Your Ideal Customer

Your Ideal Customer

Smaller is better.

Smaller is better.

“My niche is ‘startups.’”

That’s the answer I hear most when I ask agency owners who they help. It sounds like an answer, but it doesn’t get them any closer to landing clients.

The Money Maker Map

A free, 33-step map for agency owners who want to grow from $100K to $1M+ (without guessing what to work on next).

The Money Maker Map

A free, 33-step map for agency owners who want to grow from $100K to $1M+ (without guessing what to work on next).

“Startups” is a market. What agency owners actually need is an ideal customer profile, or ICP. Here’s what an ICP is and the 2 ways to choose yours.

What an ICP is

Here’s how I define it:

A group of people with a shared, expensive, painful problem who already spend money trying to solve it.

Every piece of that sentence matters, so let’s take them one at a time.

First, a group of people. “Startups” is a fine place to start, but it’s not an ICP yet. CMOs are an ICP. So are chief product officers, veterans, pick-up basketball players, and divorced moms. (Mine is design agency owners. When I ran my design agency, our ICP was design leaders at big companies.)

Second, a shared, expensive, painful problem. “Shared” means everyone in the group has it. Here’s a test I give the agency owners I coach: if your ICP said, “You decide what we should work on,” what would you pick? If the answer is different for every client, it’s not an ICP. If it’s the same, you get to make the same offer to all of them. If it works for one, it’ll work for the others.

Third, who already spend money trying to solve it. You don’t want to convince anyone the problem exists. Trying to convince people that they have a problem is one of the most expensive parts of marketing. Instead, you want them to take money they already spend and spend it with you.

Here’s what that looks like when it works. Picture a few years from now. Someone in your ICP needs help with the exact problem you solve. They ask people they trust for a name, and everyone says yours. They visit your site, and the first line describes them perfectly. Your case studies are for companies they know and look up to, and each one shows a result they want. One of the quotes is from someone they met at a conference last month.

They haven’t seen a single design from you yet, and they’re already sold. You’re clearly the right fit.

Size does matter

Here’s the goal of an ICP: pick a market you can dominate. You want to be the #1 choice, the agency everyone in that group thinks of first.

Most agency owners pick an ICP that’s too big. Bruce Henderson, who founded Boston Consulting Group, came up with the Rule of 3 and 4 in 1976. It says a steady market has only 3 big players. The leader has about twice as much of the market as #2, and about four times as much as #3. In 2011, for example, the top 3 US rental car companies held 48%, 22%, and 14% of the market.

So if your ICP is startups, being the go-to means working with something like 40% of all startups. There are an estimated 150 million startups in the world. 40% of it is 60 million companies. A busy agency works with 10 to 20 clients a year. The math ain’t mathing.

To run an agency doing $1M to $3M, you probably need around 10 to 20 clients a year. That’s also about as many as you can handle. So flip it around: pick an ICP where the entire market is only about 50 companies. That’s small enough for you to be the #1 choice. And if the #1 choice wins around 40% of a market, 40% of 50 is 20 clients, right in that range. If your ICP feels too small, you’re probably headed in the right direction. Fight the fear that this is too narrow and carry on.

2 ways to choose your ICP

When I ask agency owners who their ICP is, I usually get one of two answers:

  1. “No idea.”

  2. “I’ve got a few candidates and I don’t know how to narrow it down.”

Each answer gets its own path.

Path 1: Start with you (when you have no idea)

In Good to Great, Jim Collins describes the Hedgehog Concept (named after hedgehogs, who are really only good at one particular thing): great companies sit where 3 things overlap. They love what they do, they can be the best in the world at it, and it makes them money.

Drop the money part for now. You can make money doing almost anything if you commit to it long enough. That leaves 2 questions.

Who would you love working with every day? You’ll find this answer in your daydreams. What do you nerd out about? Landscaping? Comic books? Software? Baseball? This matters more than it sounds, because you’re going to pitch this group over and over and hear no a lot. Pick the group you can hear no from and still go, “Got it. Here’s another offer.”

Who are you the best in the world at helping? You’ll find this answer in your past. Rory Vaden, cofounder of Brand Builders Group and author of Take the Stairs, puts it this way: “You are most powerfully positioned to serve the person you once were.” What big change have you been through? Where were you once stuck that other people are stuck now?

That’s my story. I ran an agency for a decade. I know what that’s like, and now I help people who run agencies. A lot of people build their agency around their skills. I’d build yours around your experience. “We’re a Webflow agency” doesn’t promise any change. “I used to build websites the slow way, found a faster way, and now I help teams make the same switch” does.

Try it: make 2 columns. On the left, everyone you’d love to work with every day. On the right, every big change you’ve been through. Look for overlap.

Path 2: Start with your clients (when you have candidates)

If you’ve been at this a while, your 5 best clients already told you who your ICP could be.

Make a table with 4 columns:

  1. Who paid you the most?

  2. Who did you get the fastest results for?

  3. Who complained the least?

  4. Who came back for more?

A client who shows up in all 4 columns is a strong signal. (For extra credit, build the opposite table: who paid the least, complained the most, and never came back.)

Then look for the pattern underneath. Were they all women? All in one industry? All in Europe?

Go a level deeper. Who hired you? Which department? What month? What did the contract look like? If your best clients all signed in March, you know when your outreach should start: February.

What if the people who paid the most aren’t the people you liked working with most? That’s the reality for a lot of agency owners. Pick based on the season you’re in. If you need cash, follow the money. If you have money in the bank, follow what you love. Just know that switching later has a cost.

Pick one this week

Once you have a candidate, commit to it. If you’re still weighing 3 ICPs, you haven’t chosen one. You’re chasing 3 rabbits.

Write it down in one line: “I help { ICP } get { OUTCOME }.”

Then pressure-test it. Paste this into ChatGPT, Claude, Gemini, or your AI of choice, in an incognito window so it isn’t working from what it already knows about you:

I help { ICP } get { OUTCOME }.

Is this a good ICP? Good means it’s a market my agency could dominate. I work with 10–15 clients a year, 1 project per client per year, and I want those clients to be about 40% of the companies in the market, which makes me the clear leader.

Estimate how many companies fit this ICP (low-high) and say how you counted.

Calculate the share of the market I’d hold with 10 clients and with 15. Is the market too big (I’d hold under 50%, roughly more than 30 companies), too small (fewer than about 20 companies, or fewer than my clients), or about right?

Using BCG’s Rule of Three and Four (in a stable market the leader holds about twice the share of #2 and four times the share of #3), tell me which agencies already help this ICP achieve that outcome and whether the leader’s seat is open.

If this isn’t a good ICP, suggest 2–3 narrower, wider, or adjacent ICPs that would pass the same test, with a company count for each. Then give a 3-sentence verdict.

Use tables. Mark which numbers are guesses so I can verify them.
I help { ICP } get { OUTCOME }.

Is this a good ICP? Good means it’s a market my agency could dominate. I work with 10–15 clients a year, 1 project per client per year, and I want those clients to be about 40% of the companies in the market, which makes me the clear leader.

Estimate how many companies fit this ICP (low-high) and say how you counted.

Calculate the share of the market I’d hold with 10 clients and with 15. Is the market too big (I’d hold under 50%, roughly more than 30 companies), too small (fewer than about 20 companies, or fewer than my clients), or about right?

Using BCG’s Rule of Three and Four (in a stable market the leader holds about twice the share of #2 and four times the share of #3), tell me which agencies already help this ICP achieve that outcome and whether the leader’s seat is open.

If this isn’t a good ICP, suggest 2–3 narrower, wider, or adjacent ICPs that would pass the same test, with a company count for each. Then give a 3-sentence verdict.

Use tables. Mark which numbers are guesses so I can verify them.
I help { ICP } get { OUTCOME }.

Is this a good ICP? Good means it’s a market my agency could dominate. I work with 10–15 clients a year, 1 project per client per year, and I want those clients to be about 40% of the companies in the market, which makes me the clear leader.

Estimate how many companies fit this ICP (low-high) and say how you counted.

Calculate the share of the market I’d hold with 10 clients and with 15. Is the market too big (I’d hold under 50%, roughly more than 30 companies), too small (fewer than about 20 companies, or fewer than my clients), or about right?

Using BCG’s Rule of Three and Four (in a stable market the leader holds about twice the share of #2 and four times the share of #3), tell me which agencies already help this ICP achieve that outcome and whether the leader’s seat is open.

If this isn’t a good ICP, suggest 2–3 narrower, wider, or adjacent ICPs that would pass the same test, with a company count for each. Then give a 3-sentence verdict.

Use tables. Mark which numbers are guesses so I can verify them.

If it says your ICP is too big (and it probably will), take one of its narrower suggestions and run the prompt again. Keep going until one passes.

Next time someone asks who you help, you won’t say “startups.” You’ll name a group small enough to know by name. And they’ll already know yours.

read this next

Don’t let clients ruin it.

Bg Image

featured

make

27.9881° N, 86.9250° E

more

Design the agency that funds the life you want to live.

money

11.3733° N, 142.5917° E

Bg Image

featured

make

more

Design the agency that funds the life you want to live.

money

Bg Image

featured

make

more

Design the agency that funds the life you want to live.

money

Helping designers get their flowers.

45,601 followers

•

15,979 connections

45,100 followers

6,458 subscribers

•

79 videos

6,296 followers

•

221 posts

This site is typeset in Bueno by Rajesh Rajput, Thermal by Reset Type Studio, Mark by Hannes von Döhren and Christoph Koeberlin, Manteca by Emyself Design, and Inter by Rasmus Andersson.


This is version 6.0.60 of my personal website. Older versions: v5, v4, v3, v2 (no longer online), v1 (no longer online).


Black lives matter.

© Dan Mall Teaches 2005–2026. All rights reserved.

Made proudly in Philly. Thou shalt not steal—but feel free to remix.

Privacy policy.

Dan Mall

Helping designers get their flowers.

45,601 followers

•

15,979 connections

45,100 followers

6,458 subscribers

•

79 videos

6,296 followers

•

221 posts

This site is typeset in Bueno by Rajesh Rajput, Thermal by Reset Type Studio, Mark by Hannes von Döhren and Christoph Koeberlin, Manteca by Emyself Design, and Inter by Rasmus Andersson.


This is version 6.0.60 of my personal website. Older versions: v5, v4, v3, v2 (no longer online), v1 (no longer online).


Black lives matter.

© Dan Mall Teaches 2005–2026. All rights reserved.

Made proudly in Philly. Thou shalt not steal—but feel free to remix.

Privacy policy.

Dan Mall

Helping designers get their flowers.

45,601 followers

•

15,979 connections

45,100 followers

6,458 subscribers

•

79 videos

6,296 followers

•

221 posts

This site is typeset in Bueno by Rajesh Rajput, Thermal by Reset Type Studio, Mark by Hannes von Döhren and Christoph Koeberlin, Manteca by Emyself Design, and Inter by Rasmus Andersson.


This is version 6.0.60 of my personal website. Older versions: v5, v4, v3, v2 (no longer online), v1 (no longer online).


Black lives matter.

© Dan Mall Teaches 2005–2026. All rights reserved.

Made proudly in Philly. Thou shalt not steal—but feel free to remix.

Privacy policy.

Dan Mall