Hacker News new | past | comments | ask | show | jobs | submit login
Mixpanel (YC S09) API + Ruby on Rails = Improved My Webapp (with code) (bingocardcreator.com)
49 points by patio11 on Aug 8, 2009 | hide | past | favorite | 4 comments



Great write-up and kudos for sharing the code.

Mix-panel looks great and I'm excited to integrate it into a couple of our rails apps. I think it will be especially useful for people like me who run marketplaces/payment-aggregation apps. If I can get better data on our customers' customers' buying behavior, that will be a win-win.


Great article!! Thanks for your continued contributions to HN. I'll test out Mixpanel with my new merb app in a month or so.

One design decision I've made to my funnel is to not force a user to sign-up too early. My user has states not_registered, registered, and premium (paying). You have these same states. You are simply forcing a user to become registered prior to creating a bingo card. Would you get better conversions if you moved the sign-up requirement to after creating, but before being able to save a card? In my, app (nothing to do with bingo) I'm betting I will.

In order to achieve this, in merb, I assign the user_id in the session cookie to a UUID. At some later stage, once they register, I map this UUID to the DB's user id. This also solves the issue of not exposing incremental db user ids in the cookie.


Would you get better conversions if you moved the sign-up requirement to after creating, but before being able to save a card?

Possibly, but I don't know if that would help the bottom line at all. I may test it at some point but that would require more code and more risk than most A/B tests.

You may have noticed I have a guest login option. You can play around with it to see what it does, but among other opportunities it asks people to sign up while waiting for their cards to print, etc.

I don't count guest signups as conversions for the purposes of e.g. AdWords, and it is a good thing, too. For targeted AdWords traffic I get a fairly high signup to the trial -- 28% or so. (That number still amazes me, since it is due to a change I made in this last week. Landing pages -- they work.) I also get another ~20% of people signing up as guests.

About 1.5% of trials go on to pay me money. And about 0% of guests go on to pay me money... or even upgrade to the trial.

This suggests to me that time spent making my service more attractive to people not really interested in it is unlikely to be time spent wisely.


Thank you for the article. Now to get off my butt and implement it.

P.S. I really love your writing style (easy to read, funny, informative)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: