check
75%
of leads above the threshold became opportunities, up from 58%
3 / 3
top scoring gaps fixed, each with a written reason
Before the change · 24 leads above the threshold, 14 opportunities 14 ÷ 24 = 58%
Gap 1 · leads from one free download scored high, no plan to buy Points lowered
Gap 2 · good-fit leads scored low, job title field empty Points moved to company size and industry
Gap 3 · leads with high email activity scored high, no budget Email-open points removed
New criterion · pricing-page visits Added
After 60 days · 16 leads above the threshold, 12 opportunities 12 ÷ 16 = 75%
Opportunities from leads below the threshold 36% before, 20% after

Your salespeople work the leads above the threshold first.

See it run

60-day check done, your confirmation needed

Inbound leads a month
18 · Scored in HubSpot
Top scoring gaps fixed 3 / 3
Run log
  1. You
    “Check the lead scores in HubSpot against the leads that became opportunities. Fix the top gaps.”
    Your instruction
  2. Accuracy measured on 54 leads Scoring Analyst
    You chose a created opportunity as the proof of a good lead. You chose the last 90 days of data, because you get 18 leads a month. The Scoring Analyst agent took 54 scored leads from HubSpot. Of the 24 leads above the qualification threshold, 14 became opportunities, or 58%. Of the 22 opportunities, 8 came from leads below the threshold, or 36%. These two shares are your baseline, the accuracy before any change. A statistical comparison needs at least 50 leads above the threshold a month. You have about 8 a month, below that minimum. You chose to use the two shares as a rough baseline until you have more leads.
  3. 3 gaps confirmed with your salespeople Scoring Analyst
    A false positive is a lead above the threshold that did not convert. A false negative is a lead below the threshold that converted. You chose to rank the gaps by how often they happen. The Scoring Analyst agent found the top three gaps. Gap 1 is 5 false positives: leads from one free download with no plan to buy. Gap 2 is 5 false negatives: good-fit leads with an empty job title field. Gap 3 is 4 false positives: leads with high email activity and no budget. The Scoring Analyst agent attached three example leads to each gap. You showed the example leads to your salespeople. Your salespeople recognized all three gaps. Your salespeople also named pricing-page visits as a strong buying signal.
  4. Criteria and weights adjusted Model Tuner
    Fit criteria show how well a lead matches your ideal customer profile. Behavior criteria show how ready a lead is to buy. You chose to balance fit and behavior criteria, with a minimum fit score for every qualified lead. For gap 1, the Model Tuner agent lowered the points for the free download. For gap 2, the Model Tuner agent moved the job title points to the company size and industry criteria. For gap 3, the Model Tuner agent removed the points for email opens. The Model Tuner agent added pricing-page visits as a new criterion. The Model Tuner agent wrote a reason from the conversion data for each change. You approved the adjusted model on August 13.
  5. You
    Your confirmation needed You
    You used the adjusted model in HubSpot from August 14 to October 12. The Model Tuner agent compared the accuracy on 36 new leads with the baseline. Of the 16 qualified leads, 12 became opportunities, or 75%. Of the 15 opportunities, 3 came from leads below the threshold, or 20%. A higher share of leads above the threshold converted than in the baseline. A lower share of opportunities came from leads below the threshold. The Model Tuner agent recommends that you confirm the model. Confirm the scoring model.
  6. Lead generation plan used Upstream
    Your lead generation plan comes from Plan Lead Generation.
  7. Ideal customer profile used Upstream
    Your ideal customer profile comes from Define Ideal Customer Profile.
How it works

Lead scores checked against the leads that convert

You choose the event that proves a lead was truly qualified. The Scoring Analyst agent takes the scored leads and their outcomes from your CRM. The Scoring Analyst agent finds the criteria behind the false positives and false negatives. You check each gap with your salespeople. The Model Tuner agent proposes weight changes, each with a reason from the data. You approve the changes before launch. After 60 days, you confirm the model or change it again.

What it needs
Lead scores and conversion outcomes for the last 90 days, from your CRM
Input from your salespeople on which scored leads they find valuable and which they do not
The original scoring model from Operations
Enough lead volume for a statistical comparison: at least 50 qualified leads a month
What you get

An updated lead scoring model with the top scoring gaps closed and each weight change backed by conversion data. A 60-day before-and-after comparison that shows the accuracy improved.

Run sequence

Start to finish

You start from your lead generation plan and your ideal customer profile. Then you, the Scoring Analyst agent and the Model Tuner agent set up your lead scoring model in four parts.

  1. 1
    Current Scoring Accuracy

    Measure your current scoring accuracy

    You choose the event that proves qualification
    Scoring Analyst measures conversion above the threshold
    You confirm the sample is large enough
    Scoring Analyst + You
  2. 2
    Top Scoring Gaps

    Find the top scoring gaps

    Scoring Analyst finds false positives and false negatives
    Scoring Analyst attaches example leads to each gap
    You check the gaps with your salespeople
    Scoring Analyst + You
  3. 3
    Adjusted Scoring Criteria

    Adjust the scoring criteria and weights

    Model Tuner traces each gap to its criterion
    Model Tuner proposes weight changes with reasons
    You approve the adjusted model before launch
    Model Tuner + You
  4. 4
    60-Day Accuracy Check

    Check the model after 60 days

    You run the adjusted model for 60 days
    Model Tuner compares accuracy with the baseline
    You confirm the model or change it again
    Model Tuner + You
Before it starts

What Works asks you

A few questions in plain words. The answers set up the run for your business.

Which conversion event proves that a lead was truly qualified: a created opportunity or a closed deal?
YouA created opportunity, the clearest sign of real qualification
Do you have at least the minimum number of qualified leads each month for a meaningful comparison, or should you widen the time window?
YouWiden to a full quarter to reach the minimum
How many gaps should you fix this round: the top three, or only the single worst one?
YouTop three, a balanced first pass
Is there a buying signal that your salespeople rely on and the model does not capture at all today?
YouPricing-page visits, which salespeople treat as a strong sign
Should fit criteria (how well a lead fits) or behavior criteria (how ready a lead is to buy) carry more weight in your adjusted model?
YouBalanced, with fit as a firm minimum
What size of accuracy gain counts as a clear improvement worth confirming, rather than random change?
YouAny gain in both the false positive and false negative rates
Who runs it

The agents behind this workflow

The Scoring Analyst agent measures your lead scores and finds the gaps. The Model Tuner agent changes the scoring criteria and their weights.

Playbook specialist

Lead Capture and Nurture Specialist

Owns this area

Creates lead magnets and sets up lead scoring, so you capture the right leads and follow up with the best ones first.

What it can do
define scoring criteria with salesbuild and tune a lead scoring modelsend the best-fit leads to sales first
On this workflow

Scoring Analyst

Pulls lead history from the CRM, measures how well lead scores predict conversion, and finds the false positives and false negatives

What it can do
Pull scored leads and outcomes from the CRMCalculate conversion rates above and below the thresholdFind the criteria that cause false positivesFind the criteria that cause false negativesAttach example accounts to each gap

Model Tuner

Changes the scoring criteria and weights to close the confirmed gaps, and runs the 60-day comparison

What it can do
Trace each gap to the criterion that causes itPropose weight changes with reasons from the dataEstimate the effect of a change before it goes liveCompare accuracy after the change with the baselineRecommend confirming the model or changing it again
Jeanie, your right hand Works Crew

Jeanie leads a core crew of seven that runs your company day to day. They direct the specialist and workflow agents in each playbook, including the ones on this workflow.

Connect your tools

Bring what you already run

Works connects to the apps you already use.

Other apps you can use
Your spreadsheet app
  • Microsoft Excel
  • Airtable
  • Smartsheet
  • Zoho Sheet

or any spreadsheet app you use

Your CRM

or any CRM you use

These are starting points, not a limit. Connect the tools you already use.

Connect your work

The workflows before and after your lead scoring model

Linked cards are live pages. The others are part of the Marketing playbook in Works.

Works · AI business OS

One workflow inside your whole business

In Marketing, Set Up Lead Scoring is part of the Demand Generation work.

Prebuilt playbooks, plus your own
StrategyMarketingSalesCustomer SuccessPartnershipsProducts & ServicesOperationsFinancePeopleLegal & Compliance + Build your own
The Marketing playbook
74
Workflows
20
Specialist agents
147
Workflow agents
226
Templates
You're here
Set Up Lead Scoring
One of 74 Marketing workflows · run by 2 workflow agents
Templates

What comes with it

Built for this workflow
Lead Scoring Model XLSX
A sheet where you score your leads. It opens in Works after you sign up.
Scoring Criteria Matrix XLSX
A sheet where you list the criteria you score leads by. It opens in Works after you sign up.
Also included here
Demand Funnel Model XLSX
A sheet where you count the leads at each stage, from first lead to sale. It opens in Works after you sign up.
Nurture Flow Map DOCX
A document where you plan the order of the emails you send to leads. It opens in Works after you sign up.
Prospect List Inventory XLSX
A sheet where you list your prospects. It opens in Works after you sign up.
ABM Engagement Tracker XLSX
A sheet where you track how each target account responds to your account-based marketing. It opens in Works after you sign up.
Intent-to-Action Map DOCX
A document where you match each intent signal to an action. It opens in Works after you sign up.

Each template opens inside Works after you sign up.

Questions

What people ask

What counts as accurate lead scoring?

You choose the event that proves a lead was truly qualified, for example a created opportunity. The Scoring Analyst agent then calculates two shares. The first is the share of leads above the qualification threshold that convert. The second is the share of opportunities that came from leads below the threshold. These two shares are your baseline accuracy.

What is a scoring gap?

A scoring gap is a repeated case where a lead's score does not match its outcome. A false positive is a lead above the threshold that did not convert. A false negative is a lead below the threshold that converted. The Scoring Analyst agent attaches example leads to each gap. You check the gaps with your salespeople.

What happens after the 60 days?

The Model Tuner agent compares the accuracy after the change with the baseline accuracy. You decide what size of gain counts as a clear improvement. If only one gap improved, you decide whether to confirm that gain or wait until all gaps improve. The Model Tuner agent recommends that you confirm the model or change it again.

How do you use AI to set up lead scoring?

The Scoring Analyst and the Model Tuner are AI agents. The Scoring Analyst agent takes the scored leads and their outcomes from your CRM. The Scoring Analyst agent finds the criteria that cause false positives and false negatives. The Model Tuner agent proposes weight changes with reasons from the data. You approve each change before launch.

Match your lead scores to the leads that convert

Sign up to measure the accuracy of your lead scores.

Sign Up