Founder-led sales: a 30-minute weekly routine from public signals

Most founder-led sales advice assumes you have time you do not have. It describes a process designed for someone whose entire job is sales, handed to someone who is also writing the product, answering support and doing payroll.

So this is deliberately small: one 30-minute block a week, plus up to three hours of follow-through, only when something is worth it. It is not a growth strategy. It is the minimum viable habit that reliably produces conversations, and it is designed so that a bad week costs you thirty minutes rather than a guilty backlog.

The premise

You are not going to out-volume anybody. A funded competitor can send four thousand emails a week and you cannot, and if you try you will lose on every axis that matters.

What you have instead is two advantages that do not scale, which is precisely why they work:

You can answer any question about the product immediately and correctly. No rep at any size of company can do this. In a public thread, that is the difference between a reply that reads as marketing and one that reads as the person who built the thing.

You can be genuinely useful before you are commercial, because you are not being measured on a weekly number. That single freedom is what public communities reward and what every automated motion is structurally incapable of.

The routine below is built to use both and nothing else.

Monday, 30 minutes

Minutes 0–10: read the list. Whatever surfaced in the last seven days: people describing the problem you solve, complaints about the tool you replace, companies whose situation implies they need you. Read it, do not act on it.

Minutes 10–15: mark three. Not ten. Three. The criterion is not "is this relevant" but "can I say something here that nobody else could say?" If the answer is no, skip it, even if it is on-topic. A generic reply from a founder is worth less than no reply, because it burns the one credibility advantage you have.

Minutes 15–25: check the context. For each of the three: read the whole thread, not the excerpt. Check the community's rules on commercial participation. Look at whether somebody has already given a good answer. If they have, and it is right, you have nothing to add and should move on.

Minutes 25–30: rate what you saw. Mark the ones that were good and the ones that were not. This is the step everybody skips and it is the one that makes week six better than week one: an unrated feed stays calibrated to a guess somebody's model made on day one.

That is the whole block. If nothing clears the bar, the week costs you thirty minutes and you go back to building. Zero weeks are normal and are not a failure.

The follow-through, up to an hour each

Only for the threads that cleared the bar.

Write the useful version first. Answer the question completely, as if you sold nothing, including how to solve it without any tool. If you cannot write that paragraph, you do not understand the problem well enough to be in the thread, and you have just learned something more valuable than a lead.

Disclose plainly. "I build one of these, so weigh that accordingly." It costs nothing, it pre-empts the most common reason comments get removed, and it makes everything else you say more credible rather than less.

Mention the product once, late, only if it is actually the answer. If a competitor fits their described situation better, say so. This is the single most persuasive thing you can do in public and everyone reading can tell whether you meant it.

Do not link unless asked. A named product is searchable; a link is what automated moderation removes. The rules that govern this are covered in how to find customers on Reddit without getting banned.

Then reply to the replies. The conversation after the first comment is where the buying decision actually happens. A comment that answers once and disappears reads as a drive-by.

What to measure, before revenue exists

Revenue is a lagging indicator and at this stage it is also a very noisy one. Two things are worth tracking weekly, and both fit in a spreadsheet:

Replies per week where you had something unique to say. If this is zero for three weeks running, either the configuration is wrong or your market does not discuss the problem publicly, and those have different fixes. If it is above five, you are lowering your bar and the quality will show.

Conversations that continued. Somebody responded to your response. This is the leading indicator that actually predicts revenue, and it moves months before anything else does.

Both are deliberately about your output, not the tool's. A feed that surfaced forty things you could not usefully reply to is not a good week.

What this does not do

Honest limits, because the failure modes are predictable:

• It does not produce forecastable volume. Some weeks there is nothing. If you need a number by month-end, this is not the motion. Paid acquisition is.
• It does not work if your buyers do not post. Developers, founders, marketers, recruiters, agency owners and local business owners discuss their problems publicly. Enterprise procurement does not. There is a fuller test in does social listening actually work.
• It does not survive being automated. The entire advantage is that a specific person who built the thing wrote the reply. Automating it removes exactly the thing that made it work, and on most platforms it is also against the rules.
• It compounds slowly, then matters. The replies stay up and get found later. That is real and it is not a first-quarter effect.

Where the tooling fits

The routine above works with no tool at all: you can run it with saved searches and an hour of reading instead of ten minutes. The tool is buying back the reading, not the thinking.

Openpulse is built for exactly this scale. You paste your website, it works out what you sell and proposes the vocabulary, you edit it, and it reads Reddit, Hacker News, X, LinkedIn, Facebook, job boards and review sites for evidence that somebody has the problem, judging each candidate against your business rather than matching a keyword, and recording why it discarded the rest.

The Go plan is the founder plan and is priced as one: $49 a month for one listener, three workspace members, daily or weekly runs, run-summary emails, one webhook and the MCP connectors, after a 14-day trial that takes a card up front. One listener is the right number for this routine. Five would produce more than thirty minutes can absorb.

If you want to see what it finds before paying anything, the free report runs it once against your website: one per email address, no card.