Cohort Analysis & Churn

brandverity by on October 15, 2008 at 9:57 pm

Matt has an excellent post up on the blist blog on cohort analysis. I’ll post a teaser graph here to give you reason to click through and read the full post:

Cohort analysis and Churn at BrandVerity

We’ve been thinking through a lot of the points that Matt makes in his post as we try to understand a common signup pattern: A user sets up their trial account and then never returns.

Tire-kickers or Satisfied Customers?

All services have a signup drop-off, but it is particularly hard for us to write off those users as ‘tire-kickers’: their behavior is nearly identical to what our most satisfied customers do. We strive to provide only actionable information in our daily emails, so that people monitoring trademarks receive all the information (and only the information) they need to make a decision or take action. So satisfied customers rarely visit the interface and instead interact mostly with our emails.

I had a great conversation with Chris Sanderson at AMWSO earlier in the week that shed some light on a likely cause. He had been trialling the service, but hadn’t returned to the site since creating his account. As I walked him through the online interface, he helped me understand the gap between trial customers and paid accounts.

We provide a lot of depth in our emails, but that depth can be illusory. Chris had thought that the emails were the extent of the service and not only did he find the depth lacking, he also found the emails somewhat overwhelming.

Most of our paid customers had either received a walk-through of the online interface or had been with us as the interface has developed. They understand the depth of the service, and also have tweaked their settings to refine the alerts they receive.

So, what next?
Well, obviously we need to do a better job with our trial funnel. We need to introduce the service better over the trial period. We need more information on the service outside the signup page (we had assumed that since signup was so easy, sceenshots and product literature weren’t important). We also need to create a few bits of ‘feature candy’ that encourage deeper interaction in a fun way.

We’ve also given thought to increasing the barrier to signup (require a credit card, or a second step in the signup process). However, trials cost us very little and additional barriers don’t seem like the right step at this point - improving the conversion funnel does.

While we can’t do quantitative cohort analysis (not enough volume), we’ll definitely qualitatively evaluate each ‘feature cohort’ as we evolve the service.

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 2.5 License. | Dave Naffziger's BlogDave & Iva Naffziger