Best Practices

The Elusive Case of the Product KPI

Published Mar 11, 2018

“If you can’t measure something, you can’t understand it. If you can’t understand it, you can’t control it. If you can’t control it, you can’t improve it.” H. James Harrington

Product teams always start with the best intentions. In our case at CareerArc, that’s what guided us when we built our enterprise client dashboard. We wanted it to be useful for all of our users, so we filled the front page with all the stuff they could ever want: reporting summaries, activity feeds, featured resources, a content calendar, and more. You name it, it was there – everything but the kitchen sink (and that was coming in V2).

We did all this without looking at any data not only because we didn’t have time to, but also because we knew this was the correct approach. Everyone did it this way, so we should too, right? But when we took a look at what our users were actually doing, we were in for a humbling surprise. They ignored all the shiny buttons and widgets and charts that we had added and focused like a laser beam directly on a small link to access the full reporting page. Armed with this kind of data (and slightly bruised egos), we got to work on incorporating a new front page into our redesign – one that has the features and information that our clients really wanted all along.

As a product manager instinct is important, but flying by the seat of your pants and making decisions based on intuition is simply not enough. Building a ton of features is not a measure of success in and of itself, nor are good intentions. But you also can’t just measure haphazardly and hope to discover meaningful information – you need a strategy. What you choose to measure and how you choose to measure will have a great impact on your ability to make the best possible product. I’m hoping these tips will help you define your true product KPI, and then meet it. 

Start With Questions

Data is abundant, but the right metrics can vary from product to product, so the first thing to nail down is what defines your product’s success. What is your KPI? To do this, you’ll need to answer some key questions: What outcomes indicate that your product is working? How do you know that your users have accomplished their goals? What areas of your product do you want to optimize and improve?

Say you have a website that lets enterprise users manage a content calendar for social media posts going out to their various corporate pages. Success here means that your users have calendars full of engaging content. So you could answer the questions above by looking at how many posts your users create, how easy it is to create and schedule those posts (both in number of steps and time to complete), how often users are coming back in to work on their calendars, and how a user’s activity affects their likelihood to renew at the end of their contract.

Of course, your approach here could be totally different if you also offered free trials for your platform and wanted to know how effective they were, or if you had a direct-to-consumer experience that was free to use and wanted to know how your registration conversion funnel stacks up. If you start with key questions about what will make your product a success, it’s easier to pick out exactly what data you need to answer those questions.

Go Beyond the Usual Suspects

There are some obvious KPIs that you typically want to measure, whatever your product is. Generally, you’ll want to see how these stats change over time to gauge the health of your product and the effectiveness of your updates. The usual suspects are:

  • Unique users / logins
  • Page views / time on site
  • Bounce rate
  • Top referrers / campaigns

But the standard stats will only get you so far. Where it really gets interesting is when you start to measure things that matter for your specific products and users. When you see these going up and to the right, you know that you’ve made real progress towards your business goals.


If you have things that can be accomplished in your product, keep track of them. Examples of this include user-generated content and purchases. The number of these across all users and the average per user are both great stats that you can benchmark and aim to improve over time.

Funnel conversions

Similar to above, but specifically for multi-step processes that have drop-offs at each step. Where are the biggest drops and how do they change over time as you seek to address them? A/B test and iterate to watch your conversions improve.

In-app surveys

If you have the ability to ask users questions while they use your product, you’ll be collecting some very interesting data about what they want. This can be a double-edged sword, though – there’s a big difference between asking users about what they do (intent) versus looking at what they actually do (action).

Feature usage

Which brings us to the next point of feature usage, which is a specific subset of page views. Behavior doesn’t lie and is a powerful way to determine what’s working or not working in your product. This kind of data can guide you on how to improve existing workflows while also highlighting when a new feature is a hit or a flop.


Net Promoter Score is a valuable stat to track over time since it is designed to be comparable across all different types of products and services. Just be sure to ask as random a sample group as possible so that your results aren’t skewed. If you only ask people that like you, for example, then you won’t be learning anything! Be sure to follow up with everyone to get more context over why they feel the way they do.

Bring It All Together

At the end of it all, your data is only meaningful if you’re using it to measure the right things. Depending on your product and business, the things that matter most can be very different, from minimizing drop-offs in a workflow, improving feature usage, maximizing subscription renewals, onboarding users from novices to experts, converting visitors from anonymous to registered to paid, and many more. Once you have your goals in mind, you’ll be able to measure where you are today and how you improve as you make changes and upgrades. You’ll also be able to see if an experiment wasn’t effective and nip it in the bud before it has a negative effect on your business!

Product KPI Cheat Sheet:

  • Measure early and often. You may not know what’s important and what’s not right away.
  • Ask a question or form a hypothesis before looking at your data. Then, use your data to answer the question, or to confirm or deny your hypothesis.
  • Use these answers and insights to create a plan of action and provide the evidence needed for prioritization.
  • Execute your plan, test, iterate, and improve!