The Next Heroku
...Might be Fly.io. Ditching free plans has made many consider alternatives to Heroku. We dive deep into which companies are best positioned to benefit.
On August 25, 2022, Heroku announced it will be ending its long-beloved free plans and moving upmarket, with a roadmap focused squarely on enterprise readiness (security, reliability, availability, and compliance). Personally, I think this makes total sense for Salesforce. Free plans are an excellent growth lever if you can monetize them well, but if you can’t, they are a drag on margins at best, and a distraction to product teams at worst. And, every developer already knows Heroku, limiting potential awareness benefit that would arise from continuing to provide the service for free.
The move cracks open a captive audience for a number of other companies building modern backends- and platforms-as-a-service. This Hacker News thread is just one of many conversations happening on Twitter, Discord, and in DMs. I wrote a quick script to figure out which companies are mentioned most as replacements for Heroku from that HN thread. That data is below.
: HN thread. Earlier-stage startups in orange. More mature companies in blue.
The top 10 are:
Perhaps expectedly, early-stage startups dominate the conversation. Outside of AWS, four of the top five most mentioned companies are relatively new entrants to the space, including Fly, Coolify, Render, and Railway. This is a testament not just to the talented teams, but also to the dire need for a better developer experience on the backend. The Hacker News crowd skews a bit more startup-friendly, but I would have expected to see mature companies (Digital Ocean and Cloudflare especially) mentioned more frequently in any conversation about Heroku.
Of the three large public clouds (four including Cloudflare), AWS being on top comes as no surprise. Azure sees paltry mentions, perhaps due to its relevancy at the enterprise-stage, but not with the hobbyist/SME crowd that AWS and Cloudflare are famous for. GCP is a close second to Cloudflare, which seems to confirm anecdotal evidence I’ve collected from dozens of software engineer friends that GCP is developer friendly, but not quite there yet in mindshare.
And then of course, there is a long tail of companies highlighting the extreme fragmentation in this space. Heroku won the mid-late 2010s due in large part to its’ slipstreaming of the zeitgeist around Rails. If you were building web apps on Rails, as so many developers were or were planning to in 2010, you went to Heroku. It’s not clear what this generation’s Ruby on Rails is. Maybe it’s the modern webs’ demand for globally distributed data and compute to service privacy regulation and latency aversion. Or maybe it’s that the current backend tooling stack has gotten overly complicated, driven by micro-optimizations at each layer.
Whatever it is, I am eager to use, and back, the next Heroku.
I did not distinguish mentions by unique users, so the same people could be dropping a products name in the same comment/thread, resulting in double counting.
I might have missed a few startups in the long-tail.
The HN community has some selection bias, in that the most popular tools are the ones whose posts routinely rank highly on HN.