A B2B lead qualification framework ICP is a written rule set that tests an account against your ideal customer profile (ICP) first, then checks that a real person can be reached, and only then asks the buying questions that BANT, MEDDIC or CHAMP supply. The order matters. Fit can be checked from public records before anyone is contacted, so it should be settled before a seller spends a call on readiness. This guide gives a three-column criteria sheet, a five-rung ladder from ICP to CRM action, and a cost formula for wrong passes.
B2B Lead Qualification Framework ICP: The Short Answer
- Fit belongs to the account, not the lead. An ICP describes companies. Apollo's guide says so directly: an ICP is company-level traits and a buyer persona is individual-level traits. A form fill or a name on a list is only a candidate until the account passes.
- Sort criteria into Must, Should and No. Must criteria are required, Should criteria add weight, and No criteria remove the account whatever else it scores. The Must-Should-No Sheet below stores the evidence behind each one.
- A fit score is not a reachable buyer. The Fit-to-Action Ladder puts a verified person between fit and the qualification questions, because a good account with no reachable buyer cannot be worked.
- Qualification questions come last. BANT, MEDDIC and CHAMP are discovery models. The BANT vs MEDDIC guide covers choosing one. This guide covers what must be true before you use it.
- Wrong passes cost time and wrong rejections cost deals. The False-Yes Cost section shows the arithmetic with invented numbers, so a team can set its own threshold.
What Competing Guides Cover and Leave Open
Each guide below answers part of the question. The right-hand column records what the page reviewed does not state, not whether the vendor does it elsewhere.
| Source | What it covers | What it does not state |
|---|---|---|
| Highspot, lead qualification process (Oct 1, 2026) | Seven-step process; says the best framework aligns questions with ICP, buying committee, sales motion and decision process; negative scoring for red flags | Where each fit fact came from, or a check that the named person is reachable |
| Tactera Digital, ICP scoring model (Jul 30, 2026) | Two layers (account fit, buyer readiness); disqualifiers that override the score; a split between criteria verifiable before contact and those established in conversation | A person-level verification step, or what a wrong pass costs |
| WorksBuddy, ICP framework guide (May 11, 2026) | Six-step build; threshold bands; written disqualification rules; a handoff that carries the score and the attributes behind it | Evidence standards per attribute |
| ZoomInfo Pipeline, ideal customer profile guide | Five ICP components (firmographic, technographic, behavioral and intent, budget and use-case, psychographic); five build steps; quarterly review | How an ICP match becomes a qualified lead with a named person |
| Apollo, building the ideal customer profile (Apr 5, 2025) | ICP versus persona; five build steps; a template with example ranges | A rule for what disqualifies an account |
| Cognism, ideal customer profile guide (published Sep 29, 2025, updated Apr 16, 2026) | Builds the ICP from best customers, interviews and data analysis | Qualification after the profile exists |
| Demandbase Help Center, Qualification Scores | A model trained on past customers, scoring firmographics, technographics and a year of historical intent; bands of 95 and 50 | The page covers accounts only, and only accounts with a domain |
| Leadfeeder, ICP Insights | Flags website-visiting companies that match an ICP segment defined by location, industry and company size | Anything about the person or the buying question |
| 6sense, lead qualification guide | Defines lead qualification as judging fit and likelihood to buy; walks through BANT, CHAMP, ANUM and MEDDPICC | Fit testing before the discovery questions |
Across these pages the gap is the middle of the chain. Fit scoring and discovery models are well covered. The step between them, where a named person is confirmed and the evidence is stored, is not described in any of them. Opinion: that step is where a framework either works or becomes a spreadsheet nobody trusts.
The Chain From ICP to CRM Action
The path has five links, and each one can stop the lead.
- ICP. Is the account inside the profile built from your best customers? The guide to defining an ICP from won-deal data covers building it.
- Research. Is there a dated, public reason the account has the problem you solve?
- Verified contact. Is there a named person in a seat that influences the purchase, with a working address, who may be contacted?
- Qualification. Does a seller's question confirm the readiness that data cannot show?
- CRM action. Is the result written to the record with an owner, a status and a reason?
The ICP scoring guide covers weighting criteria, and the B2B lead qualification guide covers the process around it. The rest of this guide covers how to store evidence and where to put the stops.
The Must-Should-No Sheet
The Must-Should-No Sheet replaces a flat list of ICP traits with three columns and an evidence standard per row.
| Column | Meaning | Effect on the lead | Evidence to store |
|---|---|---|---|
| Must | The account cannot be a customer without it | Fails the ICP if missing | Source and date, from a registry, filing or the company's own site where possible |
| Should | The trait raises the odds but is not required | Adds weight, never decides alone | Source and date, with a note if it is a vendor value |
| No | A condition that removes the account | Removed whatever the score | Source, date and the rule it triggered |
Apollo's template and the build steps in ZoomInfo's and Cognism's guides list the traits to start from: industry, size, revenue band, tools in use, buying triggers. Those lists are the input. The sheet adds the rule for how each trait is used and what proves it.
Typical rows, written as an example:
- Must: operates in the industry you serve; headcount inside the range your product supports; sells in a country you can serve.
- Should: a recent trigger, such as a change in the function you sell to; a tool in use that your product connects to.
- No: already a customer or open opportunity; a competitor; outside the countries you can contact lawfully; a company too small to buy at your price.
Tactera's guide makes the same split in its own terms, separating criteria that can be verified before contact from those only established in conversation. The sheet adds one rule: a Must trait needs a stored source, and a Should trait taken from a vendor field is marked as a vendor value.
The firmographic and technographic data guide shows how to label each field as Recorded, Detected or Inferred before it goes in a column.
The Fit-to-Action Ladder
The Fit-to-Action Ladder turns the chain into five rungs. A lead moves up only when the rung's check passes, and every stop writes a reason so that rejected leads can be reviewed.
| Rung | Check | Pass writes | Stop writes |
|---|---|---|---|
| 1. Fit | All Must traits met, no No trait present | Account marked in ICP, with the evidence date | Not in ICP, with the failed trait |
| 2. Reason | A dated public signal ties the account to the problem | Signal text, link and date | Fit only, no current reason |
| 3. Person | A named person in a relevant seat, address verified, contact permitted | Name, role, verification result and date | No reachable buyer |
| 4. Question | The seller's answer confirms need, authority and timing | Accepted, with the answers | Returned with the missing answer |
| 5. Action | The record is created or updated | Owner, status, next step | Recycled or closed with a reason |
Two rungs are the unusual ones. Rung 3 is separate from rung 1 because an ICP is about companies and a lead is a person. A fit account can still have no reachable buyer, and the committee is large: Forrester's Buyers' Journey Survey, 2025, as reported on its blog in February 2026, found that 73% of purchases involve three or more departments, with an average of 13 people inside the buyer's organization and nine from outside. The buying committee guide covers which seat to name. One name at one account is a starting point, not a committee.
Rung 3 is also where contact checks happen. The contact verification guide separates a valid address from a delivered message. The FTC's CAN-SPAM guide says the law covers all commercial email and makes no exception for business-to-business email, and that opt-out requests must be honored within 10 business days. Contact permission is therefore a check on the person, not on the account, and it belongs on this rung.
Rung 4 is where BANT, MEDDIC or CHAMP enter. Those models ask a person about budget, authority, need, timing, metrics or priorities. By this rung the seller is confirming, not discovering, which is the shape Highspot's guide describes when it says to align questions with the ICP and the buying committee.
The False-Yes Cost: What a Wrong Pass Costs
A false yes is a lead the ladder passed that should have stopped. A false no is a lead it stopped that would have bought. Both have a price, and the threshold is a choice between them.
The numbers below belong to an invented company called Kestrel Row, which sells dock-scheduling software to mid-sized logistics operators. They are made up to show the arithmetic and are not a result from any company.
- A false yes costs 35 minutes of seller time: research, two touches and a call preparation.
- Seller time costs $60 an hour, so a false yes costs $35.
- An average first-year contract is $18,000, and fit accounts that reach rung 4 win 8% of the time, so a missed true lead has an expected value of $1,440.
Dividing the two gives the break-even: $1,440 divided by $35 is about 41. Tightening a rule only pays if it removes more than 41 false yeses for each true lead it also removes. A rule that blocks 30 false yeses a week and loses one true lead every month passes that test. A rule that loses one true lead for every 10 false yeses blocked does not.
| Kestrel Row, invented funnel | Accounts |
|---|---|
| Pass rung 1 on firmographics alone | 100 |
| Removed by No traits (customers, competitors, out-of-territory) | 9 |
| Pass rung 1 | 91 |
| Pass rung 2 (dated reason found) | 54 |
| Pass rung 3 (named, verified, permitted person) | 41 |
| Accepted at rung 4 | 17 |
Kestrel Row's seller meets 17 accounts that have a reason, a person and a confirmed answer, rather than 100 that matched a filter. Whether the extra stops are worth the speed depends on the two prices above. Replace the invented numbers with your own and the break-even changes.
Where Lead Seeker, Percepture and Pyra Fit
Lead Seeker publishes this guide. It starts from your ICP: Lead Compass reads your website, ICP and intent themes to surface opportunity picks with public sources where available, and a search returns verified people with the public signal behind each one. Read how Lead Seeker works for the flow, and the best B2B contact database guide for the contact layer that rung 3 depends on.
Percepture, where Bob Generale is President, runs B2B intent data services. Its guide on intent data to identify sales qualified leads covers the timing signal that supports rung 2. Pyra, which Bob co-founded, sells a paid pilot of one workflow with approval gates; it fits where one rung, such as writing rung 5 to the CRM, has enough volume to test an agent.
Limitations
- The ladder is an editorial method, not a benchmark. No page compared here publishes a win-rate or conversion figure that links fit tiers to closed-won deals with a method, so this guide gives none.
- Vendor models are not tested here. Demandbase's score needs training accounts and a domain, and Leadfeeder's segment uses three attributes. Those are descriptions from their own pages, not measured performance.
- The threshold is yours. The break-even depends on your seller cost, deal size and win rate, which this guide does not know.
- Legal checks depend on where you sell. CAN-SPAM is a US law. Rules in other countries and states differ.
- Fit predicts a conversation, not a purchase. A lead that passes all five rungs has met your criteria. It has not bought.
Frequently Asked Questions
What is a B2B lead qualification framework ICP?
It is a written set of rules that tests an account against your ideal customer profile, confirms a reachable person, and then applies discovery questions such as BANT or MEDDIC. Fit is checked first because it can be read from public data before anyone is contacted.
Should ICP fit come before BANT or MEDDIC?
Yes, in the order this guide recommends. ICP fit is a property of the account that data can usually show, while BANT and MEDDIC are questions a person answers. Running the fit check first keeps the question time for accounts that can buy.
What are good disqualification criteria for an ICP?
Use binary conditions that no score should outweigh: an existing customer or open deal, a competitor, a country you cannot contact lawfully, and a company outside the size your product supports. Store the rule that fired so rejected leads can be reviewed.
How is an ICP different from a buyer persona?
An ICP describes the company and a persona describes the person. Apollo's guide puts it as company-level traits for targeting and individual-level traits for messaging. You need both, and the qualification ladder tests the account before the person.
Where does a lead qualification framework write to the CRM?
At the last rung. Write an owner, a status, the evidence behind each rung and a reason code for any stop, so that a seller can see why the lead was accepted and a manager can review why one was not.
How to Read the Tables
The Must-Should-No Sheet, the Fit-to-Action Ladder, the False-Yes Cost and the Kestrel Row example are this guide's own editorial methods, not vendor documentation or benchmarks. Kestrel Row is invented and its numbers are illustrative. Vendor descriptions are not performance testing, and "does not state" means the page reviewed does not say, not that the vendor lacks the feature. Source dates are the dates each source gives for itself. This is an October 2026 snapshot.
Sources
- Highspot, Lead Qualification Process: The 2026 Sales Checklist (Oct 1, 2026).
- Tactera Digital, B2B Lead Qualification Framework ICP: The Scoring Model (Jul 30, 2026).
- WorksBuddy, ICP-Based B2B Lead Qualification Framework for IT Companies (May 11, 2026).
- ZoomInfo Pipeline, Ideal Customer Profile.
- Apollo, Building the Ideal Customer Profile for Your Sales Strategy (Apr 5, 2025).
- Cognism, How to Create an Ideal Customer Profile (ICP) With Template (published Sep 29, 2025; updated Apr 16, 2026).
- Demandbase Help Center, Understanding Qualification Scores.
- Leadfeeder, ICP Insights.
- 6sense, The Perfect Guide to Lead Qualification.
- Forrester, Three Realities About B2B Buying Networks (Feb 23, 2026), citing the Buyers' Journey Survey, 2025.
- Federal Trade Commission, CAN-SPAM Act: A Compliance Guide for Business.
- Percepture, How to Use Intent Data to Identify Sales Qualified Leads and B2B Intent Data Services. Percepture is a company Bob Generale leads.
- Pyra, Pricing and paid pilot page. Pyra is a company the author co-founded.
- Lead Seeker, Lead Compass and Pricing: the free-leads offer, the 14-day Pilot and "No auto-spend". Lead Seeker publishes this page.
About the Author
Bob Generale is President of Percepture. He works across SEO, AI search, digital PR, sales intelligence and AI-powered revenue systems, with a focus on connecting visibility, buyer intent and sales action.
Next Steps
Write your Must, Should and No columns on one page, attach a source and date to every Must row, and add a reason code for every stop. To start from your ICP rather than a lead form, try Lead Compass or start a free Lead Seeker trial.