How to Build an Ideal Customer Profile from Firmographics

To build an ideal customer profile (ICP) from firmographic data, start with your best existing customers, record the same company facts for each one (industry, size, location, age, ownership and so on), look for the traits they share, then turn those traits into filters you can apply to a list of companies. The result is a written description of the kind of company most likely to buy from you and stay, plus a set of rules precise enough that two people would pick the same companies from the same list.

Most teams skip the first step and write an ICP from instinct. This post walks through a version based on evidence, which works even if you only have a few dozen customers.

What is an ideal customer profile?

An ICP describes a type of company. It answers the question: which organisations get the most value from what we sell, buy without excessive effort, and stay? Buyer personas, which describe the individuals inside those companies, come afterwards.

A good ICP is:

• Specific enough to filter with. "Mid-sized companies" is too vague. "Manufacturers with 50 to 250 employees, registered in Germany, Austria or Switzerland" can be applied to a list.
• Based on outcomes, meaning revenue, retention and ease of sale, rather than on who replied most often.
• Written down and dated, so you can revise it when the evidence changes.

What is firmographic data?

Firmographics are to companies what demographics are to people: descriptive facts about the organisation. The common ones are:

• Industry, ideally as a standard code such as NACE or NAICS as well as a plain label.
• Size, by employee count, revenue band or both.
• Location, meaning country of registration and where the company actually operates.
• Age, from the incorporation date.
• Legal form and ownership, such as private limited company, partnership, subsidiary of a group, or public company.
• Status, whether the company is active, dormant or being wound up.
• Web presence, such as whether it has a working website and what that site says it does.

Technographic data (which software a company uses) and intent data (signs that it is researching a purchase) are useful extras, but firmographics are the foundation because they are relatively stable and widely available.

Step 1: Which of your customers are the best ones?

Export your customer list with three numbers per account: revenue to date, how long they have been a customer, and a rough measure of how hard they were to win and serve. Then sort.

Pick the top group, the customers you would gladly clone. Also note the ones that went badly: churned quickly, needed constant support, or never paid on time. You will compare the two groups.

If you have fewer than twenty customers, include strong late-stage prospects and recent lost deals where you know why they were lost. A small sample is still better than a guess, as long as you treat the result as a first draft.

Step 2: What firmographic facts do you record for each?

For every company in both groups, fill in the same columns: industry code, employee band, revenue band if available, country, incorporation year, legal form, whether it belongs to a group, and its website.

Record where each fact came from and when you checked it. Company size and status change, and a column of employee counts from mixed sources and mixed years is a weak foundation. Official business registers, the company's own website and its filed accounts are the most reliable starting points.

Step 3: What traits do the best customers share?

Now compare. Look for traits that appear far more often among the best customers than among the poor ones:

• Are most of them in two or three industries?
• Do they cluster in a size band?
• Are they mostly in certain countries or regions?
• Are they older, established companies or young ones?
• Do they tend to be independent or part of a group?

Also look for disqualifiers: traits that appear mostly among the bad accounts. These are often as useful as the positive traits, because they stop you wasting effort.

Step 4: How do you turn traits into filters?

Write each trait as a rule that a database or spreadsheet can apply:

1. Industry code in a defined list.
2. Employee count between two numbers.
3. Country of registration in a defined list.
4. Status is active.
5. Incorporated at least a set number of years ago.
6. Has a working company website.
7. Not a subsidiary of a group, if that proved to be a disqualifier.

Keep the must-have rules short. Every extra rule shrinks the list, and too many rules leave you with almost nothing. Put softer preferences into a score instead.

Step 5: How do you score and rank the results?

Apply the must-have filters to a broad company list, then score what remains on the softer traits. A simple points system is enough: a few points for a preferred industry, a few for the ideal size band, a few for a website that mentions a relevant product or service.

Sort by score and review the top of the list by hand. If the companies at the top look right to someone who knows your customers, the profile is working. If they look wrong, a rule is probably too loose or a trait is missing.

How do you check the ICP is working?

Treat the first version as a hypothesis. Over the next few months, track:

• Reply and meeting rates from ICP companies compared with others.
• Deal size and time to close.
• Retention of customers who match versus those who do not.

Revise the profile when the numbers say so, and record the date of each version.

When you start reaching out to the list, the contact data needs the same care as the company data. Our posts on list hygiene and bounce rates and email authentication cover the sending side.

What mistakes should you avoid?

• Building the ICP from your largest logo rather than from the pattern across your best accounts.
• Using stale or unsourced firmographics, so the filters run on numbers nobody can trace.
• Writing an ICP too broad to filter with, which turns into "any business that might need us".
• Never revisiting it after the market or your offer changes.

Working with Syntora Ai

Syntora Ai builds company lists filtered to a defined customer profile, with the source and check date kept for each fact, as part of our growth and data practice. If you want help turning your customer history into an ICP and a target list, write to hello@syntorahq.ai.