You can't manage what you can't measure(blog.fray.it)
blog.fray.it
You can't manage what you can't measure
http://blog.fray.it/post/27834642686/you-cant-manage-what-you-cant-measure
2 comments
Feel free to comment,
Bas
Bas
http://blog.fray.it/post/27834642686/you-cant-manage-what-you-cant-measure
However here are few things I think worked well for us and might help other people make more informed decisions:
1. Track everything. It's important as you may never know what kind of information you'll need in the future. As every day pass and you're still alive (congratulation for not being part of the deadpool yet), you are constantly expanding your history. There were several occurence where we needed some kind of data and because we opted to track everything, we had that data available.
2. Talk to people. It's imperative that you build an entourage of smart people. People who've been in the trenches like you, people who are building something similar or build something completely different but in the same market. There will be time where you'll need to make some decisions based on your gutts and talking to people will give you that extra re-assurance you're doing the right decision (or not).
3. Research. I usually like to head to quora and search data that are related to my company. Sometimes I find nuggets there regarding some KPIs and then compare them to ours. You can tell when some of that data is bullshit but most of the data is accurate. I also like to ask closed community such as HN, Forrst or the various stackexchange communities sites for advices on certain things I'm working on. (twitter and facebook is good too, in short engage people on the net - and search for questions that have already been answered in the past).
4. Make assumptions and then challenge them. Some of the time you'll need to just follow your gutts and assume that something needs to be a certain way (that's true to startup who are building innovative products that have no comparison to draw from). You'll need to start from somewhere and do some A/B testing, as long as you go to the right and up (and preferrably in a hockey stick fashion), you're good.
5. Split your users into Cohorts and stop looking at vanity metrics. That's another subject and most likely one that needs a series of blog posts. But the basic is that you need to split your users based on events or thing they did in a certain period. Put your users into cohorts of month they signed up. See how they behave over time, correlate that data with new features you've introduced and/or promotion you ran. You'll soon find out that while you thought that your product is doing alright, some metrics are going downward and you need to find the root cause.
6. Go beyond Google analytics. It's somewhat related to my previous point. Basically Google analytics is a great tool (and it's free) but you can only get so far with it, you'll need to tie lot of that data you're tracking to users. That will eventually lead you to split users into different group and engage them differently. (It will also lead you to do a lot of conversion optimization based on the data you'll find).
I can go on and on, but I'll stop there and let other discuss/comment as well.