How to Find Early Adopters for a New Business

Maciej Stolarski · August 24, 2026 · 9 min read

All posts

The early adopters box is where most Lean Canvases stall. People write a demographic, something like 'small business owners aged 30 to 50 who value efficiency', and move on. A demographic cannot be messaged, cannot be interviewed and cannot buy anything.

An early adopter is a person you can name, find and contact this week. This guide gives the test that separates them from the rest of the segment, a method for building a list of twenty names, the places they actually gather, a first message that gets replies, and the five questions worth asking once they answer.

Table of contents

The three-part test

A person qualifies as an early adopter when all three of these are true:

  1. They feel the problem now. Not last year, not in theory. In the last month they lost time, money or a client to it.
  2. They already pay for a workaround. A spreadsheet they maintain on Sunday evening, an assistant's hours, a subscription they half use, a nephew who is good with computers. Somebody paying nothing for a workaround usually does not have the problem yet.
  3. You can reach them this week. A group they post in, a conference they attend, a supplier who knows them, a public profile you can message.

Without the first you are selling to someone with no urgency. Without the second you have no price anchor, because the workaround cost is what your price gets compared against. Without the third the segment stays theoretical however well you describe it.

Write the test as three columns and score ten candidates. Those scoring three out of three are your first conversations.

From a segment to twenty names

Start from the customer segment you already wrote and narrow it with facts you can verify from the outside:

Narrowing dimension Weak version Usable version
Business type Service companies Single-location beauty salons
Size Small businesses Two to five staff
Location Poland Cities over 100 000 people
Behaviour Digitally active Publishes prices on Instagram and answers booking messages personally
Trigger Growing Hired a second person in the last six months

Each row cuts the list. Four rows usually take a category of thousands down to a few dozen you can actually see, the right size for a first round.

Build the list as a spreadsheet with five columns: name, business, where you found them, evidence they have the problem, contact route. The evidence column is the discipline. With nothing to write in it, the row is a guess and belongs at the bottom.

Twenty rows is enough. Ten replies from twenty is a good round, and five real conversations tell you more than a month of desk research.

Where they actually gather

The people who feel a problem badly leave traces. Look in these places, in this order, because the earlier ones carry evidence with the name:

  • Industry groups on Facebook and LinkedIn. Search inside the group for the problem in the customer's words. The complaints are your evidence column.
  • One and two star reviews of the workaround. Reviews of the tool people currently use name both the problem and the person.
  • Question threads. Reddit, Quora, national trade forums. Search the phrase a frustrated person types at 11pm.
  • Trade associations and chambers. Member lists are public and filtered by the criteria you narrowed on.
  • Suppliers who sell to your segment. A wholesaler or an accountant serving 40 salons knows which ones have the problem, and their introduction beats any cold message.
  • Conference attendee lists. People who paid to attend a session about the problem have declared it.
  • Job adverts. A business advertising for the role that does the manual work has the problem, has budgeted for it, and said so publicly.
  • Your own network two steps out. Ask five people who they know in the segment; each usually produces two names.

Two sources worked properly beat eight skimmed. Pick the two with the strongest evidence and work them until they run dry.

The first message

The message has one job, which is to get a twenty minute conversation. It is not a pitch, and mentioning your product costs you replies.

Four lines is enough:

Hi [name], I saw your post in [group] about [the specific thing they said].

I am researching how [segment] handle [problem] and you clearly deal with it more than most.

Could I ask you three questions about how you handle it now? Twenty minutes, no pitch, and I will send you a summary of what I learn from everyone else.

[Your name], [one line about who you are]

What makes it work: a specific reference proving you read what they wrote, a request for their expertise, a bounded time commitment, and something in return. The summary you promise costs you one afternoon and is useful to them.

What to cut: 'quick call', any adjective about your product, a calendar link in the first message, and the word 'opportunity'. Reply rates run 20 to 40 percent inside a group where you are a visible member, and under 10 percent cold. Becoming visible first is worth the two weeks.

The twenty minute conversation

Ask about the past, not the future. What someone did last month is a fact, and what they would do next year is a guess that flatters you.

Five questions, in this order:

  1. Walk me through the last time [problem] happened. Let them tell the story. Take note of the tools, the people and the hours.
  2. What did you do about it? This is the existing alternative, and it is your real competitor.
  3. What did that cost you? Money, hours, a lost client, a weekend. Push for a number even if it is rough.
  4. What have you tried before that did not work? This tells you which promises they no longer believe.
  5. Who else has this worse than you? The referral question. It doubles your list at zero cost.

Keep your solution out of it until the questions are done. If you must, give it one sentence at the end and watch the face rather than the words.

Three things end a conversation early, and all three are signals: no memory of the last occurrence, no cost estimate, or an immediate list of features to build. The third feels like enthusiasm and usually means someone who enjoys product discussions more than the problem.

Reading the answers

After five conversations, sort the evidence:

Signals that confirm. The same story appears in at least three interviews with different names. Somebody already pays for a workaround and can say how much. Somebody asks when they can try it, unprompted. Somebody offers to introduce you to two more people.

Signals that kill. Every interview describes a different problem. The cost estimate is always 'not much'. People say 'that sounds useful' with no example from their own month. Nobody gives you a referral.

The strongest single signal is a stranger who asks to be told when it is ready and gives you their contact details for that purpose. The second strongest is somebody who offers to pay before the thing exists, and that one moves you straight from research to a presale.

How many conversations before you decide

Ten conversations is the smallest number that tells you anything. Twenty is where patterns become reliable for a first decision. Beyond thirty, in an early segment, you are usually postponing the decision rather than improving it.

Set the threshold before you start: 'If seven of the first ten name the same problem and can put a cost on it, I build the smallest version. If fewer than four, I move the segment.' A threshold written in advance stops the interviews from becoming a comfortable substitute for building anything.

Run the round inside two weeks. Stretched over two months it mixes answers from different market conditions and loses the momentum that gets replies.

What changes on the canvas afterwards

Rewrite the canvas the day the round ends, in this order:

  • Early adopters becomes a description of the people you actually spoke to, in their language.
  • Problem gets replaced by the three sentences you heard most often, in their words.
  • Existing alternatives gets the tools and habits they described, with the costs they quoted.
  • Unique value proposition gets rewritten against the workaround, not against a competitor's website.
  • Key metrics gets one number that shows whether the promise landed.

Keep the previous version. Two canvases side by side, before and after the interviews, are the clearest record of what you learned. Our guide on how to fill in a Lean Canvas covers the order for the remaining boxes.

When early adopters and paying customers are different people

In some markets the person who tries first is not the person who signs. A practice manager tests the tool, the owner approves the invoice. A teacher adopts the app, the school pays.

Keep both in the canvas: early adopters describes who tries and reports back, revenue streams describes who signs. The first round of interviews goes to the person who feels the problem, the second to whoever controls the budget, because they answer a different question, which is what evidence they need before releasing money. Skipping that second round is the most common reason a product with enthusiastic users has no revenue in month six.

Five mistakes that waste the whole exercise

Pitching instead of asking. The moment you describe the product, people become polite and the data stops.

Interviewing friends. They want you to succeed, so they confirm. Use them only for practice runs and referrals.

Asking about the future. 'Would you pay 50 euros a month for this?' produces a number with no information in it.

Talking to whoever replies. The people with time to talk are often the ones with the least pressing problem. Score against the three-part test before booking.

Not writing it down within an hour. Notes made the next day keep the flattering parts and lose the awkward details, which is where the useful information sits.

FAQ

What is an early adopter in a Lean Canvas?

The narrower group inside your customer segment who already feel the problem, already pay for some workaround, and can be reached now. They are named people rather than a described demographic.

How many early adopters do I need before building?

Ten conversations is the minimum for a signal and twenty is where patterns become reliable. Set the decision threshold before the round starts, for example seven out of ten naming the same problem with a cost attached.

Where do I find early adopters with no audience of my own?

Industry groups, one and two star reviews of the tool people currently use, question threads, association member lists, and suppliers who already serve your segment. A supplier introduction converts better than any cold message.

What if nobody has the problem badly enough?

Move the segment rather than the product. The same capability often finds an urgent problem one segment over, and that is a cheaper change than persuading people who are comfortable.

See also

Put the names into the plan

Twenty names with evidence beside them change every other box in the plan. The free GenerateBizPlan generator builds the longer document from the same segment, problem and numbers, so the people you interviewed stay visible in the plan a bank or a co-founder eventually reads.

Need a ready-to-use business plan?

Generate business plan