Metric-Driven Products
It is now probably more important than ever for product managers to work with data. Doing this could be daunting when considering the large amount that we often have to contend with. But it’s something we need to be at home with to deliver the right features or solutions.
Seasoned product managers make use of metrics to effectively guide the work of teams. The quantitative measures not only provide useful insights on the progress being made (if any) but can also help us to solve problems fast. However, building the right features isn’t as easy as simply tracking metrics. Product teams often struggle to use them effectively. Getting features right demands being more specific or zeroing in on what matters more.
Gut Feelings Aren’t Enough
Every product professional that knows their onions understands that they can’t afford to rely too much on impulse. But this is not to say there won’t be the occasional temptation to want to rely on intuition or instincts when making decisions. We could find ourselves thinking we already know enough – from experience – than to bother so much about what metrics could tell us. This explains why many people do not factor data into their decision-making despite acknowledging how helpful doing that would be.
One thing about gut feelings or instincts is that they aren’t always accurate. The numbers often provide a different picture. This makes it important to rely less on our intuition (along with potential biases) when looking to build a great product. Even if we have been successful doing things a certain way in the past, there is no guarantee that we would get the same results when we use them now or later.
The Problem with High-Level Metrics
Metrics help to assess how well our product is doing relative to our initial expectations. Are we on track or are we going off it? That’s one question we could find an answer to by using metrics. But to get the best out of these quantitative measurements, we need to find a way to dig deeper.
Many businesses are familiar with and use Dave McClure’s popular AARRR metrics, also called the Pirate Metrics. AARRR is an acronym for Acquisition, Activation, Retention, Referral, and Revenue. These are high-level metrics. While they can provide us a hint on how successful we are, high-level metrics aren’t that useful when it comes to figuring what to do better. They provide us a basis for assessment and comparison but not so much about what we can do to improve the numbers. High-level metrics tell us what is happening without helping us to know why.
Take the Revenue metric, for example. It would be clear that our revenue is falling or rising. But we are not provided with information on why the numbers are as they are. Several factors could be responsible for what we are seeing. If we do not have an idea of why revenue is falling, it may be difficult to identify what to do better. We could find ourselves spending weeks or even months trying to unravel the underlying factors.
A Better Way to Use Metrics
Based on the foregoing, there is a need for us to go granular when using metrics to decide the right features. We need to make the numbers we get more actionable. Making metrics more useful requires that we limit them to a more specific area. Retention or churn, for instance, may be too broad to address more effectively because of several factors that could be responsible. We could instead hone in on a specific factor that may help us understand what’s happening better.
We should be looking more toward specific experiences instead of the general picture to make metrics more actionable. For instance, we could track the number of key user actions per session or period – say, the number of clicks. This can serve as a basis of comparison between the customers that we managed to retain and those that churned.
A key indicator that could help make more meaning out of the churn metric is users per action or feature usage. This lets us know how many people are using a specific feature. Retention issues could come up if users don’t know, and so aren’t using, the valuable features on offer. They may consider a solution incomplete or not up to the task in such a case.
Specificity can make it easier to get our features and products right. It doesn’t mean that localized metrics tell us the exact problems. But they make it easier to pin down the issues as a result of the reduced scope.
