← Back to blog

Launch traffic research · July 18, 2026

How much traffic do Product Hunt, Hacker News, and Lobsters send?

There is no honest universal average. Platforms do not promise clicks, founders measure different windows, and the launches that publish numbers are already a selected sample. What we can do is put the reported cases on one scale, preserve their attribution labels, and use them as test scenarios—not forecasts.

The short answer

Product Hunt300–10,000published visitor observations

The widest spread: a solo CMS launch, an AI wrapper, and an older SaaS outlier.

Hacker News500–6,493direct visitors in measured cases

Front-page rank matters, but product fit and the story still change the outcome.

Lobsters150exact visits in the clearest #1 case

Public case data is sparse; this is a technical community, not a launch directory.

The chart below uses a logarithmic axis because the smallest and largest observations differ by nearly two orders of magnitude. A dot is a published case, not a platform benchmark. The chatWise point is derived from the founder’s 2,058 seven-day visitors and 55% Product Hunt attribution; every other dot uses the reported visitor count directly.

Published visitor observationsLog scale · 100 to 10,000 visitors
Published visitor observations for Product Hunt, Hacker News, and Lobsters1001K10KProduct HuntHacker NewsLobstersWisp CMSchatWiseReply.ioAidlabUnixism · #21Unixism · #8Unixism · #7SyftenDots use published or explicitly derived visitor counts. Different launch windows are not normalized.

01 · Product Hunt

Product Hunt can convert, but the spread is enormous

Product Hunt is explicitly a launch platform for makers and early adopters. Its own guide highlights companies including Notion, Framer, and Loom. But a submission is not guaranteed homepage distribution: the current featuring guidelines say not every product is featured and prioritize live, useful, novel, high-craft products.

The three public funnels below span a solo-founder CMS, a small AI product, and an older SaaS launch. Bars are normalized within each company so conversion remains legible; the labels retain the actual counts.

Wisp CMS

Finished #6 with 420 upvotes. Its founder reported 300+ site visitors, 65 signups, and two paid customers during the 24-hour launch. This is the most useful small, contemporary solo-founder case in the sample.

Founder postmortem ↗
chatWise

Reported 2,058 unique landing visitors over seven days, 55% from Product Hunt, 510 app users, and seven paid customers. The author says the first-day numbers roughly doubled over the week.

Seven-day numbers ↗
Reply.io

In a 2017 interview, the founder recalled 10,000 Product Hunt visitors over two days, 600 signups, and roughly 60 customers. Treat it as a breakout historical case, not a current baseline.

Interview transcript ↗

02 · Hacker News

Hacker News is the clearest source of a sharp technical burst

Unixism published analytics for three articles that finished at #21, #8, and #7. Across the following seven days, direct Hacker News traffic ranged from 2,481 to 6,493 unique users. This small sample suggests rank matters, but it does not establish a conversion formula.

Unixism case sampleDirect HN unique users over 7 days
Direct Hacker News users for three Unixism stories by final daily rank02K4K6K2,481finished #214,336finished #86,493finished #7One publisher, three technical articles; useful for shape, too small for a universal rank-to-traffic rule.

Supabase’s alpha postmortem shows the high end of launch impact: after reaching the top of HN and remaining on the front page for more than 24 hours, it recorded 30,000 new website visitors and more than 1,400 signups during launch week. Those are total launch-week numbers, not clean HN referral counts, so they are deliberately excluded from the direct-traffic chart.

A smaller Aidlab Show HN reported 500+ direct HN visitors, about 6,000 total page views, 170 shop visitors, no direct sales, and four inbound enquiries. That is the useful commercial lesson: technically engaged traffic can explore deeply without buying immediately.

03 · Lobsters

Lobsters is smaller, focused, and not a generic startup launch channel

The cleanest published case we found is Syften: a technical article reached #1 on Lobsters and produced 150 visits and three signups. That is useful evidence, but one observation is not a distribution.

Syften’s #1 Lobsters storyPublished funnel
Syften received 150 visits and 3 signups after reaching number one on Lobsters150 visits3 signups2% visit-to-signup

A second example illustrates why referral claims get messy. The creator of Killed by Google reported roughly 4,000 unique users on the day, with Lobsters as the leading referrer at 37% of referral traffic. The denominator was referral traffic—not all visitors—so we cannot turn 37% into an exact Lobsters visitor count. We show it as contextual evidence rather than inventing a number.

Lobsters describes itself as a computing-focused, invitation-based community. Its rules say self-promotion should stay below 25% of a member’s stories and comments, analytics domains are banned, and tracking parameters are stripped. A deep engineering write-up, open-source release, or postmortem fits; a generic “we launched” page often does not.

Read the official Lobsters guidelines ↗

What a Railway founder should actually prepare for

Do not size a service from upvotes or from a seven-day visitor total. A static marketing page, an authenticated SaaS flow, a checkout, and an AI streaming endpoint turn the same visitor count into very different CPU, memory, database, and egress demand.

01Keep three traffic envelopes

Use a small featured-launch case, a strong front-page case, and a breakout case. The evidence above offers inputs for those scenarios without pretending one is “average.”

02Model the actual entry path

HN may enter through a technical post or open demo; Product Hunt through a product page; Lobsters through a write-up. Test the landing path plus the next action, not only the homepage.

03Measure the first hour

The published windows are mostly 24 hours to seven days. Your own analytics must capture the peak hour and operation mix before those totals become a capacity profile.

CapacityLab’s Railway workload model turns peak users per hour into a reviewable resource hypothesis. A controlled run then replaces that hypothesis with observed latency, errors, resource series, and a reproducible decision.

Model a Railway workload

Method and source table

All product numbers are self-reported by founders or company teams. We preserve the original time window and attribution language. “Direct” means the source explicitly named referral traffic; “launch window” may include secondary sharing, search, social posts, and unattributed visits. The sample is intentionally not averaged because its collection process is not random.

PlatformCompanyPublished resultWindow / attributionSource
Product HuntWisp CMS300+ visitors, 65 signups, 2 paid customers24h launch windowOpen ↗
Product HuntchatWise2,058 visitors; 55% attributed to PH; 510 app users; 7 paid7 daysOpen ↗
Product HuntReply.io10,000 PH visitors, 600 signups, about 60 customers2 daysOpen ↗
Hacker NewsUnixism · three posts2,481–6,493 direct HN users7 days eachOpen ↗
Hacker NewsAidlab500+ direct HN visitors; 6,000 total page views; 0 direct saleslaunch plus tailOpen ↗
Hacker NewsSupabase30,000 new site visitors and 1,400+ signups7-day launch window; not HN-onlyOpen ↗
LobstersSyften150 visits and 3 signups from a #1 storyauthor-reportedOpen ↗