If you want to find people who are actively solving a problem — not just browsing, but actively solving it — Hacker News is one of the best places to look.

The signal-to-noise ratio on HN is unusually high. The audience skews toward technical founders, early adopters, and people with real purchasing authority. When someone posts an Ask HN or comments on a relevant thread, they are often describing a genuine problem they are working on right now.

The problem is that HN moves fast. A thread that's relevant to your product today might be buried under 50 newer posts by tomorrow. Most founders either check manually and inconsistently, or miss it entirely.

What High-Intent HN Signals Look Like

Not all HN activity is equally useful. The highest-signal patterns tend to cluster around a few types:

Ask HN threads: "Ask HN: How do you discover early demand before building?" or "Ask HN: What tools do you use to find early customers?" These are explicit problem statements. Someone is not venting — they are asking for a solution.

Comments describing a workflow pain: A comment buried in a product launch thread that says "we still do this manually and it takes hours every week" is more valuable than a top-level upvote. The person is describing a problem they live with.

Tool comparisons: "We tried X, it didn't do Y" often signals someone who has already bought something, evaluated it, and found a gap. That gap might be exactly what you solve.

Requests for recommendations: "Anyone know a tool that does X?" is the most direct demand signal available. The person wants to buy. They just don't know what.

Why Manual Monitoring Breaks Down

Most founders start by setting up a Google Alert or checking the HN search page when they remember to. This breaks down in a few ways:

Timing: HN threads move fast. Commenting on a three-day-old thread is rarely worth the effort — the conversation has moved on.
Coverage: A single keyword search misses adjacent terms. "Demand discovery" and "finding early users" and "customer signal" all describe the same problem.
Scoring: Not all results are equally relevant. A thread mentioning your keyword in a tangential context is not the same as someone actively looking for a solution.

The gap is between knowing the signal exists and having a structured way to surface, score, and act on it before it goes cold.

What a Systematic Approach Looks Like

The HN Algolia search API (hn.algolia.com/api/v1/search) is public, well-documented, and covers the full archive. You can query it by keyword, filter by date, and retrieve comments alongside their parent post context — which matters because a comment read without the thread context is often ambiguous.

A minimal monitoring loop looks like this:

Fetch: Query HN search API every 24 hours per keyword, with a date filter to only get new results since the last run.
Contextualize: For each comment, fetch the parent thread to understand what question the person was answering.
Score: Apply your ICP filter — industry, company size signals, the nature of the problem described. Not all mentions are worth responding to.
Queue: Surface only the high-scoring signals for human review. A good rule of thumb: if you wouldn't want to write a reply to it, it shouldn't be in your queue.

The scoring step is where most manual approaches fail. Without it, you're looking at a list of mentions that varies from highly relevant to completely useless, and spending equal time on both.

Turning Signals Into Conversations

When you find a high-intent HN signal, the window for engagement is usually 24–48 hours. After that, the comment thread is dormant and a reply is unlikely to get much attention.

A few principles that hold up in practice:

Lead with the problem, not the product. If someone posted "we still do X manually," your reply should acknowledge that X is genuinely painful, maybe share what you've learned about why it's hard, and only mention Glean in passing if it's directly relevant. The goal is a conversation, not a conversion.

Be specific about the context. Generic replies ("I work on a tool that might help with this") get ignored. Referencing something specific from their comment ("you mentioned the scoring part is the bottleneck — that's exactly the piece we focused on") demonstrates you actually read what they wrote.

One ask only. End with a single low-friction ask: "Want me to run a quick sample on your category and share the results?" Not a demo booking, not a pricing page. One ask.

The Compound Effect

The value of monitoring HN consistently compounds over time. Early signals give you a view into what language your potential users actually use to describe their problems — which improves your marketing copy, your onboarding flow, and your scoring model.

A founder who has been systematically monitoring HN for three months has something that no amount of user research surveys can replicate: a real-time map of the problems people are actively trying to solve, in their own words, at the moment they are trying to solve them.

That map is the foundation of founder-led sales that scales.

Glean monitors Hacker News, GitHub Issues, RSS feeds, and manually imported links from Threads and Quora. Signals are scored 1–10 for intent and queued for human review before any reply is sent. Try it free →