Raised a seed round but no P/MF yet? Here's your work plan
Short answer: Go back to the market before you build anything else. If you have customers, read their usage, split them into three groups (power users, light users and the inactive), and talk to every one of them. Their answers tell you who your ICP really is and what value they're after. If you have no customers yet, test two or three ICP and value proposition pairs against each other and look for the pair where you're a lifeline. Solve the ICP and the value proposition first, then make sure the product satisfies them. The business model comes last.
You raised the round. The pressure is on. And the instinct is to build: more features, more marketing, more hires.
Don't. Not yet. First, find out what the market is actually telling you.
The plan starts with one question: do you have customers?
If you have customers
Step 1: Read the usage
Open your analytics. Look at retention and engagement.
I can't give you one magic metric, because it depends on the product. A weekly digest gets opened once a week. An annual tax tool gets used once a year. So first decide how often a customer should come back if the product works for them. Then check whether they do.
If it's a daily product: are they using it daily? Are they using the features you gave them? Nobody uses every feature, and that's fine. The real question is whether they're getting the value.
What you're after: what value users actually get from the product, and how often.
Step 2: Split your customers into three groups
| Group | What they do | What you need to find out |
|---|---|---|
| Power users | Use the product to its full extent. They get it. | Are they your ICP, and can they pay? |
| Light users | Do the first couple of easy things, then stop. | Is that all they need, or are they missing value they don't know how to get? |
| The inactive | Signed up, logged in once, did nothing. | Why did they sign up, and what was more important? |
Step 3: Talk to every one of them
If you're early and have 10, 20, 50 customers, reach out to all of them. Not everyone will accept. You'll get maybe 30 meetings. Run them in a week or two.
You'll come out way smarter. Nothing I can tell you will show you what's in their heads.
Step 4: Ask, don't tell
They should be doing the talking. Your job is to ask:
- Why did you sign up?
- What were you trying to achieve?
- Are you achieving it?
- Are you trying something else? Are you combining our product with something else?
- Did you know you can do this?
- What else could we do that would make the product more useful for you?
Two things you're trying to understand: where they're actually trying to get (not what your product delivers), and how they see your product as the way to get there.
Step 5: Read each group's answers
Power users. Highly motivated, so most likely your ICP. But check they can pay. If someone invests a lot of time in your product, either it's on their critical path, or their scope is small and this is all they do all day. Don't confuse a small customer with a great one.
Light users. They're investing time, so something is pushing them. They have a pain. Maybe what you already give them is enough. Maybe they don't know what else the product can do, or it's too complex to get there. Find the obstacle.
Sometimes they'll tell you: "I want to get from A to G. You take me to E. I use you to get to E, then I find my own way from E to G." That's fine. It's actually good news. It shows you what's missing, and it gives you a more honest message: "We get you to E" instead of a false claim that you get them to G.
The inactive. Ask first. Don't pitch. Why did they sign up, and what did they hope to get? Often what they wanted is exactly what you sold them. They just aren't doing it. That means they don't really have the pain. It's a phantom pain: they complain about it, they say they want it fixed, but it isn't their priority.
Real pain moves. Phantom pain complains.
Step 6: Name your ICP
Put the answers side by side, segmented by who said what, and the ICP shows up.
It's not all of them. The person who wants your value proposition but does nothing to get it is not your ICP. Maybe they're too small, or too big. Maybe someone else handles the problem. Maybe they don't own it. Whatever the reason, not your customer.
Here's more on how to define your ICP with signals, not labels.
If you don't have customers yet
Then it's a different story. You start by validating the market.
Write down two or three variations of the ICP you have in mind. For each one, match at least one value proposition. Then test the pairs against each other.
Be careful, because it grows fast:
| ICPs | Value propositions per ICP | Experiments |
|---|---|---|
| 2 (A, B) | 2 (A1, A2, B1, B2) | 4 |
| 5 | 2 | 10 |
So focus your energy. Run two, three, maybe four experiments, and benchmark one against the other. The goal is to find the ICP and value proposition that work best together.
The signals are traction. They say "yes, I want this". They come to your webinar. They jump on a call. They ask for a demo. They want to start a trial. They don't care that it's buggy, they want the beta, because the pain is that big.
The lifeline test
That's what you're looking for: being a lifeline.
You're a new startup. Why should anyone take the risk on you? Only if you're throwing them a lifeline. And when you are, they don't care who's throwing it. They just want out of the pit they're in.
If you're getting some traction but not enough, either move on to the next pair, or dig deeper and find out exactly what lifeline they're expecting. Your product always needs to be the lifeline. Otherwise they won't try you.
Keep the order
- ICP and value proposition first. Who it's for, and why they'd care.
- Then the product. Make sure it actually satisfies them.
- Business model last. You still need one, and the unit economics need to work, but it comes after the first two.
This is the same order as the four steps to P/MF. Change an earlier step and everything after it moves.
Common mistakes
- Building your way out. More features for the wrong customer just makes the wrong product bigger.
- Talking instead of listening. A customer interview where you pitch is a sales call. You learn nothing.
- Treating all customers as one group. Averages hide the power users, and the power users hide the ICP.
- Too many experiments at once. Five ICPs times two value propositions is ten experiments. You'll learn a little about everything and a lot about nothing.
- Fixing the business model first. Pricing can't save a value proposition the market doesn't want.
Your work plan
- Week 1: Pull usage and retention. Decide what "active" means for your product, and split customers into power, light and inactive.
- Weeks 1 to 2: Reach out to every customer. Aim for around 30 conversations. Ask, don't tell.
- End of week 2: Put the answers side by side by group. Write down your ICP and the value they're actually after.
- Then: Adjust the message and the product for that ICP. Only after that, work on the business model.
- No customers yet? Pick two to four ICP and value proposition pairs, test them against each other, and look for the lifeline.
Not sure whether you're there yet? Here's how to know you've reached P/MF.
Want a second pair of eyes on your plan before you spend the round? That's what a P/MF Sanity Check is for.