Start free
ProductHow it worksPlatform tourSafety and controlsYour first 30 daysCompetitor trackingLinkedIn signalsDone for youHubSpotSalesforceSmartleadAll integrationsFounder-led salesBefore the first SDRAccount-based marketingClosed-lost reactivation
Free toolsFree ICP assessmentSandbox workspaceSPF, DKIM and DMARC checkOutbound cost calculatorB2B outbound glossaryRoast my cold emailOutlevo CommunityICP generatorCold email generatorSubject line graderEmail word checkerBlacklist checkUse it from Claude, Cursor, LovableDeveloper hubAPI reference
IntegrationsFor AI buildersPricingResources
Theme
Log inStart free

Guides

ICP Examples for B2B Teams: Fields and Worked Cases

Worked B2B ICP examples with firmographic, behavioral and disqualifying fields, plus steps to turn past customers into a testable hypothesis.

An ICP example that reads "mid-market SaaS companies with growth goals" cannot be tested, and it cannot disqualify anyone.

A usable ICP is a short set of fields. Each field either matches a company or it does not. This guide shows four worked examples, each built from a different pattern in a team's past deals. It does not give you industry thresholds or benchmark win rates, because those numbers only mean something when they come from your own customers. The examples use blanks, and you fill the blanks from your own data.

Who this is for

This guide is for a founder, a sales lead or a marketer who has a list of past customers and wants to turn it into a profile a team can act on. It fits you if:

  • You have closed some deals, won and lost, in a CRM or a spreadsheet.
  • Your current ICP is a sentence, not a set of fields.
  • You want a profile you can test before you spend on outreach or ads.

If you have no closed deals yet, example 4 below is for you. For the wider process, see How to Define an Ideal Customer Profile. For the glossary definition, see Ideal customer profile (ICP).

The three field groups

Every example below uses the same three groups.

Firmographic fields describe the company: what it is, how big it is, where it is. A CSV import in Outlevo maps your columns to domain, name, country, size, industry and deal fields such as stage, amount, outcome and loss reason, for files of up to 200,000 rows. [[1]](#sources) In HubSpot, the company fields read are name, domain, owner and lifecycle stage. [[2]](#sources)

Behavioral fields describe what the company did. Inside a CRM, that is mostly deal behavior: which pipeline, which stage, what amount. [[2]](#sources) On your website, it is which pages a company read. The website visitor script shows which companies visit, which pages they read and whether they fit your ICP. It never identifies a person. [[3]](#sources) Those are company-level signals only, so write behavioral fields at the company level.

Disqualifying fields describe what rules a company out. The best source is the closed-lost reason on lost deals, which HubSpot deals carry alongside pipeline, stage and amount. [[2]](#sources) A profile without this group matches everyone.

Worked example 1: the firmographic cluster

The pattern in the data. Your won companies look alike on paper. Open the won list and the same industry and the same size band keep coming up.

The profile card:

  • Firmographic: industry is [the industry shared by most won companies]. Size is [the band that holds most won companies]. Country is [a country you already sell into].
  • Behavioral: reached [the deal stage where won deals usually accelerate].
  • Disqualifying: size outside the band, or an industry that appears only in lost deals.

The hypothesis sentence: "Companies in [industry] and [size band] close at a higher rate for us than companies outside that profile."

This is the easiest example to write and the easiest to get wrong. If you count deals, one large customer with five renewals looks like five wins. Outlevo's win/loss analysis counts companies, not deals, and it is a good habit to copy by hand if you work in a spreadsheet. [[4]](#sources)

On the country line, write only the countries you chose to sell into. Outlevo asks which countries a customer sells into and never proposes markets. [[4]](#sources) An ICP describes who fits within the markets you already serve. It does not decide the markets for you.

Worked example 2: the behavioral separator

The pattern in the data. Won and lost companies look the same on firmographics. Same industries, same sizes. But the wins did something the losses did not.

The profile card:

  • Firmographic: [the broad band you already sell to; keep it loose].
  • Behavioral: visited [the page the won companies read before they became deals, such as pricing or an integration page]. Reached [a specific stage] within [a time window you choose from your own history].
  • Disqualifying: no visit to that page and no reply after [the number of touches you set in your policy].

The hypothesis sentence: "Among companies that fit the firmographic band, those that read [page] are the ones worth a conversation first."

Website visitor tracking costs 1 credit per company identified per day, however many pages that company reads. [[3]](#sources) Browsers that send Do Not Track or Global Privacy Control, and automated browsers, send nothing at all. [[3]](#sources) So the behavioral field describes the visits you can see, not every visit that happened. Say so in the profile, so nobody reads a missing visit as a missing interest.

Worked example 3: the disqualifier-led profile

The pattern in the data. Wins are scattered, but the losses repeat. The same closed-lost reasons keep coming back.

The profile card:

  • Firmographic: [loose; the wins do not cluster].
  • Behavioral: [none yet; leave this blank rather than guess].
  • Disqualifying: closed-lost reason is [the most common reason on your lost list, such as a missing capability or a tool they already use]. Lost to [a named competitor] in more than [a count you pick] companies.

The hypothesis sentence: "Excluding companies that show [the disqualifier] raises the share of conversations that become deals."

Disqualifiers can be checked before anyone is contacted. Competitor losses are worth their own line. Outlevo's competitor tracking covers quoted facts from competitors' public pages, changes, win/loss by competitor and competitor signals on accounts. [[5]](#sources) For tools in that category, see Competitor tracking software for B2B.

Treat a closed-lost reason as a lead, not a verdict. Read a good sample of them before you turn one into a rule.

Worked example 4: too few wins

The pattern in the data. You have a handful of customers, or none. Any pattern you see could be luck.

The profile card:

  • Firmographic: [what your own site says you serve, taken from the page text].
  • Behavioral: [leave blank until you have data].
  • Disqualifying: [what your own team already knows you cannot serve].

The hypothesis sentence: "We think [segment] is a fit, and we will treat that as unproven until we have results."

Outlevo labels win/loss samples by size, and under 10 won companies is very low confidence. [[4]](#sources) That label is useful even if you never use the tool. Write it at the top of the profile so nobody mistakes a guess for a finding.

Here the free assessment helps. It reads up to 5 public pages of a website and returns facts labelled as verified observations or AI inferences, and up to 5 experimental segment hypotheses. [[6]](#sources) The assessment also asks which countries you sell into. [[7]](#sources) You can run it at the free ICP assessment.

Steps

Here is how a team turns a list of past customers into a testable hypothesis.

  1. Gather the closed deals. Outlevo reads accounts, opportunities and their stages from Salesforce for the ICP and win/loss analysis. [[8]](#sources) It reads companies, deals and owners from HubSpot. [[2]](#sources) If your deals live in a spreadsheet, use a CSV import. [[1]](#sources) Importing past deals from a CSV or your CRM gives you a win/loss analysis that proposes ICP hypotheses. [[9]](#sources)
  1. Collapse deals to companies. One row per company, marked won or lost. Do this before you look for patterns. A company that bought three times is one data point.
  1. Check the sample. Count the won companies. Under 10 means your profile is a guess, and it should be labelled as one. [[4]](#sources)
  1. Look for the pattern. Sort won companies by one field at a time: industry, size, country. Then try the behavioral fields. Then read the lost reasons. Match what you see to one of the four examples above. Expect one example to fit your data better than the others.
  1. Write the profile card. Fill the three groups with real values. Keep each field checkable: a value, a band, or a yes/no. If a field needs a person's judgment to evaluate, rewrite it.
  1. Write the hypothesis in one sentence, and the test beside it. Name the segment, name the comparison group, and decide in advance what result would change your mind. Write the result you will count, such as replies or meetings, before you look at any numbers. Outlevo counts each Calendly or Cal.com booking with one of your contacts as a meeting booked. [[10]](#sources) [[11]](#sources)
  1. Get a person to approve it. AI-proposed ICP segments are labelled experimental until a person approves them. [[4]](#sources) Apply the same rule to hypotheses a teammate wrote on a whiteboard.
  1. Set the policy. Policies set who can be contacted, on which channels, how often and with what budget, and nothing is sent outside a policy that a person approved. [[12]](#sources) Your disqualifying fields belong in that policy as well as in the profile.
  1. Qualify accounts against the profile. Each qualified account gets a Readiness Card with fit, intent and data quality scores. [[13]](#sources) Qualification costs 3 credits, and light qualification costs 1 credit. [[14]](#sources) For the method, see How to Qualify B2B Accounts.
  1. Write the result back. With HubSpot, Outlevo writes status, fit and intent scores, ICP version, reason, next action and due date to company fields it owns. Nothing is written until an owner or admin approves the field mapping. [[2]](#sources) Standard HubSpot fields are never edited. [[2]](#sources)
  1. Read the test, then change one thing. If the segment does not beat the comparison group, change one field and run again. Keep the old version of the profile so you can see what changed.

What to check

Before you call the profile done, run through these questions.

Is the sample honest? Count companies, not deals. If the won list is under 10 companies, label the profile very low confidence. [[4]](#sources) One favorite customer is a story, not a segment.

Is every field checkable? "Innovative" is not a field. "Industry is X" is. If a teammate cannot check a field in under a minute from public information, it is not ready. If you cannot find a value for a field on most accounts, drop it.

Does the profile describe the company, not the buyer? An ICP describes the company. A persona describes the person. Keep them in separate documents, or your firmographic field turns into a job-title list.

Is each fact backed by evidence? A Readiness Card needs at least three observations, each with a source and a date. [[13]](#sources) Facts read from a website are kept only when their quote is found on the page, and guesses are labelled as hypotheses. [[13]](#sources) Hold your own profile to the same bar: mark each line as observed or inferred. Until a segment has been tested and a person has approved it, label it experimental.

Does the profile have disqualifiers? A profile with no disqualifying fields will match almost any company. Add at least one before you approve it, and draw it from your lost deals as well as your wins.

Did you name the markets yourself? The countries in the profile should come from you. Outlevo never proposes markets. [[4]](#sources)

Do your contact rules match the profile? A company that fits the ICP is not automatically a company you may email. Two outside rules are worth reading:

  • The FTC's CAN-SPAM guide says: "The law makes no exception for business-to-business email." [[15]](#sources)
  • The UK's ICO says on PECR: "You can email or text any corporate body (a company, Scottish partnership, limited liability partnership or government body)." [[16]](#sources)

Those two statements cover different rules, and neither one is the whole picture. Treat them as prompts to check your own obligations, and put the result in the policy a person approves. [[12]](#sources)

Is data quality visible? The Readiness Card carries a data quality score next to fit and intent. [[13]](#sources) A high fit score on poor data deserves less trust than the same score on good data.

Is the profile versioned? Outlevo writes an ICP version to the CRM company record. [[2]](#sources) If you work by hand, add a date and a version number to the profile card, so results can be tied to the version that produced them.

Next step

Pick the example that matches your data and fill in its profile card with real values from your own closed deals. If you have fewer than 10 won companies, start with the free ICP assessment to get up to 5 experimental segment hypotheses to test. [[6]](#sources)

If you already have a CRM full of deals, connect HubSpot or Salesforce, or import a CSV, and read the win/loss analysis against the profile card you wrote by hand. Connecting and syncing a CRM uses no credits. [[2]](#sources) [[8]](#sources) An owner or admin connects the CRM and approves the field mapping. [[2]](#sources) [[8]](#sources)

For the full method behind these examples, read How to Define an Ideal Customer Profile. To see how Outlevo handles approval and limits, see How Outlevo works.

Sources

  1. CSV and Excel integration
  2. HubSpot integration
  3. Website visitors integration
  4. ICP from evidence
  5. Release 2026-10-19
  6. Free ICP assessment
  7. Release 2026-10-19
  8. Salesforce integration
  9. Release 2026-10-19
  10. Calendly integration
  11. Cal.com integration
  12. How Outlevo works
  13. Readiness Cards
  14. Outlevo pricing
  15. FTC: CAN-SPAM Act compliance guide for business
  16. ICO: Electronic mail marketing (PECR)

Related

See what Outlevo finds on your own website

Free, no sign-up, about a minute.

Get a free ICP assessment