Since our launch in November, things have changed a lot for us at Colloq. Our service grew a lot, we gained new users, some of which pay happily for our service already to get extra features. And we built more features to match your needs.
But as we grow, things change. Our costs grew and we’re nearer to spending more money than we get from our customers. Some people argue this is the way a fresh company starts, but we weren’t happy about this. So we spend some thoughts on whether we can save money somewhere and let go two things we don’t necessarily need and feel better again.
When we made an inventory of service costs we suddenly realized that we stopped using one of our servers over two months ago but still paid for it and kept it running. We stopped it, made on configuration change and by doing so stop wasting money for things we don’t need.
The second thing is technically a bit more interesting. Until now, we used a hosted Continuous Integration (CI) service to run our tests and do the deployments. While we were happy with the service, the service costs increased a lot recently due to more and more tests and changes to our platform. But when we were forced to think about doing infrastructural changes to test PHP7.2 on the CI we realized that we’re spending a lot of money for a service that isn’t reflecting our environment at all. While the CI was running some PHP 7 version on a Linux host, we are actually running a different PHP version on FreeBSD. And this is only one of many differences that questions the usefulness of the CI reports. As a result, we researched what we can do and decided to cancel our paid plans at the service provider and will shift to a more pragmatic testing and deployment plan until we come up with a relatively simple and more reliable way to run automated integration tests and deployments. For now, we will simply install git hooks and run tests before a deployment can be made from each of our development machines.
Overall, it is interesting how quick one forgets about running costs and how little we think about saving costs, even if we don’t have money to waste. With relatively simple changes and little consequences we went from a subjectively unhealthy state into a healthy state again.
Interestingly, Jason Fried from Basecamp backed these thoughts entirely with his fresh published latest article Outlasting retroactively. I’m glad to read similar thoughts at other companies that validate our ideas which sometimes feel a bit excentric.
Photo credit: rawpixel.com