When is the best time to post on Hacker News?

How ranking works, and why timing matters at all

Hacker News explains its ranking in one sentence: the basic algorithm divides a story’s points by a power of the time since it was submitted (Hacker News FAQ). The same page lists other factors, including user flags, software that demotes overheated discussions, and moderator action. So a post needs votes early: the older it gets, the more votes it needs to stay up. More readers online means more possible votes, but also more posts competing for them.

What the published analyses found

Each used a different dataset and a different definition of success.

• Show HN posts, 2009 to 2025. Benjamin Derville at Myriade looked at more than 157,000 Show HN posts from the public Hacker News dataset on BigQuery and counted how often a post reached 30 votes, which was the top 10% (Myriade, July 2025). In its results Sunday did best at 11.75%, Saturday next at 11.08%, and weekdays ranged from 9.45% to 9.90%. By hour, the same analysis found 12:00 UTC did best at 12.2%, and 3:00 to 7:00 UTC did worst. It also found the pattern shifts in summer, when late evening US time does better.
• All stories, January 2018 to May 2019. A post on chanind’s blog used the same BigQuery dataset, counted a story as “frontpage” at 50 or more points, and measured the share of stories reaching it for each hour of the week (chanind, May 2019). That post found stories posted on Sunday at 6am UTC were 2.5 times as likely to reach the front page as stories posted on Wednesday at 9am UTC, and noted the tradeoff: a front-page post at a quiet time will likely get fewer views.
• All submissions, 2007 to 2014. Max Woolf downloaded 1,265,114 submissions from the official API (Max Woolf, February 2014). He found activity peaks around 12 PM Eastern on weekdays and that submission times were uncorrelated with the number of viral posts, concluding that good content matters more. In an update on the same page he notes that measuring the proportion of viral posts would favour weekends, since far fewer posts are submitted then.

Where they disagree

The best hour differs: 12:00 UTC across all days in one, early Sunday morning UTC in another. They also measure different things. A higher chance of reaching 30 or 50 points is not the same as more visitors, and chanind says so directly. And they cover different years; the site and its audience have changed since 2014. What they share is weaker: quieter periods, weekends in particular, give a small edge in the odds of a post taking off.

What matters more than the hour

• The post. Woolf’s conclusion and the size of the differences above point the same way: a few percentage points between days is small next to what the title and the thing you are showing do. How to write it is on the Show HN page.
• Being there. Pick a time when you can answer comments for the next few hours.
• A second chance. If a good post sinks, Hacker News moderators and reviewers pick overlooked submissions for a “second-chance pool” that places them, one at a time, on the lower part of the front page, and dang, one of the moderators, invites people to email [email protected] with suggestions (dang on the second-chance pool). The FAQ also allows a small number of reposts when a story has not had significant attention in the last year or so (Hacker News FAQ).

What to measure

Tag the submitted link, for example ?utm_source=hn&utm_campaign=show-hn, so visits from apps and digests that send no referrer still count as Hacker News. If you post twice at different times, give each post its own campaign name. Then compare signups and payments per post after a month, not the peak visitors on the day. That is the only dataset that really applies to your product.

With Nordva Launch

Launch’s analytics records each visit’s referring site and UTM tags, with a live view of the last five minutes. A signup through a Launch waitlist keeps the source it came with, your own form can link a person to earlier visits with identify, and with Stripe connected (read-only, every plan) a payment is traced to its first visit, so you can see whether the post brought paying customers. You can start free, no card needed.