Shoutt — Analytics - Keeping track of the pulse of your startup

A startup’s existence hinges on the ability to see what’s working (or not working) and then quickly improve your product based on what you learned.

That means you must MEASURE ALL THINGS. Well… not exactly. Although measuring everything sounds good in principle, it quickly overwhelms you and you are left floundering about trying to decide what the value of a particular metric means. Its not what you do in an early stage startup because all your focus is on finding the product-market fit for your product.

The importance of measuring things rather than just going “by your gut” cannot be understated. If you do not have the ability to measure a feature - do not ship it. Its important to elaborate further on what I exactly mean by measuring a feature. If you cannot make a decision on whether a particular feature is affecting the overall performance of your app/product then you are not measuring it correctly. Your metrics should always be actionable.

Its getting the last part that is extremely difficult for beginners. There are several wrong ways of measuring your metrics. Its also really easy to measure things so that it almost always confirms whatever “your gut said” (in scientific terms its called confirmation bias). Its equally easy to fall into the trap of measuring vanity metrics. These metrics make good soundbites… but they don’t really matter since they do not measure your product. A good example of a vanity metric is MAU (Monthly Active Users). This metric is often touted by many companies as a way to say they have a large and active user base but it doesn’t really matter because the performance of most mobile/web apps (not all) is dependent on DAU (Daily Active Users). There are, of course, exceptions to the rule but they are relatively rare.

First Among Equals

Your product will have one, maybe two, and in extremely rare cases, three, core metrics that are indicators of its overall performance. To determine whether that metric is a core metric or not, you should ask the following question:

In most cases, the metric is very specific to your product, so you need to have a very good understanding of what your product is trying to accomplish. For Twitter, a good candidate for that one core metric could be Number of Tweets per Active User. It captures the act of content creation that is critical for sustaining the whole eco-system. If the number of tweets generated per active user goes down, twitter is in deep trouble. Another good metric to measure would be the user retention, i.e. how many users come back after a certain number of time. It measures the engagement of a user, and indirectly the quality of the content being shared.

All the other metrics like the average number of followers for each user, number of photos attached, number of logins per day, length of a session for each login and so on, are secondary. They will not matter if the core metric is showing signs of trouble.

I have only talked about business metrics in this post. There is whole other set of metrics which is almost equally important. They are the technical metrics of your app. They are almost always interwoven with the business metrics. For example, the speed at which twitter’s page loads might directly correlate with the number of tweets generated per active user (Its a good to time to bring up the fact that correlation does not equal to causation). It is really helpful to be aware about these kind of relationships in your product. Getting into the nitty gritty of those kind of metrics and how they might affect your product is left for another post.

At Shoutt, we use analytics continuously to monitor the health of our product. I hope that this post was informative on how analytics matter and why its extremely critical to get them right in order for your startup to succeed.

Discuss on HN.

— The Grumpy Cat