The Raw Numbers of a Mediocre Product Hunt Launch


So a month ago, we launched the public beta of Braid on Product Hunt. For a minimally viable product with no external funding, we did pretty well! Over 170 upvotes in the first day, with hundreds of new signups, and small — but real — conversion to paid after the 30 day free trials expired.

In a previous post, I compared Braid’s Product Hunt launch to the 2009 TechCrunch post for the last startup I did.

In this post, I’ll go through all the numbers that we have so that other companies and founders will have a realistic idea of what a Product Hunt launch can — and cannot — do.

The numbers for the first 24 hours:

  • 2,181 total visits to Braid’s homepage (per Product Hunt data)
  • 2,099 unique visits to Braid’s homepage (per Product Hunt data)
  • 1,368 unique visitors to Braid’s homepage (per Google Analytics)

The Product Hunt numbers were generously provided by Ryan Hoover, founder of Product Hunt. Technically, the numbers from Product Hunt’s side are users that clicked on the GET IT button to visit Braid OR visited via Product Hunt’s unique short URL (https://www.producthunt.com/r/b56c18d57654cc/41051), which shows up on our side at https://braidhq.com/?ref=producthunt.

So the first thing to investigate is the 3:2 discrepancy between the traffic they sent and the traffic we recorded.

After digging, this is probably due to the fact that the traffic Product Hunt sent — even though a couple thousand isn’t a lot — is more than Namecheap’s shared hosting could handle for our marketing site. The Braid app itself runs off a much more robust, and much more expensive, Amazon Web Services instance. Another possible explanation is that the Segment Wordpress plugin we were using was acting up or unable to fire completely. (We weren’t using GA’s code at the time, but we’ve switched out Segment in favor of Google’s native code as a result.) This is probably because page loading times were so slow with the additional Product Hunt traffic that people bounced before the Segment Javascript could load.

But! Even with the slow site, we signed up 130 new users directly attributable to Product Hunt in the first day (about a 10% conversion rate). After a month, Product Hunt continues to send 10–20 uniques to our site per day, of which we continue to get around a 10% conversion rate to free trial customers. (This is defined by people who install the Chrome extension, navigate to Gmail, subsequently choose to create a Braid account, and authorize Braid to access their Google data. That’s a fair number of steps to signup.)

Braid has a 30-day free trial period (we don’t do freemium), so it took a month to see what the results really were. Turns out that the majority of people coming from Product Hunt liked Braid, but it wasn’t worth converting from free. (That’s to be expected, of course.) But we were able to convert the 130 users into a license for 10 paid users in one customer account. What’s interesting is that the customer paid for a license for ten users, but only two users inside the organization came from Product Hunt.

We’re going to take the feedback we got from surveying those who didn’t up for a paid plan and make our onboarding simpler and get to the “aha moment” faster. There’s also some interesting feature requests we’re investigating to find the true root motivation, and create feature hypotheses to hopefully solve those root problems.

So, in sum, Product Hunt is pretty awesome, but it’s not going to radically transform your company’s trajectory in any way. From my perspective, here are the wins and the deltas:

What we did right:

  • The product being good and matching the benefit you choose to hype matters a lot! People understood our core value proposition — that at some point, email doesn’t really scale. But a lot of people don’t really need full-fledged task managers for everything. So something that complements both Gmail as well as, say, Trello is unique and interesting. (In fact, we use Trello to build Braid’s product but use Braid to log and share all important non-technical developments.)
  • Once the Product Hunt listing was live, frantically emailing and Gchatting people about our posting made a big difference to get us on the front page above the fold, and from there, the Braid listing stayed all day. In fact, it seems that other social networks such as Twitter, Facebook, and LinkedIn posts did nothing.

What we did wrong:

  • I didn’t queue up folks in advance to tweet about our link, and I didn’t email our list because I thought it was spammy. Maybe these are noble intentions, but we didn’t crack the top five for the day so we didn’t make the daily email that supposedly brings another huge boost. (Algolia for Places in the morning and YouTube Creator in the afternoon took two of the available spots.)
  • I never explicitly asked people to upvote our listing. It’s against the rules, after all. However, I can’t help but think that it could have helped enough to get us into that top five. I’m torn on this one — it was the right thing to do but I also have an obligation to the company to do everything I can to make it successful.
  • I didn’t make sure the site could handle the load. I thought that we’d be able to render just fine with a few thousand visits, but the site did slow to a crawl. It also doesn’t help that Braid has lots of opportunity for compressing images and Javascript that we didn’t take advantage of before. (Now, it’s probably not worth it as the site loads reasonably and half our traffic is from the Chrome Store directly.)