Why Heroku?

Heroku is fantastic. I've been using it for years. It's reliable and hides all the stuff that you don't want to worry about as an app developer (failover, load balancing, process monitoring, backups, security patches, etc, etc).

Why NOT Heroku?

Heroku can be expensive.

To run a Rails app with 1 GB RAM ($34), a database ($9), a worker thread ($34) and SSL support ($20) will cost $97 per month on Heroku.

If you have a profitable app with many paying customers that is probably ok. I'd say it is worth $97 per month not having to worry about servers and instead focus on your what's more important - your apps, customers and business.

But what if your app is not making any money yet? And what if you have many apps like this?

If you know your way around a server, and have some time to configure and maintain it, you can save some money by doing it yourself.

Why DigitalOcean?

DigitalOcean is a relatively new cloud server provider (started 2012) and it has been getting a log of attention due to its all SSD storage and low priced servers starting at $5/mo.

To run a Rails app like the example above, 2GB RAM should be enough. That will be $20 per month on DigitalOcean.

To compare, 2 GB of RAM would cost $40 on Linode, and $44 on Amazon EC2.

Please note that I am not taking storage, bandwith, transfer costs or support level into account here. If you have special needs this needs to be factored in.

Better performance?

I compared load times from the exact same app running on Heroku and DigitalOcean. I did 30 requests to each service to get the load time of a page. The page in this case did a few database queries.

Surprisingly, DigitalOcean has 40-50% faster load times.

Heroku

Min:        0.538s
Max:        1.175s
Average:    0.811s

DigitalOcean

Min:        0.248s
Max:        0.798s
Average:    0.502s

Better uptime?

Too early for me to say. I'd have to give it 3-6 months to know.

How is my server set up?

It's a long answer. It requires another blog post or two to explain. Coming soon, I hope.

Conclusion

If you:

• Have an app without enough paying users
• Have time to set up a Linux server
• Think fiddling with servers is fun!

Then YES, a move to DigitalOcean can save you money today! If you use this link you get a $10 credit!

If you like this you should follow me on Twitter.

Note: I included affiliate links to DigitalOcean and Linode. It does not affect my opinions or recommendations of those services. Use the links if you want to support future blog posts like this.

What is runstar?

The runstar™ application can best be described as a running application for Android with a focus on simplicity and nice graphics. You can get more details about the app here.

The app was coded by me, in my company Bitbetter, and designed by my friends over at Color Monkey. We launched the free version at the end of 2009, then we took a short break, and launched the paid version with more features early 2011.

The free version has had 330 000 downloads with 23% active users. The free version has no income streams, but I still included it to show the amount of downloads, since they convert over to the paid app. Currently the conversion rate is 2.3%. However, that number is actually too low, since the free version has been downloading for more than a year more.

The paid version has had 8 000 downloads with 73% active users. It is priced at $4.99 and we have not changed the price during this time.

During 2011 we made $24 800 in revenue.

Earnings in 2011

The chart shows the accumulated revenue for the paid version of the runstar™ app.

If you look closely, you see a few distinct places where the the line bends up or down. Those places coincide with certain events that proved to have a large impact on sales.

Integrated with RunKeeper and Dailymile - June 23

On June 23, runstar™ released a new version of the paid app that supported logging into and uploading all runs from the runstar™ app to RunKeeper and Dailymile. This was also mentioned in a blog post here. And it was also twittered about here.

RunKeeper has a great iPhone and Android app. RunKeeper also built a platform to hold data, which they later opened up for third party users, like runstar™. Dailymile is another great looking and easy-to-use open running platform that has been around for a while.

Uploading runs online was a very requested feature that we did not have the time to develop ourselves. So when this was released, we saw a 100% increase in sales, and looking at the chart you will see that the line is a bit steeper (=increased daily sales) after June 23.

I’m pretty sure that the boost we saw was not only from just adding new features, but also hooking into the Dailymile and RunKeeper networks.

Featured on Android Market - October 8

On October 8 I found runstar™ featured on Android Market. Amazing! We saw an immediate surge in downloads and it is pretty obvious on the chart when that happened. Our revenue increased with 1000% that first week.

Yep, you heard me right. One. Thousand. Percent.

If you are interested in the details then read my other post on getting featured on Android Market.

Drops off Android Market top 25 list - November 17

As with all good things, they come to an end. Also the crazy runstar™ sales. On November 17 runstar™ dropped off the top 25 list in the health category on Android Market, which brought the sales back to pre-getting-featured-on-Android-Market-levels.

If I was a bit more prepared here, we could have done a sale or something to keep us up there, but we will save that for later.

Lessons Learned

Launching the paid version of runstar™ has taught me a few things:

An app is not a company. If you want to make a living from mobile apps, either build a platform and service with an app promoting it. Or build multiple apps. One app is not enough for me to live on, anyway.

Apps don’t sell themselves. Well, they actually do, but not as fast as they could. So any activity that exposes your app to new people, e.g. hooking it into Facebook or getting featured in an app store, was super important for me.

Next Steps

Even though it did not make me rich, I’m happy with runstar™’s first year. It made $24 800 in nine months, got featured in Android Market, and received tons of positive feedback from users.

But for 2012 I want to step it up a notch. My goal is to reach $50 000 for the year. It probably means some extra marketing activities. And possibly adding some features. And most likely introducing ads in some way.

I never thought ads could be a good source of revenue, but is seems to be working for Droid-Blog, KreCi and Making Money With Android. Very inspired by those guys, and I’ll try it out for runstar™.

Stay tuned for more reports,
Martin