Hacker News new | ask | show | jobs
by ransom1538 2850 days ago
1) Create the conversion funnel. EG: clicked ad -> installed app -> registered -> engaged with new user flow -> used product part a -> purchased product part b. What parts should they see and use all broken down. As they progress in the funnel they are more engaged.

2) Fill in all the stats of the conversion funnels. This will give you a baseline to answer: Did my modifications help or make it worse. EG: clicked ad -> installed app will be a %50 drop. installed app -> registered was a 20% drop. Amplitude works fine here.

3) Get a steady stream of users in. Ads, forums, social anything. Just a handful so you can see if you are changing things in 2).

4) Setup user testing. (UXCam <- makes VIDEOS of mobile sessions) This will let you see videos of EVERYTHING a person did. Watch each user's behavior 10 times. Over and over. Get to the point where you know what happens next.

5) This is the hard part. Make some change. See if it changes 2). Changes made in the early part of the funnel will change how many people TRY your app. Changes made in the later portion will change how many people STAY in your app.

6) Do 5) until you give up. Over and over. Then if you can't make 2) better quit.

7) If your cost of user acquisition is less than your profit. Run to a VC as fast as possible and pour gas all over it [or not].

2 comments

5) seems scary to me. Isn't that something you would need the user to opt in to? Or does EDPR not apply to apps?
GDPR does not prohibit tracking. GDPR prohibits not disclosing that you're doing it (and naming 3rd parties that you use).

I like to look at it as a rethink moment for when you build in the mechanism to track. Do I really need this? Is there data I don't need to collect here? Do I need a user's consent?

This is actually a very insightful comment. I don't have experience with the VC part at the end, but tight iteration and measurement loops are key, especially early on.