What Is an Ideal Customer Profile (ICP)? With Examples (2026)
An ICP describes the company you sell to, not the person inside it. Here is the definition, the test that separates a real ICP from a wish, and 3 worked examples.
What Is an Ideal Customer Profile (ICP)? With Examples (2026)
An ideal customer profile is a description of the company you sell to: its size, industry, geography, and the specific problem it has that your product solves. It is a company-level filter, not a person. A good ICP is precise enough to turn into a query that returns a findable list of real accounts you can contact this week.
Most founders write an ICP that reads like a horoscope. "Forward-thinking mid-market companies that value innovation." That is not an ideal customer profile. It is a mood board, and you cannot build a target list from a mood.
Here is the distinction that fixes it: an ICP describes the COMPANY you sell to, a buyer persona describes the PERSON inside it. Founders who blur the two ship lists nobody can act on. a16z Growth calls the ICP "the north star for your entire company", the prerequisite for any repeatable B2B sales motion. At seed, though, it is something more concrete: a learnable list, not a market size.
What is an ideal customer profile? (the definition that actually filters)
An ideal customer profile is the set of company-level attributes that predict a fast, high-value sale, tight enough to resolve into a findable list of accounts.
The ideal customer profile definition that matters at seed has one property the generic ones miss: every attribute is something you can filter a database on. a16z Growth's ICP framework lists the high-fidelity attributes as company size or revenue range, type of business, geography, industries served, job titles, and a definable, solvable problem plus the conditions that create it.
Notice what is on that list and what is not. Company size is there. "Innovative culture" is not. The test:
If you cannot turn your ICP into a query that returns a findable list of companies this week, it is not an ICP yet. It is a fantasy.
ICP vs buyer persona: the mistake that kills your target list
The ICP vs buyer persona confusion is the single most expensive targeting error founders make at seed. The ICP is the company. The persona is the human inside it who signs.
You define the company first, then find the people inside it. Do it backwards and you get HubSpot's cautionary case: a SaaS company that acquired good-fit individuals at the wrong companies, buyers who could never actually deploy the product.
| ICP (the company) | Buyer persona (the person) | |
|---|---|---|
| Answers | Which accounts do we target? | Who inside do we talk to? |
| Attributes | Size, industry, geography, problem | Job title, goals, objections |
| Example | 50-person EU fintech on legacy rails | VP Engineering who owns the budget |
| Built | First | Second, inside each account |
Define the ICP first. The persona lives inside it, never the other way around.
3 ICP examples: filterable attributes vs vanity attributes
The difference between a real ICP and a wish is whether each attribute is filterable and findable. Here are three worked ICP examples for common seed archetypes, with the vanity attributes called out.
Vertical SaaS (dental scheduling):
- Filterable: US dental practices, 3 to 15 locations, currently on paper or spreadsheet scheduling.
- Vanity, cut it: "practices that care about patient experience." You cannot filter on it.
Devtool (API observability):
- Filterable: Series A to B software companies, 20 to 200 engineers, running a microservices architecture on AWS.
- Vanity, cut it: "engineering teams that value reliability." Every team says yes; it narrows nothing.
Services startup (fractional CFO):
- Filterable: US e-commerce brands, $2M to $10M revenue, no full-time finance hire yet.
- Vanity, cut it: "ambitious founders who want to scale." Not a filter, a compliment.
Lenny Rachitsky's ICP playbook is to pick the three most unique and important characteristics of the customer, the smallest set that returns a usable list. Three sharp filters beat ten fuzzy ones every time.
Why an ICP matters more in 2026
Buyer precision is now a defining GTM variable, not a nice-to-have. The buyer landscape is shifting fast: per PitchBook-NVCA data, acquisitions of AI startups made up over a third of venture-ecosystem M&A deals in 2025. If your buyers are being absorbed into AI-native categories, a stale ICP quietly stops resolving to real accounts.
YC's Michael Seibel tells founders chasing their first ten customers to start with a target narrow enough to call or email personally, not a broad TAM. At 0 to 3 users, your ICP is not a market-sizing slide. It is a list of 30 companies you can name today.
If you have a rough description but not the list, tools like the Causo ICP generator turn a fuzzy profile into a findable set of matching companies fast. That is the whole game at seed: get from "I think I sell to fintechs" to a spreadsheet of 40 real ones.
From ICP to a real customer list
A sharp ICP is only worth something once it resolves into named accounts you can email this week. The gap between "US dental practices with 3 to 15 locations" and an actual spreadsheet of them is where most founders stall, because the filterable attributes you defined are exactly what a research agent can go find. Causo takes your ICP as a query, searches the live open internet, and returns matching companies rather than a stale scraped list you rented from Apollo or ZoomInfo.
Once the accounts exist, you still need the person inside each one. Causo pulls the decision-makers that fit your buyer persona and attaches verified emails so you are not guessing at addresses or bouncing outreach, closing the loop from company-level ICP to the human who actually signs. It runs the find, enrich, and reach steps in one place instead of ten browser tabs and three tools that never sync.
Then it drafts the outreach in your own voice, and you stay in the loop to approve every message before it sends. That is the honest version of what an ICP buys you: not a full-auto SDR firing blind, but a researched list and a first draft you can trust. Point the Causo ICP generator at your fuzzy profile and get back the real accounts it resolves to.
FAQ
What is an ideal customer profile (ICP)? An ideal customer profile is a description of the company you sell to: its size, industry, geography, and the specific problem it has that your product solves. It is a company-level filter, not a person. A good ICP is precise enough to turn into a query that returns a findable list of real accounts you can contact this week.
What is the difference between an ICP and a buyer persona? An ICP describes the company (a 50-person fintech in the EU running on legacy infrastructure). A buyer persona describes the person inside that company you talk to (the VP of Engineering who owns the budget). You define the ICP first, then find the personas inside it. Blur the two and you build lists nobody can act on.
What should an ideal customer profile include? Filterable, findable attributes: company size or revenue range, industry, geography, business type, and the specific solvable problem plus the conditions that create it. a16z's framework also lists the job titles you sell into. Leave out vanity attributes like "innovative" or "growth-minded" that you cannot filter a database on.
What is a good example of an ideal customer profile? For a vertical SaaS startup: US dental practices with 3 to 15 locations still using paper or spreadsheet scheduling. Every attribute there is filterable and findable, so it resolves to a real list of accounts. That is the test of a good ICP: it produces a usable target list, not a mood board.
Related on the hub
- Go to market strategy seed founders can execute in 2026 — for when the playbook turns into a raise.
- How to Find Customers for Your Startup (2026) — Related sales guide.
- Build a repeatable B2B sales process at seed (2026) — Related sales guide.
- The H1 2026 AI Sales Outreach Report — Related cold outreach guide.