Key takeaways
An ideal customer profile, or ICP, is a written description of the companies most likely to buy from you and stay. It names the company facts, such as industry, size and location, the situation that creates the need, and the signs that rule a company out. You build it from your best customers and research, then use it to decide which companies to contact first.
- An ideal customer profile describes the companies most likely to buy and stay; a buyer persona describes the people inside them.
- Build it from evidence: your best customers, the deals you lost and cited research on the segments you are unsure about.
- Write must-haves, good signs and exclusions in plain words, and keep the first version to one page.
- Turn it into a list in stages: research, then matching companies, then the right people at the best of them.
- Treat the profile as a hypothesis. Test it on a short list, record what the replies teach you and revise it.
What an ideal customer profile is and is not
An ideal customer profile describes a type of company, not a person. It answers one question: which companies should we spend our time on first? A good one fits on a page. It lists the facts a company must match, the facts that make it a better fit, and the facts that rule it out.
Three nearby ideas are often mixed up with it. A buyer persona describes the people inside those companies: their role, goals and worries. A customer segment is a group of companies or contacts you treat the same way, for example in a campaign. A target account list is the named companies you have chosen to pursue this quarter. The profile comes first, because it is the rule the other three are built from.
In this guide ICP means ideal customer profile. It is a working document, not a one-off exercise: you write a first version from what you know, then correct it as real conversations teach you more.
Why a written profile saves time in prospecting
Without a written profile, every person on the team carries a slightly different picture of the ideal customer. Lists grow long, research spreads thin and messages try to speak to everyone. The cost shows up as weeks spent on companies that were never going to buy.
Research aimed at the right companies
You read about markets and accounts that match the profile, instead of whatever comes up first.
Shorter lists with fewer wasted leads
Each lead you find costs time and, in Pebbles Ai, one lead from your wallet. A profile keeps the list to companies worth that cost.
Messages with a real angle
When you know the situation that creates the need, the opening line can name it.
A settled answer instead of weekly debate
The team agrees the profile once, in writing, and changes it on evidence rather than opinion.
The profile also makes AI help useful. An assistant can find companies that match a clear description. It cannot guess which companies you quietly meant. The clearer the profile, the better the search, the research and the drafts that follow from it.
How to build an ideal customer profile in five steps
Start from evidence you already have, write the profile in plain words, and test it on a small list before you trust it. The five steps below work for a first profile and for a revision.
1. Collect the evidence
List your best customers and a few deals you lost or regretted. Note industry, size, location, what they bought and what was happening when they bought.
2. Describe the company
Write the facts a company must match: industry, headcount or revenue range, head office country or region, and the business model.
3. Describe the situation
Write what creates the need: a stage, a change, a tool in use, a job being hired for, a problem the company has in public.
4. Write the exclusions
Name the signs that rule a company out, however attractive it looks: too small to pay, a sector you cannot serve, a region you cannot support.
5. Test it on a short list
Find ten to twenty companies that match, research two or three, and contact a handful. Keep what the replies teach you and revise the profile.
Keep the first version short. A profile with three must-haves, three good signs and three exclusions is easier to use and to test than a two-page specification. Add detail only when a real deal shows you it matters.
In Pebbles Ai, the Strategy assistant can draft the profile from your Company Intel, where Brand voice already describes your Audience, and from the research and notes you give it. You review and correct the draft. The CRM, in Beta, keeps ideal customer profiles beside your contacts, segments and deals, so the team works from the same one.
From profile to prospect list
A profile earns its keep when it turns into a list of companies and people to contact. The path has three stages: research the parts you are unsure about, find companies that match, then find the right people at the best of them.
Research before you commit
Where the evidence is thin, ask for cited research on the market or the segment. Decide on facts you can check, not on a hunch.
Find matching companies
Describe the profile to the Marketing or Strategy assistant. It finds companies by keyword, country or region, industry and headcount, 25 by default and up to 50.
Find the right people
Pick up to ten companies and start Find fresh leads with the roles and locations from your profile. The run finds up to 100 leads and saves them as CRM contacts in a segment.
Each lead comes with personalisation angles drawn from their public LinkedIn posts, or from their role and company when there are no posts. That is where the situation you wrote in step three pays off: the opening line can speak to it.
Starting a lead search asks for your approval in Manual mode, the web app default. Each lead found uses one lead from your Pebbles wallet, so a tight profile keeps the cost of a run in step with its value.
Limits to keep in mind
An ideal customer profile is a hypothesis about who will buy, not a fact. The evidence behind a first version is small, so treat the profile as something to test and revise, not a rule to defend.
AI help has limits too. An assistant can draft a profile from your company context and research, and find companies that match a description. It does not know how your customers fared after they bought, and it cannot see data behind a login. Company search returns up to 50 companies per search, people search runs inside up to ten named companies per run, and contact details depend on what is public. A run can finish with fewer leads than you asked for.
Keep the profile where the team can see it and put a date on it. Revisit it after each round of outreach, after a pricing or product change, and at least once a quarter. When two different kinds of company keep buying, write two profiles rather than one vague one.
How an ICP differs from a buyer persona
Three documents that are often confused. Each answers a different question, and the profile comes first because the other two are built from it.
| Ideal customer profile | Buyer persona | Target account list | |
|---|---|---|---|
| Describes | A type of company | A type of person inside those companies | Named companies chosen for a period |
| Answers | Which companies should we spend time on first? | Who do we talk to there, and what do they care about? | Which companies are we working this quarter? |
| Built from | Best customers, lost deals and cited research | Conversations, calls and the roles that bought | A search for companies that match the profile |
| Changes | Slowly, on evidence from real deals | When the buying group or their worries shift | Every period, as accounts are won, lost or parked |
| In Pebbles Ai | Drafted with the Strategy assistant from Company Intel; kept in the CRM, in Beta | The roles and locations you give a Find fresh leads run; a persona for each campaign piece | The companies your assistant finds, up to 50 per search, and the up to ten you pick for a lead search |
A general description of three working documents, not a method every team must follow. Small teams often keep all three on one page.
Worked example:
from best customers to a lead list
A ten-person software company sells to operations teams and wants to stop chasing every inbound lead. It needs a profile it can act on within a week.
- List the ten best customers and three lost deals. For each, note industry, headcount, head office country, what they bought and what was happening when they bought.
- Ask the Strategy assistant to draft a one-page profile from Company Intel and those notes: three must-haves, three good signs and three exclusions. Correct anything that does not match what the team knows.
- Where the evidence is thin, such as whether a second sector is worth pursuing, ask for cited research on that segment and decide on the facts.
- Describe the profile to the assistant to find matching companies, up to 50 per search. Pick the ten that fit best and record why.
- Start Find fresh leads at those ten companies with the roles and locations from the profile, and approve the run in Manual mode. The leads land as CRM contacts in a segment, each with personalisation angles.
- After the first replies, write down what matched and what did not, revise the profile and date the new version.
These steps describe a working method, not a promise of results. Company search returns up to 50 companies, people search runs inside up to ten named companies, each lead found uses one lead, and actions that reach the outside world follow the execution mode.
Common mistakes
Describing everyone who could buy
A profile that fits most companies in your market is a market description, not a profile. Narrow it until it tells you who to call first.
Writing a persona and calling it a profile
A description of the decision maker is useful, but it does not tell you which companies to look at. Write the company first, then the people.
Borrowing a competitor's profile
Their best customers were won with their product, pricing and story. Start from your own customers and lost deals.
Skipping the exclusions
Without written exclusions, attractive but wrong companies creep back onto every list. The exclusions are the part the team will thank you for.
Treating the first version as final
The evidence behind a first profile is small. Test it on a short list, record what the replies teach you and revise it on a date you set.
Practical checklist
- Can we name our ten best customers and say what they had in common when they bought?
- Does the profile state three must-haves, three good signs and three exclusions in plain words?
- Which parts of the profile rest on evidence, and which on assumption that cited research could check?
- Can the profile be turned into search filters: keyword, industry, headcount, head office country or region?
- Which roles and locations will we look for at the companies that match?
- Who owns the profile, where does the team find it, and when is the next review?
Key terms
- Ideal customer profile (ICP)
- A written description of the type of company most likely to buy from you, succeed with you and stay: the facts it must match and the signs that rule it out.
- Buyer persona
- A description of a person inside a target company: their role, goals, worries and how they decide. A profile describes the company; a persona describes the people.
- Firmographics
- The facts that describe a company rather than a person: industry, headcount, revenue, location, business model and age.
- Trigger
- A change or situation that creates the need for what you sell, such as a funding round, a new hire, a new market or a tool reaching its limits.
- Exclusion
- A sign that rules a company out of the profile, however attractive it looks: a sector you cannot serve, a size that cannot pay, a region you cannot support.
- Customer segment
- A group of companies or contacts you treat the same way. In Pebbles Ai, leads land as CRM contacts in a segment, and a template can be sent to a segment.
- Target account list
- The named companies you have chosen to pursue in a given period, selected because they match the profile.
- Company context
- A reviewed description of a business that an AI assistant reads before it works: what the company does, who it serves, what it offers and how it communicates.
Common questions
What is an ideal customer profile in B2B?
An ideal customer profile, or ICP, is a short written description of the type of company most likely to buy from you and stay. It names the facts a company must match, such as industry, size and location, the situation that creates the need, and the signs that rule a company out. Teams use it to decide which companies to research and contact first.
What is the difference between an ICP and a buyer persona?
An ideal customer profile describes a company. A buyer persona describes a person inside that company: their role, goals and worries. You need both, but the profile comes first. It tells you which companies to look at, and the persona tells you who to talk to there and what to say.
How do I create an ideal customer profile step by step?
Follow five steps. First, collect evidence from your best customers and a few lost deals. Second, describe the company: industry, size, location and business model. Third, describe the situation that creates the need. Fourth, write the exclusions that rule a company out. Fifth, test the profile on a short list of companies and revise it from what the replies teach you.
What should an ideal customer profile include?
Three short lists. Must-haves are the facts a company has to match, such as industry, headcount range and head office region. Good signs make a company a better fit, such as a trigger like a new hire or a funding round. Exclusions rule a company out, however attractive it looks. Add the roles you usually sell to, so the profile can turn into a lead search.
How many ideal customer profiles should a B2B company have?
Start with one. A single clear profile is easier to test and to act on than three vague ones. Write a second profile only when a different kind of company keeps buying for a different reason, or when you sell a second offer to a different market. Keep each one to a page.
How do I find companies that match my ideal customer profile with AI?
Describe the profile to an assistant that can search for companies. In Pebbles Ai, the Marketing or Strategy assistant finds companies by keyword, head office country or region, industry and headcount. A search returns 25 companies by default and up to 50. Then pick up to ten companies and start Find fresh leads for the roles in your profile. The leads land as CRM contacts in a segment.
Can AI write my ideal customer profile for me?
It can draft one, if you give it evidence. In Pebbles Ai, the Strategy assistant drafts a profile from your Company Intel, where Brand voice describes your Audience, and from the customer notes and research you provide. The draft is only as good as that evidence, so review it against what your team knows about real deals before you use it.
How often should I update my ideal customer profile?
Review it after each round of outreach, after a pricing or product change, and at least once a quarter. Change it on evidence, such as who replied, who bought and who churned, not on opinion. Put a date on each version so the team knows which one is current.
Where should the team keep the ideal customer profile?
Somewhere everyone who researches, prospects or writes can see it. In Pebbles Ai, the Audience section of Brand voice in Company Intel describes who you serve, and the CRM, in Beta, keeps ideal customer profiles beside contacts, segments and deals. Keep one dated version, and point the team to it rather than to copies.
In Pebbles Ai
- CapabilityProspecting & customer recordsFind leads at named companies, ask for companies that match your ideal customer and keep contacts, segments and deals in the CRM.
- CapabilityResearch & intelligenceRead company websites and public sources, pull out what matters and build research you can trace back to its sources.
- FoundationCompany IntelBusiness knowledge and brand voice that every assistant works from, built from your website, company information and files.
About this guide. Written by the Pebbles Ai team and checked against the current product before publication. Product details describe the web app on the update date shown above; each linked platform page carries the latest detail. Guides are revised when the product changes, and the date changes only with a verified revision.
