4. Community listening

§The question this chapter answers: how do people talk about this problem when nobody is selling them anything?

§Every interview you will ever run is shaped by the fact that you are in the room. People are polite, they perform, they answer the question you asked rather than the one they have. Communities are the one source where the problem is discussed in your absence, and that makes them worth an unhurried week before you say a word.

§ 4.1Where to listen#

§Wherever your target already gathers to complain. For developers: GitHub issues, Hacker News threads, the Discord of the tool they use, the subreddit for their language. For operators: industry forums, Facebook and WhatsApp groups, trade association boards, LinkedIn threads under a well-known operator's post. For professionals: Slack communities, the comments under a newsletter they all read. For consumers: Reddit, product review sites, TikTok comments, the app store.

§Two places are consistently richer than they look. The support forum of the incumbent tool, where every unmet need is written down by someone frustrated enough to type it. And the review threads of adjacent tools, where people describe using something for a purpose it was not built for. Those are the workarounds.

§ 4.2What you are collecting#

§Language. How does the person with the problem name it? Not your term, theirs. Founders say "inventory optimization." Restaurant owners say "we ran out of the thing on Saturday again." The customer conversations in the next chapter will be ten times better if you ask about running out of the thing on Saturday. Indi Young calls this listening for reasoning rather than requests: you want the sentences people use when they explain themselves to each other.

§Workarounds. Every real problem has a folk solution: the spreadsheet with the macro, the group chat where the shift is coordinated, the general-purpose AI model someone has taught to do a specialized job with a saved prompt. Workarounds are the strongest evidence of a problem that pays, because someone already spent effort on it. They are also your first competitor.

§Frequency and heat. How often does the problem come up unprompted? How angry are people when it does? A problem mentioned weekly with profanity is a different market from one mentioned yearly with a shrug.

§ 4.3Never pitch#

§The temptation is enormous. You have read forty posts describing exactly the problem you intend to solve, and you have a link. Do not post it.

§The first reason is that you will be removed, and deserve it. The second is that the moment you become a seller, everyone in the thread becomes a prospect, and prospects do not tell you the truth. The third is that you are not ready. You do not yet know whether the problem you read about is the problem they have, and a pitch commits you to a framing before you have earned it.

§If you must speak, ask. "I keep seeing people mention running out of stock on weekends. What do you do about it now?" is welcome almost everywhere. "I built a tool for that" is welcome almost nowhere, and it closes the thread.

§ 4.4When to come back#

§Communities are not a phase either. When you have a prototype, they are where you will recruit strangers. When you have a product, they are where the first hundred users will come from, one at a time, in the way Paul Graham described. But the sequence matters: listen first, ask second, offer last, and only when someone has asked you to.

§ 4.5What you leave with#

§A glossary, in the customer's words, of how the problem is named, with the phrases you will use in the next chapter's conversations. A list of the workarounds you found, each with a one-line note on what it costs the person using it. A rough sense of frequency and heat. And a list of the communities you will return to when you have something to show, which is not now.