Saturday, 2 February 2013
Some thoughts about StatDiary
It's ages since I talked about StatDiary. In the last month there is some activity, so I share my experience with the long term maintenance strategy.
If you don't know StatDiary, it's a personal data tracking service. Since the beginning it gained some fans, which I'm the most happy about. There are many things tracked on the service - most are health and lifestyle related. Let me state here - I only know certain type of statistics when a user tells me. I usually ask them what they use the service for, so I can do the changes and always keep it in the right direction.
When I started StatDiary I was super excited, as anybody would be. I thought there will be millions of people tracking everything and I need to clone myself. It didn't happen, in fact, however there is a little 'community' I really like. There is 2 ways of contacting in case of errors or new ideas. I have my email and a Wufoo form which acts as a feedback/bug report facility. Almost all the contacts came from the Wufoo from and I'm pretty satisfied with that.
In general 70% of the mails were about feature request. In general I managed to implement 50-60 percent of them, where the request was general enough. The rest 30% was real bug report - where immediate action was needed. So far I managed to solve them all. The last I've got 2 days ago. Drupal update did something with sessions and Services changed it's behavior towards the mobile app. The result was failed login attempts. Pretty bad. But thanks to my super lazy organizing skills I'm sitting in the middle of London without any chance getting to my new apartment - so the coffee break was a great time for the fix.
I've made some statistics through the 1.5 year of StatDiary lifetime. It turned out people usually use it for a short term. I said usually, there is a small core who use it regularly (just like me). There are couple of hundreds of users, couple of thousands of data entered. Activity distributions mimics the very standard long-tail diagram. I'm not really a entrepreneur type. I don't have the passion to make business. However I like having the site. It costs $10/month - which is the server cost. And of course couple of hours a month for updates. This setup is pretty fine for me. People are using the service. There are new ones once in a while. And most importantly the few who use the site regularly seems to like it. Again - including me.
One interesting conclusion I realized is that the iPhone app that makes the most traffic. It's easier to get your phone and enter the daily stats, makes a lot of sense. However the current infrastructure is really decoupled. The app is for data upload and a minimal chart visualization. For the better displays you need to visit the website.
What I was always curious about is what outcome they expect from the system. Data tracking can be boring after a time. Got many feedbacks and now it's pretty obvious - people want to see correlations. They want to see how one part of their life changes compare to an other component. How the mood change according to their coffee intake. How their running performance goes with the sleep hours. At the moment there is a global mashup page, where you can see all your stats on one chart, and select which one you want to display. That's something - but not perfect. There has to be decoupled ranges and some dynamic feature. So definitely that's gonna be the next item in my todo list. When I get there.
There are several other parts of StatDiary which worth mentioning, I'll talk about them in another post.