New here? Learn about Bountify and follow @bountify to get notified of new bounties! x
← All posts

Making the front page of Hacker News and the value of good press

Posted by bevan on Nov 05

At 2:30am one night two weeks ago, I was surprised to see two users sign up on Bountify within a minute of each other. Bountify hadn't gotten a new user in several hours, so its sudden popularity was unexpected. I fired up Google Analytics, which revealed a steady stream of visitors coming from news.ycombinator.com (thanks @danabramov!).

Why I didn't post it myself

I was beaten to the punch. I'd been waiting for "just the right time" to post it myself. I didn't think Bountify could have made the front page in its then state because of its modest design (it uses Bootstrap elements) and the relatively small number of bounties posted. Those supposed drawbacks didn't seem to matter, so I probably could have posted it a lot sooner.

Traffic

The HN post had two upvotes when I saw it on the /new page. I upvoted the link and waited. I half hoped it would drop off the /new page quickly so I could go to bed; I'd worked late and was definitely ready for sleep (2:30am).

No such luck. The post made the front page within minutes, hitting the #3 spot with 5 votes. I grabbed a pear from the kitchen and settled in for the long haul, my fingers command-tabbing between the real-time analytics display, my inbox, and the HN post, in a chair I'd occupy for the next twelve hours.

Traffic blew up. Real-time users topped out at 200 and plateaued at around 100 (according to Google Analytics). There seemed to be bumps in traffic when the east and west coasts (of the US) got to work. After hitting the #2 position early on, it steadily descended the front page until the early afternoon. Bountify got about 8000 visits during its 12-hour stint on the front page.

Heroku performed well. Requests weren't queueing up, even with just 2 web dynos. Performance was "excellent" or "good" for the US and Europe (according to New Relic). I scaled it up to 4 dynos "just to be safe" (read: to feel cool).

Feedback

I got a lot of feedback from the HN post, much of which was positive. The most popular critical suggestion was to remove Bootstrap elements from the site, which I'm currently working on. I got several practical feature suggestions as well, most of which I've since deployed (such as a filter to view active bounties, an RSS feed, the "I'm working on it" feature, an "increase the bounty" feature, and others). There was also a bug report or two, which were thankfully minor. There was also some not-so-constructive pessimism about the service, but after getting a death threat for promoting a bounty on Craigslist awhile ago, the mild trolling didn't phase me. I'm very thankful for the feedback I received.

I'm glad Bountify had been tested for a few weeks before the link made the HN front page. Even a easy-to-fix bug would have been cumbersome to deal with during that time, because it was the middle of the night and I was a semi-flustered by the glut of traffic. Also, deploying a fix would cost a good minute of uptime, which could have costed several signups and left bad impressions. Thanks to Bountify's early users for helping get the app in shape.

Aftermath

The aftermath of the HN post brought in even more hits and users. Soon after the post dropped off of HN's front page, the British site SmashingMag tweeted Bountify to their 600,000 users, resulting in another cascade of hits (thanks @smashingmag!). There was also a few Facebook posts that day that generated lots of traffic as well.

A day later, the Russian site Habrahabr.ru wrote a thoughtful blog post on Bountify, which led to many more hits and signups. In fact, Habrahabr.ru's post generated less traffic than the HN post, but three times as many signups (I credit the favorable writeup). It's very rare for people to realistically consider the perspectives and motivations of both types of Bountify users (bounty posters and solution providers) before offering feedback on Bountify, so I was very grateful to have someone clearly explain those dynamics to a lot of people that I'd otherwise be unlikely to reach.

What I'd do differently

If I could go back in time to before the HN post I'd put more effort into optimizing the front page for signup conversions. The signup conversion rate from HN seemed low (perhaps only in comparison to Habrahabr.ru's rate) and I lost an opportunity to gather some useful data. A dynamic solution (like a multi-armed bandit A/B optimizer) could have been very useful. A bigger signup button would have been an improvement too.

Thanks!

The new users from HN, Twitter, and Habrahabr have been great additions to the Bountify community. More bounties have been posted and great solutions received. Thanks to those who provided feedback and who helped spread the word!

← All posts
Painlessly Crowdsource Small Coding Tasks with Bountify
  1. Post a bounty
  2. Verify solutions
  3. Pick a winner
About | FAQ | Sign up