I would argue that 99% of companies that are really good at developing tech products do these three things:
Let me also acknowledge that I have never seen a company that has 100% of its tracking set up correctly.
Something is always broken, out of date, or missing. It’s just the reality of tech products, unfortunately.
If this topic fills you with dread, it’s likely that your tracking infrastructure is broken somewhere. That’s fine, literally everyone’s is to some degree.
That said, every subscription company should have a basic hierarchy of metrics, understand what those metrics mean and try to improve them.
Regardless of the type of product that you run, there are 4 core things that happen in product development:
Needless to say, if you aren’t measuring the impact of your work, you can’t move to step 3 or 4 and just play whack-a-mole with new ideas and bugs.
Very early on in a startup's development, product development is guided by the instincts of the founders.
The founders have some sort of unique insight in the market and they use this insight to guide product development and ship a product. They are probably only looking at user sign ups, revenue and trying to move as fast as they can.
Over time, however, as the company grows larger, the founders need to spend more of their focus building the company and the market evolves, this instinctive style of development stops being as effective.
Companies at this point need to switch to a more structured approach, where they have product teams who set goal metrics and then try to ship projects to move these goals.
If you don’t have a clear hierarchy of the metrics that govern your product, it’s too easy to change your goals every quarter or let teams set goals that don’t scale up to the company’s goals.
The biggest danger that I see in growing $1M - $5M ARR companies, is that they have something that works, they mistakenly break it and don’t realize it for many months.
Best case scenario is that they figure it out and just lose out on many months of growth. The Worst case is that this kills them.
This is a basic hierarchy of metrics that the majority of subscription products can use, which breaks into 3 core categories.
Note here that I am only going to lightly touch on the core product metrics (activation, retention, etc) as those are highly dependent on the product that you’re creating.
I am going to focus instead on the “business” metrics and the revenue that the product produces.
I am going to break each group into:
This bucket is effectively attracting new users to your product and getting them in the door.
They start to vary from product to product, but I’ll list some ones here that are commonly useful.
This bucket is how well the product is retaining users. As you’re a subscription product, you always want to increase these numbers.
As noted in other articles, many products have a natural “cap” on retention based on the problem they are solving for users. That said, you should try to figure out that limit and get retention to that point.
Ideally, you are looking at this data on a cohort level. So for every month of new paying users, you should see how many months they stay around.
Usage retention will likely drive payment retention, so when someone stops using something they will eventually stop paying.
It’s your job to figure out how to keep driving usage of the product.
You should look at both forms of retention both as a table and as a chart. They’ll look something like this.
Most of these metrics are governed by the core product that your problem is solving. This is a framework that I learned in Regorge’s “Engagement and Retention” course, which is great.
I can’t go nearly into the depth that this course can, but on a high level, you’re trying to measure the path that users have to “retained”, which means they will keep coming back to your product.
This path goes like this:
Improving engagement and retention is the core job of product teams and is essentially a never-ending process.
Users will leave your product in a few ways and it’s important to measure cancellation according to these groups.
Before we get into metrics, there are two ways that users will leave your product
In the example of language learning, if I was paying for a product like Babble to learn Spanish, I either
Within the “Unhappy” bucket, I would break this again into two additional buckets
There is a lot in this post, so don’t get overwhelmed and try to implement everything at once.
As metrics always break, it helps to anchor to the sources of truth for the thing that you care about the most. These are going to be the key lifecycle events that are captured by other systems.
For the average subscription company, these are:
This is especially useful when your tracking data breaks or is suspect. It allows you to quickly hand calculate numbers like your free to paid conversion rate by taking your email sign-ups/ your new payments.
Pick the primary metrics that you want to track and ensure that whatever feeds your reporting systems roughly equals the SOT data.
These won’t be exactly equal, but they should be close.
So when you look at your free to paid conversion data, it should be roughly equal to hand calculated conversion rate.
Then, pick 1-2 operation metrics and do the same gut checks. You will likely find things that are broken and it’s time to start prioritizing things to fix.
Even though most of these numbers won’t move daily, you’ll be able to start to draw a relationship between the projects that you see shipping and the way that numbers move.
You’ll also be able to build your intuition for the seasonality of your business.
Typically when a company is not monitoring metrics, it is driven by one or more of these reasons
Figure out which of these are your core problems and work from there.