Founder-led GTM,
from zero.

Before product-market fit, paid ads burn cash for a structural reason: you don't yet know which words make your buyers move, and ads charge you per experiment. SEO compounds too slowly to matter in your first year. Cold outreach lands in a folder. What's left is the one channel where the buyer starts the conversation: public threads where someone describes the exact problem you solve.

Every day on Reddit, X, and Bluesky, people post things like "is there a tool that does X?" or "how do you all handle Y, this is eating my evenings." A founder answering that thread isn't marketing — it's the most welcome message on the internet. It also does triple duty: it can win a user today, it teaches you the words buyers actually use (the messaging your future ads will need), and the thread keeps converting from search traffic for years after you replied.

The math founders get wrong: this channel doesn't scale with money, it scales with consistency. Three good replies a day, five days a week, is ~60 conversations a month with people who were already asking. Nobody's first 100 customers came from anywhere better.

Founders fail here in two opposite ways: watching nothing (and missing every thread that mattered) or watching everything (and drowning in keyword noise until they quit). The fix is a small, deliberate watchlist with three layers:

• Brand + competitor names — anyone mentioning you or the tool you replace is the warmest possible thread.
• Problem phrases — the words people use when they have your problem but don't know solutions exist (“drowning in support emails”, not “helpdesk software”). These outnumber solution-seekers 10:1 and nobody else is answering them.
• Watering holes — the 4–6 communities where your buyers actually hang out (subreddits, hashtags). Quality beats coverage: six well-chosen communities beat twenty random ones, because you learn each community's norms well enough to be welcome.

Pick problem phrases by listening, not brainstorming: read 20 threads in your space and write down the exact phrases OPs use. Your guesses will be wrong in revealing ways — that gap is also your messaging homework.

The fear that stops most founders is the ban. Here's the actual line, and it's sharper than people think: platforms ban automated posting and spammy content — not tools, not founders, not product mentions. Reddit's and X's enforcement looks at behavior: accounts that post faster than humans, identical text across threads, fresh accounts that only ever link one domain.

The heuristics that keep you permanently safe:

• You hit send. Never let software post as you — that's the bright line in every platform's rules, and it's also the moment your judgment and authenticity enter the reply.
• Answer first, mention second. The reply must be genuinely useful with the product mention deleted. If it isn't, don't post it.
• Disclose. “I build a tool in this space” costs you nothing and buys you trust + rule compliance in most communities.
• Respect per-community norms. Some subreddits welcome self-promo Saturdays, some allow disclosed mentions, some ban all of it. Read the sidebar once per community; it takes two minutes.
• Mention your product in maybe 1 of 3 replies. The other two build the account history that makes the third credible.

Auto-posting reply bots invert every one of these — which is why communities smell them instantly, downvote them, and report the account. The five minutes a human spends reviewing is not overhead; it's the moat.

You have 15 minutes; the feed has 200 threads. The skill that makes this channel work is triage, and triage is learnable. Before replying, score the thread on four axes:

• Intent — is the OP describing a problem, asking for a tool, comparing options, or just venting? “What do you all use for X” is gold; a meme about X is noise.
• Fit — is this person actually your buyer? A student asking hypothetically and a founder asking with budget read very differently in the same thread.
• Author — account age, posting history, whether they engage with answers. A 6-year-old account that replies to commenters is worth ten drive-by posts.
• Freshness — most replies earn their reads in the first hours, but high-search threads (“best tool for…”) convert for years. Old + high-search beats new + low-search.

Manually, this is the judgment you build by doing it badly for a month. Write your own rubric down — seriously, in a note — and score 10 threads a day against it. Your rubric converges fast, and it becomes the most valuable marketing asset you own: a written definition of who's about to buy.

Consistency dies when the workflow has too many steps. The unsustainable version: scroll five tabs at 11pm, paste context into ChatGPT per reply, lose track of what you sent. The sustainable version is a single daily batch:

• One pass, same time every day. Morning works: threads from overnight are still fresh enough to be early on.
• Triage first, write second. Score the batch (chapter 4), pick the top 3–5, ignore the rest without guilt.
• Draft from your product truth. Your reply quality is grounded in what your product actually does today — keep your own docs/pricing/changelog in front of you so you never promise stale features.
• Sound like yourself. Your real voice — typos in the bones, your phrasing — outperforms polished marketing English in communities every single time.
• Send from your own account, in your own browser, and log what you sent where.

Fifteen minutes. The whole system stands or falls on whether tomorrow-you repeats it, so cut anything — a metric, a tool, a step — that makes the loop heavier.

The channel produces two kinds of numbers, and founders track the wrong one. Upvotes and impressions are weather. The signal is: which replies turned into signups, and what did those threads have in common?

• Track reply → visit → signup per thread, even crudely. A spreadsheet with URL, community, date, and “did anything happen” beats no attribution.
• Attribute patiently — community traffic converts on a lag (people lurk, come back via search a week later). Judge communities on a month, not a day.
• Feed it back: double down on the 2 communities and 3 problem-phrases that convert; drop the ones that only produce upvotes.
• Harvest the words. Every converting thread contains your buyers' exact vocabulary — that's the copy for your landing page and, later, your ads.

This loop — reply, measure, refine the watchlist — is what makes month three dramatically better than month one with the same 15 minutes a day. The receipts also become your build-in-public content, which compounds the whole thing again.