In short
An ideal customer profile (ICP) describes the kind of company that gets the most value from a product and brings the most value back: it stays, pays and does not need constant help. Sales and marketing teams use it to decide which companies to look for and which to leave alone. A good ICP is not a wish. It is built from evidence, usually from the company's own best and worst customers.
In this lesson, learners work with the customer data of Fabrikam, a fictional company that sells online appointment booking and staff scheduling software to outpatient clinics. They sort twelve customers, look for patterns and write Fabrikam's ICP, including an anti-ICP: the companies Fabrikam should not chase.
Level: vocational school (for example industrial or office management apprentices), bachelor courses in business, continuing education in sales and marketing. Time: 90 minutes. Groups: three or four learners.
Running the lesson
90 min- 10 min
Warm-up: "Who is a good customer for a bakery's catering service, and who is a bad one?" Collect answers on the board in two columns.
- 15 min
Input: ICP versus persona, the four kinds of criteria, fit and anti-ICP. Use the talking points below.
- 25 min
Group work, part 1: tasks 1 to 3. Sort Fabrikam's twelve customers, calculate revenue to date and find patterns.
- 15 min
Group work, part 2: task 4. Write Fabrikam's ICP and anti-ICP on the canvas.
- 15 min
Two groups present, the others compare. Then everyone tests their ICP on eight new leads (the sorting exercise below).
- 10 min
Reflection and exit ticket: why does an anti-ICP save money? Explain the homework.
Materials
This page on a projector, or printed: the customer table and the tasks.
One printed ICP canvas per group, from the ICP and persona canvas, or a flipchart with the same fields.
Calculators or a spreadsheet for task 2.
Optional: devices for the sorting exercise at the end.
Talking points for the input
ICP versus persona. The ICP describes a company (a dental group with five practices). A buyer persona describes a person in that company (the practice manager who plans the rota). You need both, but the ICP comes first: the right person at the wrong company is still a bad lead.
Firmographics are facts about a company, the way demographics are facts about a person: industry, number of employees, number of locations, country, legal form, age of the company, revenue.
Technographics are the technology a company already uses, such as the software it would have to connect to.
Situation and pains. A pain is a problem the customer feels and wants solved, in their own words: "Our phone line is blocked every morning." Two companies with the same firmographics can have very different pains.
Fit. A company fits when the product solves its pain, it can afford it, it can start using it without huge effort, and it can buy it in a reasonable time.
Anti-ICP (also called a negative ICP). A short list of companies you deliberately do not target, because they churn, cost too much support or never sign. Saying no early saves time for both sides.
Evidence. Every criterion should point back to real customers. "Because it feels right" is not evidence.
Fabrikam's customers
Fabrikam charges 200 euros per clinic location per month. The table shows twelve customers, numbered C01 to C12. All companies and figures are invented for teaching.
Customer | Kind of clinic | Locations | Staff | Country | Connects to their practice software? | Months as customer | Monthly fee (€) | Support tickets, last 12 months | Status |
|---|---|---|---|---|---|---|---|---|---|
C01 Lamna Healthcare | Physiotherapy group | 8 | 95 | Germany | Yes | 34 | 1,600 | 6 | Active |
C02 | Dental group | 5 | 60 | Austria | Yes | 28 | 1,000 | 4 | Active |
C03 | Physiotherapy group | 3 | 28 | Germany | Yes | 22 | 600 | 3 | Active |
C04 | Dental group | 12 | 150 | Switzerland | Yes | 19 | 2,400 | 9 | Active |
C05 | Physiotherapy group | 4 | 40 | Germany | Yes | 31 | 800 | 2 | Active |
C06 | Single dental practice | 1 | 4 | Germany | No | 5 | 200 | 7 | Left |
C07 | Single physiotherapy practice | 1 | 3 | Austria | No | 3 | 200 | 5 | Left |
C08 | Hospital outpatient department | 1 | 2,000 | Germany | No | 14 | 900 (discount) | 41 | Left |
C09 | GP practice (family doctor) | 2 | 18 | Germany | Yes | 12 | 400 | 12 | Active |
C10 | Cosmetic clinic chain | 6 | 45 | Germany | No | 7 | 1,200 | 15 | Left |
C11 | Veterinary group | 3 | 30 | Netherlands | No | 9 | 600 | 6 | Active |
C12 | Dental group | 2 | 22 | Switzerland | Yes | 24 | 400 | 1 | Active |
What the customers said (from sales notes and exit interviews):
C01: "Our phone line was blocked every morning. Since patients book online, the front desk can breathe, and reminders cut our no-shows."
C02: "We wanted new patients to book online in the evening, when we are closed."
C03: "My front desk staff spent half the day on the phone."
C04: "We needed one staff rota across all twelve practices."
C05: "Short-notice cancellations left gaps we could not fill."
C06: "Too expensive for one dentist. I also had no time to set it up."
C07: "I never really finished the setup."
C08: "Our procurement took eleven months, and IT asked for interfaces to our hospital system that did not exist."
C09: "Most of our patients are older and prefer to call. Hardly anyone books online."
C10: "We bought it to run marketing campaigns, and it does not really do that."
C11: "Booking works well, but we need animal records, which it does not have."
C12: "It connected to our practice software on day one."
Tasks for the groups
Sort. Use these rules. A best customer is still active after at least 18 months and had fewer than 10 support tickets. A worst customer left within 15 months or had 12 or more tickets. Which customers are best, which are worst, and which fit neither rule?
Calculate. Revenue to date is the monthly fee times the months as a customer. Calculate it for every customer. Then work out the average revenue to date and the average number of support tickets for the best and for the worst group.
Find patterns. What do the best customers have in common? Look at the kind of clinic, locations, staff, country, software and what they said. What do the worst customers have in common? Find at least four patterns.
Write the ICP. Fill in the canvas: industry, size, region, technology, situation and pains, how they buy, and an anti-ICP with at least three kinds of companies Fabrikam should not target. Next to each criterion, write the customers that back it up.
Solutions to tasks 1 and 2
Task 1. Best: C01, C02, C03, C04, C05 and C12 (six customers). Worst: C06, C07, C08, C09 and C10 (five customers). C09 is still active but had 12 tickets, so it counts as worst. C11 fits neither rule: it is active and has few tickets, but after 9 months it is too early to say. Good answers point out that "too early to tell" is a result too.
Task 2. Revenue to date:
Customer | Calculation | Revenue to date (€) |
|---|---|---|
C01 | 1,600 x 34 | 54,400 |
C02 | 1,000 x 28 | 28,000 |
C03 | 600 x 22 | 13,200 |
C04 | 2,400 x 19 | 45,600 |
C05 | 800 x 31 | 24,800 |
C06 | 200 x 5 | 1,000 |
C07 | 200 x 3 | 600 |
C08 | 900 x 14 | 12,600 |
C09 | 400 x 12 | 4,800 |
C10 | 1,200 x 7 | 8,400 |
C11 | 600 x 9 | 5,400 |
C12 | 400 x 24 | 9,600 |
Best group: 175,600 euros in total, on average about 29,267 euros per customer (175,600 / 6 = 29,266.67), and 25 tickets, about 4.2 per customer.
Worst group: 27,400 euros in total, 5,480 euros per customer on average (27,400 / 5), and 80 tickets, 16 per customer.
A best customer has brought in more than five times as much so far and needed about a quarter of the support.
Solutions to tasks 3 and 4
Patterns in the best group
All are physiotherapy or dental groups: clinics where patients book appointments in advance.
All have 2 to 12 locations and 22 to 150 staff.
All are in Germany, Austria or Switzerland, where Fabrikam offers support in German.
All already use practice software that Fabrikam connects to.
Their pains match the product: overloaded phone lines, no-shows and short-notice cancellations, one rota across several sites.
Patterns in the worst group
Very small (one location, three or four staff): the fee is high for them and nobody has time to set it up (C06, C07).
Very large and public (C08): a long buying process, special interfaces, many support tickets.
The pain does not match: patients who prefer to call (C09), a wish for marketing features (C10).
Most of them could not connect Fabrikam to their own software.
Model ICP for Fabrikam
Field | Fabrikam's ICP | Evidence |
|---|---|---|
Industry | Outpatient physiotherapy and dental groups | C01 to C05, C12 |
Size | 2 to 12 locations, about 20 to 150 staff | All best customers |
Region | Germany, Austria, Switzerland | All best customers |
Technology | Uses practice software Fabrikam connects to | All best customers, against C06, C07, C08, C10 |
Situation and pains | Phone lines overloaded, many no-shows or late cancellations, rota across several sites is hard to plan | Notes of C01, C03, C04, C05 |
How they buy | Owner or practice manager decides within weeks, no public tender | Contrast with C08 |
Anti-ICP: who Fabrikam should not target
Single practitioners with one location: too small to benefit, likely to leave.
Hospitals and public bodies that buy through long tenders and need custom interfaces.
Clinics that want marketing tools rather than booking and scheduling.
Practices whose patients mostly book by phone and do not want to change.
Open questions (a strong group will name them): veterinary groups (C11) might fit later, but there is not enough evidence yet. Twelve customers are a small sample, so the ICP should be checked again after the next quarter.
Test your ICP
After the groups have presented, everyone sorts eight new leads. Use the model ICP above or the class's own.
Good fit or bad fit for Fabrikam?
Good fit
Bad fit
Differentiation
Easier
Give out the table with the best and worst groups already marked, and skip task 2 or give the revenue column.
Give a word bank for task 3: size, locations, country, software, pain, buying process.
Fill only five canvas fields: industry, size, region, pain and one anti-ICP line.
Harder
Weight the criteria. Give each one a weight from 1 to 3 and justify it. Which single criterion would you drop first if you had to?
Calculate revenue to date per support ticket for each group. What does it tell you about the cost of serving a bad-fit customer?
Name the data that is missing for a solid ICP, for example the time each deal took to close, the cost of winning each customer, or why the active customers stay.
Write three questions a salesperson could ask in the first call to find out quickly whether a clinic fits.
Homework
Write a short ICP for a company you know well, such as your training company or a business in your family. Name at least four firmographic criteria, two pains it solves for its customers and two kinds of companies it should not target. Use descriptions, not names of real customers, and leave out any personal data.
Notes for the teacher
Learners often start with size only ("big companies are best"). Point them to C08 and C10: size without the right pain does not make a good customer.
Some groups will want to put C11 in the best group. Ask what would change their mind in six months.
An ICP describes companies, not people. In a real company, the data for this analysis comes from the CRM (customer relationship management system) and stays inside the company.
Follow-up lessons: describe the people in the buying decision with the buyer personas exercise, turn the ICP into points with the lead scoring worksheet, or size the market with the market segmentation exercise.