Why did Parse fail?

On January, 28th, Facebook announced the retirement of Parse. Parse was one of the leaders in the “mobile backend as a service” (MBaaS) space. It gave app developers a hosted alternative to building a custom backend from scratch for data storage and some logic around that data. Parse was acquired in 2013 by Facebook for $85 million and was powering 500,000 apps in 2014 according to Wikipedia. So what turned the success story into a failure?

Facebook did not explain why they retire Parse leaving room for speculation: What are the possible reasons for Facebook leaving MBaaS space?

  • While Parse was a nice way to get in touch with app developers, Facebook failed to make them use their social network. There was no incentive for Parse apps to use Facebook’s social network.
  • Parse may not have been profitable (enough). The business wan’t strong enough. But even if Parse wouldn’t be profitable, Facebook would have had enough money to keep it going for decades. It seemed not worth it because it did not help other businesses.
  • Parse wasn’t Facebook’s core business. Unlike Google, Amazon, and Microsoft, Facebook does have a could computing platform. For example, Firebase can make use of Google’s existing infrastructure. On the contrary, Parse with its Node.js and MongoDb stack were alien to Facebook’s technology stack.
  • Facebook now earns money with ads. Today is different than 2013 when Facebook acquired Parse. One day before the announcement regarding Parse, Facebook published their financial results for 2015. With a net income of $3.69 billion, Facebook has a solid business based on (mostly mobile) advertisement on which it will focus. But back in 2013, it was not the case yet. Facebook was looking for expansion eagerly to compete with Google’s and Amazon’s offerings.
  • Lock-ins scare people. Hosting is an external dependency which you cannot control. Big players are no exception to the try&error mentality. Think about how many Google products were discontinued because they were not successful enough. And Parse just showed us how right people are to be careful here.
  • Entrusting another company with the heart of your product may not be as hip as it was a couple of years ago. It might even be a factor of concern for end users. Today, the ownership of data and technology is a strategic advantage and can be a crucial factor of a companies’ value.
  • Parse’s technology was limited: While it’s great to build prototypes and MVPs quickly, developers may hit the limit once an app gets successful and requires more advanced features. When using Parse, you don’t have the power of a database like SQLite on the client and you were very limited to what you could do on the server. In performance tests we made, Parse was around 25 times slower than SQLite/greenDAO. It even crashed the app when handling larger amounts of data.

On the bright side, Facebook is making it easy for developers to “move on” as they call it. There will be a 1 year grace period for the hosted service, a migration tool and they open sourced the server code. This allows developers to continue with Parse if they are willing to do the hosting themselves. Open Sourcing Parse does not look like the new strategy for a bright future, however. It’s rather damage control done well.

Does Facebooks decision make sense? Will Firebase survive? Feel free to comment!

Update: Justin Beckwith, a Product Manager on the App Engine team at Google, just seized the opportunity and posted an article on how to run Parse to Google’s cloud infrastructure using App Engine (Node.js) and MongoLab. Well played, Google!



Spread the love
Posted in Cloud and tagged , , , , .