By now you have most likely heard the phrase “build it, they will come”. This was introduced in the 1989 movie Field of Dreams but gained popularity in tech in the early days of the dotcom boom when the players were limited and early adopters eager to experience all the web had to offer. Back then, entrepreneurs would regularly receive this terrible advice (myself included). But the days of “build it, they will come” are long gone. I am not suggesting that identifying a real problem and building a solution to alleviate the pain-point should not be pursued. However, the barriers to entry are very low these days (i.e “There is an App for that”) and simply building a product doesn’t guarantee a successful business. Once you have a solution to a problem (i.e., a Good product), building a business is a science that requires methodical analysis of data to guide the product to its ultimate success (or as Geoffry Moore would suggest, refining the product to “cross the chasm” in order to get to the mass market — early and late adopters).
That’s where having an operational dashboard that provides “real-time” view of key KPIs is critical to ensure you are making product decisions that align with your users and business goals.
Before we talk about our experience building a dashboard (in an upcoming post), it’s good to have a primer on one key metric that is often referred to as a critical event.
What’s a critical event and how to set out to find it?
A critical event is an action (or chain of actions) that when performed satisfies a user need, increases the probability that the user will come back to the product, and contributes to the business (i.e., It aligns with your core value prop). For example in the context of an ecommerce app/website, one critical event might be a users making a purchase within the first 3/7/14 days after they have been exposed to the product. In this case, the user’s need to purchase was satisfied, and depending on how the experience was, the user would come back to the product and by performing this action revenue is generated for the company. For Facebook, the critical event identified early on was “creating a connection with 5 friends”. Meaning, activity and retention rate increased once users established their first 5 connections because the value of the platform suddenly became apparent and encouraged users to become active participants. Similarly at Uber, the critical event for the drivers side is completing at least 25 trips within the first 30 days. Once you have identified your critical event, you can then design your product, marketing and promotion activities to drive users towards performing the desired action or chain of actions. However, finding this critical event is not that simple and requires a good amount of analysis and experimentation (and some patience along the day).
Steps to finding your critical event?
- Keeping your product’s core value prop (review our previous post Improve Your Results with Clear Value Propositions) in mind, define a list of critical actions users can take within your product. For example for an ecommerce app, the list could be: Registration, adding credit card to profile, viewing catalog items, adding item(s) to the shopping card, making a purchase, writing a product review, etc.
- Form a list of hypothesis based on the events identified. For example: a) Users who add a credit card to their profile within x number of days increases probability of purchase by y%. b) Users who add as item to their shopping card make a purchase within x days
- Analyze your 3/7/14 day key KPIs for retention, sessions, activation and revenue for new/existing user cohorts based on the hypothesis you developed. For example: Analyze the purchase pattern of users who added a credit card to their profile. Are they more likely to purchase than those who haven’t? Do they come back to the app more often than those who haven’t? Are they viewing more items from the catalog than those who haven’t? How many days after adding a credit card do they make a purchase?
- Repeat step 3 for each hypothesis until you have narrowed down your choice of critical events
- Now that you have narrowed down your choices, run as many “cheap” experiments as possible to try to validate your findings. The intention here is to separate correlation vs causation because this can’t be determined from past data. For example, let’s say you find that users who add a credit card are more likely to make a purchase. Then you can easily send an email or push notification to those who haven’t added a credit card encouraging them to do so and observe their purchase pattern.
- Once you find your critical event(s), then align all your marketing and product activity to drive users to perform that action. Some examples: Onboarding/in-app tutorials; Push notifications; Email campaigns; Gamification (Example: 25% of your profile is completed. People in similar roles have 80% completed); Incentives (Example: Invite 10 friends and get points)
What to look out for?
- As you are analyzing your data, keep track of any other variables that may impact your analysis. For example, a product release with a new feature might impact your analysis so make sure to compare the right set of users
- Knowing your product usage interval is very important to this exercise. A product usage interval is the frequency with which you expect users to use the product. For example, a social app might be used every hour whereas an e-commerce product might be used once a month. Knowing this helps you analyze your data.
Not all events/actions are created equal. Just as defining the formula that drives your business connects the dots between the product and its business impact, finding your critical event/action connects the dots between the product and users. Before creating dashboard and reports, you have to know which events/actions within your product result in engaged users so that you can make investment in the area of the application that provides the most return.
The opinions expressed in these posts are solely my own and do not express the views or opinions of my employer.