We've actually had retention since we launched nearly three years ago, but this release gives it a lot more love.

The new stuff:

1. You can calculate birth cohort retention based on anything (signup date, date of first payment, app install date, etc). Basically, the grouping of your users is configurable.

2. You can filter and segment retention based on the properties you send. This lets you look at how things like A/B test version, Ad version, or user type impact retention rates.

3. The visualization is prettier!

We're really excited about how flexible Retention is now. We'd hacked in a little flexibility before, but this is dramatically better.

Ah. Sorry about that. I'd update the title, but don't have edit rights anymore.

I just got the email today and while I'm not accusing Mixpanel or anything, it was amusing to me how similar the screenshot looks like that of KISSMetrics. I'm currently a KISSMetrics customer but Mixpanel always looked like a good alternative/addition. How does Mixpanel's cohort analysis compare to KISSMetrics' in practice?

Just looked up the Kissmetrics site. They do look strikingly similar. http://www.kissmetrics.com/features/cohort-analysis

But how many ways can you really present something like that in a pleasing manner.

I assume they work almost the same. I'm on the Mixpanel side of things so would be interested in hearing how Kiss is in comparison. Anyone use both?

Yeah, getting people to use cohort[1] is really hard unfortunately. Which is really sad because, as Shopify[2] pointed out recently, without it you can't really answer basic questions, such as: "is retention going up or down?", or even "is our work having any affect on our retention?".

[1] http://techcrunch.com/2011/09/23/kissmetrics-helps-you-hone-...

[2] http://www.shopify.com/technology/4018382-defining-churn-rat...