You’ve created an app that your beta testers use and your launch was a huge success. Every day the number of registered users increases and you feel like you’re making progress and moving closer to your goals.
It would be a mistake to hunker down and focus singularly on hammering out new features. Now that you’ve launched you have significantly more people visiting your homepage, creating accounts and using your app. You have much more data to help you prioritize what to do next.
I’ll focus on three important metrics that are extremely important and especially so after launching. These metrics are all conversion metrics and include 1) vistors creating an account, 2) becoming an active user and 3) becoming a paying user. While there are exceptions, the order is important: a user can’t become an active user without creating an account. Likewise, paying users are most likely active users.
Create An Account
What percent of visitors create accounts? There many variables that will influence the percent of visitors that create an account and they include:
Maybe visitors are arriving at your site searching for something else. If that’s the case you might have a high bounce rate (e.g. users coming to your site and immediately leaving). To fix this issue investigate ways to improve your SEO and ensure that your keywords are relevant to your product.
Clear Selling Point
Have you made it as clear as possible why your product is needed? Ask colleagues unfamiliar with your product to evaluate you’re site’s copy and it’s clarity.
Have you made it obvious how they can create an account? To test this you can use a service like Optimizely to A/B test different variants of your front page and your call-to-action button.
Maybe your pricing page is too complicated, or your prices are too high, and as a result users don’t even bother creating an account. This is an issue that you should look into before you launch. Ideally you should speak with an appropriate number of potential customers to figure out if your price is within a reasonable range.
Become An Active User
What percent of users become active users (e.g. use your app every day/week)? Your definition of an active user is dependent on your service. If you’re creating a new email app, or a social network, an active user might be someone who uses your app multiple times per day. If you’re creating a tax app, an active user might be someone who uses it once a year for a week. In order to measure active users first you must define who they are (duh!).
Problem: users create an account and then never use your app again. It could be that you have a very interesting idea, but your implementation is off. Likewise you could be attracting a lot of curious users that have no intention of actually using your app, but just want to see what the fuss is about.
I’d recommend following-up on a user if they haven’t used your app in a week after registering. Ask them why they didn’t end up using it, and what *one* feature they’d need to use it. Your response rate will likely be between 5 - 10% but hopefully that should be enough to pick up trends. If possible I’d automate this step to ensure that emails are sent out consistently and you don’t waste your time sending out copy and pasted emails.
Problem: they use your app for a couple of weeks and then never use it again. While still bad, at least you have more data to work with. Check your database and determine what parts of your app they were using. For an email app, were they sending emails, but not creating contacts? Maybe creating contacts was too tricky and they gave up and stopped using your app.
Analyze your data carefully and see if you can pick patterns that can point you to areas that need to be improved. Alternatively there could be temporal patterns, for instance, people who only use your app once a week quickly stop using it relative to those that use it several times a week. Finding patterns of use and comparing active users to non active users can shed further light on potential problems.
Become A Paying User
What percent of your active users become paying users? Comparing patterns of use between users who’ve converted, and those who have not, in the same cohort is useful for elucidating causality in conversions.
Unlike the previous step you should have much more data to conduct your analysis. More data means that it should be easier to find statistically significant patterns, but it may be more challenging and time consuming to do analysis. I’d recommend generating fewer than five hypotheses before you start your analysis. This will both limit the complexiy of analysis and reduce the multiple comparison problem (e.g. with enough comparisons you’re bound to get significant differences that are due to chance alone).
Analyzing what your users are doing, why some are creating accounts, why some are becoming paid users and why others are not is extremely important. If you love coding, and adding new features, it can seem like a big time waster, however, doing the above will help you prioritize better.
Although I haven’t tried either myself both Kissmetrics or Google Analytics’s Conversion feature should help you with the above. It’s crucial to be able to quickly determine why some users never become active, or never become paying users. Use hypotheses, data and outcomes to determine how to spend your time efficiently. Time is something you have little of, use it wisely.