Hacker News lead generation

Where buyers show up on HN

• Ask HN threads: “Ask HN: What do you use for X?” The comments are a list of buyers and their constraints.
• Show HN comments: people ask “how is this different from X?” or “does it work with Y?” Those are evaluation questions.
• Migration and outage threads: “We are moving off X” or “X is down again” attract people with an active problem.
• “Who is hiring” threads: useful for market research, not lead generation. Do not pitch there.

The title is often a topic, not a question. The buying signal is in the comments, which is why a title-only alert misses most of the value.

Search comments, not stories

HN’s search is powered by Algolia. Search comments, sort by date, and filter by recency. The web UI is enough for a weekly check; the API is better if you want to automate it.

https://hn.algolia.com/?query=alternative%20to&type=comment&sort=byDate

https://hn.algolia.com/api/v1/search?query=alternative%20to&tags=comment&numericFilters=created_at_i>1700000000

• tags=comment searches comments; tags=story searches titles.
• tags=ask_hn narrows to Ask HN.
• numericFilters=created_at_i>… keeps results recent.
• hitsPerPage=100 is the maximum per request; paginate with page.

Timing is the whole game

HN threads decay in a day or two. A comment on a three-day-old thread is usually invisible. Check the timestamp before you write, and prefer the first 24 hours. If a thread already has hundreds of comments, your reply may be read by nobody.

Read the comment like a buyer

HN users are unusually good at stating constraints. That makes scoring easier than on other platforms: specificity is the signal, and a vague complaint is usually not a buying conversation.

Reply without getting flagged

• Add information, not a pitch. If the comment would be useful with your product removed, it is a good comment.
• Disclose that you work on the product, in the same comment, without a headline.
• Do not link unless asked. HN users will find the product if the comment is useful.
• Write in a plain register. Short, specific, no marketing voice.
• Do not argue about downvotes or ask for upvotes.

LaunchEngine’s drafts follow the same rules: no product names, no links, and no claims that cannot be checked. The draft is a starting point; the judgment and the posting stay with you.

A small, repeatable HN workflow

1. Save three Algolia searches: one for your problem, one for your category, one for the tool you replace.
2. Check the last 48 hours every weekday. It takes ten minutes.
3. Open the comment, read the parent thread, and decide whether you have something useful to add.
4. Write the answer first; add the disclosure second; add a link only if asked.
5. Log the thread and the outcome. Do not count impressions.

What to automate

HN is a good source for an intent system because the text is public and the constraints are explicit. It is a bad source for auto-posting: the community detects it, and the account cost is permanent.

LaunchEngine scans HN alongside Reddit and the open web, scores each post, and drafts a reply a human posts. You can also call the same data over MCP or REST if your workflow lives in an agent.