Once you get to bring in users (free or paid) to use your product, as discussed in my previous article, your next concern is related to the engagement of these users, that is, are they using the product? Are they solving the problem that the product is supposed to solve? How many times a day (or a week, or a month) your product is being used? For how long? How is it being used?
It is very important to find metrics to measure engagement. For instance, for a product to send e-mail marketing, some engagement and usage metrics are:
Note that each product has different engagement and usage metrics. Each product manager must select metrics to track his/her specific product.
Have you ever stopped and thought how many times a day do you use your cell phone? What do you do when you unlock it? WhatsApp? Facebook? Instagram? Can you tell that you are deeply engaged with these applications?
Promoting the engagement should be one of the concerns of the product manager. In 2013, Nir Eyal launched his book called Hooked: How to Build Habit-Forming Products, in which he explains the theory behind these products that just end up entering our daily lives. It is a great book to understand more about this theme.
There are some strategies that can help you to increase your product’s engagement and usage. These techniques are called lock-in.
Another very important metric is the churn, that is, the number of users and clients who are no longer users or clients. It is important to know how many are they, and the reasons why that happened because you need this data to improve your software product, to reduce the churn.
This number is very important in any company that has the business model of continuous usage, especially those based on subscriptions. It is usually measured as a percentage as follows:
Monthly churn = amount of clients who cancelled on this month / total number of clients of the last day of the month.
There is also the annual churn, that is calculated the same way, only by dividing the number of clients who canceled on a certain year by the total number of clients on the last day of the previous years.
Churn is a number that holds a lot of information but, for being only a unique number, it leaves several questions on the air. Iíve heard affirmations like this: “if the churn is 20%, in five months we won’t have any clients, so it is not worth to invest in promoting this product.” Although, tracking churn, in isolation, is not a good practice. For example, even though the churn stays on 20% for several months, even for more than 5 months, the total number of clients can keep growing. How? Just by having a higher amount of new users than cancellations. Promoting helps a lot to do so. Check out the example:
Although the churn is higher than 20% every month, the annual growth result was 73 new clients.
Why even with a high monthly churn it is possible to grow?
Two are the reasons. The first, that I’ve already mentioned, is that it is necessary to have a higher number of new customers than customers leaving.
The second one is that the churn varies according to the client’s age. It is common to find cases in which the churn is high in the first month because the client didn’t like the service and decided to cancel right away. Or on the third or sixth month, if the charge is every three or every six months. Some people call it premature churn.
Although it is common, the premature churn is something that can and must be reduced. You do that by:
The concepts of churn and engagement walk side by side because the more engaged a user is, the smaller are the chances of this user canceling the service. Therefore, a good way to foresee the churn of a given client is to track the engagement.
For example, if you launched a distance learning product and tracks the use of this product, you’ll probably see that the cancelation ratio is higher with customers who never attended a class. Review the previous topic on lock-in to see the tactics to raise the engagement and reduce the churn.
In recent years the terms data science, machine learning and artificial inteligence have appeared recurrently and abundantly. These terms are very important for product managers. No wonder I dedicate 5 chapters of my product management book to data and metrics.
As I commented in the previous article, the product manager should be a data geek, that is, a person who is always thinking about how to learn more with data. What is a person’s behavior in the months and days before you unsubscribe from your product? What about the behavior of a person who does upgrade? What is the behavior of a user who says they are satisfied with their product? And what is said very satisfied? If your product has multiple features, which one is the most popular? Which generates greater satisfaction? What is the typical usage pattern of your product? If an atypical use pattern appears, what does that mean? These are examples of some questions that the product manager can ask and who will have their answers in the product metrics. And with each new response you get, it’s very likely that the product manager will want to ask you more questions.
In order to find the answers to your questions, it is important that the product manager knows data science techniques and how to extract the answers to her questions herself, either through data extraction and visualization tools or by running SQL queries in the product database. If the product manager does not have that independence and need to have other people extract the data for her, that may disrupt the product evolution.
As this learning from the data happens, it is likely that the product manager will begin to realize opportunities to embed such learning within the product. For example, a CRM software product manager may realize, after doing analysis with product usage and engagement data, that customers end up canceling less when they are using the business proposal generation feature. Once this discovery is made, it can promote a change in her product to make it easier and immediate to use that feature, and thereby decrease the churn of customers by making them more engaged. This is a way to insert data in your product.
Machine learning, which is nothing more than a way of implementing artificial intelligence, is when we program the machines to learn from the data and the more data the machine has in its hands, the more it will learn. It’s a way to insert data in the product to make it better. The more you use a particular product, the more data is available to the team that develops the product to know the user and how he or she uses that product. For example, the more you shop at a virtual store, the more it learns about your shopping habits and the easier it gets for store software to make recommendations that interest you. The same goes for Netflix and Spotify suggestions. In such cases, it is common for the store to compare someone’s use with the use of people who show similar behavior to make suggestions such as “who bought this item also bought these other items”.
This is why the product manager and the team that develops the product must know and know how to use data science, machine learning and artificial intelligence in their daily lives. They are powerful tools to help you increase the chances of building a successful product.
In the next article, we will move forward in the metrics theme, focusing on financial and long term metrics. We will discover the concept of negative churn, the “Holy Graal” of subscription business model products.
Do you work with digital products? Do you want to know more about how to manage a digital product to increase its chances of success? Check out my book Product Management: How to increase the chances of success of your digital product, based on my almost 30 years of experience in creating and managing digital products.