Show HN: What Actually Gets Upvoted (41,828 Posts Analyzed)

The distribution: extreme power law

Show HN upvotes follow an extreme power law. The average post receives 16.2 upvotes. The median is likely 2-3. Only 1.77% of posts in our 41,828 sample reach 100+ points. In the full 150,764-post dataset, 3.97% cross that threshold.

The top 0.1% of posts (roughly 150 out of 150,764) receive 500+ upvotes and generate front-page traffic for 24+ hours. Everyone else gets a brief spike or nothing at all.

A 2026 study measuring 188,085 HN posts found the median Show HN score is 2 points. A separate analysis of a 12-month window found 35.01% of Show HN posts never receive a second vote beyond the submitter's own upvote.

Title patterns that work

Successful Show HN titles share measurable patterns. Length: 50-70 characters (full visibility on mobile and desktop). Structure: "Show HN: [Name] - [specific capability]." Language: technical terms, not marketing terms.

Our HN model's highest-weighted title features include "open source," "sqlite," "graphql," "nginx," and "haskell." These are not recommendations to use these words. They indicate that posts naming specific technologies outperform posts using generic product language.

• Optimal length: 50-70 characters.
• Format: "Show HN: [Product] - [specific thing it does]."
• Include the technology or problem domain in the title.
• Avoid: "revolutionary," "game-changing," "excited to share."

Technical authenticity vs. marketing speak

This is the single strongest predictor in our HN model. Posts written as technical explanations outperform posts written as product announcements by a wide margin. The model assigns "excited" a positive coefficient, but this reflects a small subset of genuine technical excitement, not marketing enthusiasm.

The pattern is consistent: posts that link to source code, documentation, or technical blog posts outperform posts that link to marketing landing pages. HN readers click through to evaluate technical substance, not sales copy.

• Write your first comment as a technical README, not a press release.
• Link to GitHub, docs, or a demo, not a landing page with pricing.
• Explain tradeoffs: "We chose Rust over Go because..."
• Share what you learned building it, not just what it does.

First-hour velocity: the 90-minute threshold

Posts that reach 15+ upvotes in the first 90 minutes have a significantly higher probability of reaching the front page. Posts stuck at 3-5 upvotes after 60 minutes rarely recover. HN's ranking algorithm heavily weights early velocity.

This is not an argument for vote manipulation. It is an argument for having genuine readers ready when you post. Notify 5-10 people who would genuinely find your project interesting, not 500 people in an upvote group.

Categories that resonate

Developer tools and infrastructure outperform consumer products on Show HN by a wide margin. Open source projects outperform closed-source products. Technical blog posts about engineering problems outperform product launches.

In our full dataset, 47.2% of posts are classified as "high-traction" and 52.8% as "low-traction." But this aggregate number hides category-level variation. Developer tools likely have a 60%+ high-traction rate while consumer apps are below 30%.

• Strongest categories: dev tools, open source, infrastructure, databases.
• Moderate categories: productivity tools with technical depth, data tools.
• Weakest categories: consumer social apps, marketing tools, no-code platforms.
• Blog posts about technical problems often outperform product launches.

What the ML model learned

Our Hacker News classifier evaluates text features trained on 150,764 posts. Average prediction probability is 49.4%, reflecting the balanced class split. Top positive signals: "open source," "https docs," "filesystem," "ide," "sqlite," "graphql."

The model evaluates title and description text only. It cannot see your HN karma, GitHub stars, or posting history. A first-time poster with a well-written technical description can score as high as a serial launcher.

The Show HN playbook: 8 actionable steps

Based on the analysis above, here is a step-by-step playbook for Show HN launches. These steps are ordered by impact based on our data analysis.

• 1. Score your title and description at launchintel.pro before posting.
• 2. Write your first comment as a technical explanation (500+ words).
• 3. Post Tuesday-Thursday between 9am-12pm EST.
• 4. Link to source code, docs, or a live demo in your first comment.
• 5. Notify 5-10 genuine technical users who would find it interesting.
• 6. Respond to every comment within 15 minutes for the first 2 hours.
• 7. Load-test your server before posting (front page = 10k-50k visitors).
• 8. Capture emails from traffic spike. HN visitors rarely convert immediately.

After the spike: what to do with Show HN traffic

A successful Show HN post sends a traffic spike lasting 24-48 hours. Founders who capture value from this spike set up email collection, prepare onboarding for technical users, and follow up with commenters personally.

Founders who treat the spike as the outcome itself gain nothing lasting. The 742 posts that reached 100+ points generated millions of page views collectively, but page views without email capture or product activation are vanity metrics.

• Add email capture to your landing page before posting.
• Track signup source with UTM parameters.
• Email every person who commented to thank them and ask for feedback.
• Write a retrospective blog post 1 week later and share it on HN.

Related guides