I had a PM ask me recently what metrics they should choose for their main dashboard.

There's a ton of data available to most PMs today. You track pageviews, clicks, signups, etc.

The definition of a user is someone who uses your product. So, do these metrics mean someone is using your product? In some cases yes, but in many cases no.

For some companies, it might be harder than you'd think to figure out this key metric. Start by asking a few questions:

  1. Why would someone use your product?
  2. What's the key behavior that someone engages in when using your product?
  3. How often do people use your product?

Answering these 3 questions crisply will lead you to a metric phrased something like "our average user does X every Y days/weeks/months."

At Lyft for example, we would say "Someone uses Lyft to get from A to B. They do this by booking a ride. They do this on average 2 times per week."

Booking a ride is our key metric. Not website pageviews, not app installs, not signups. Rides.

It's also a good idea to track this metric for your best users.

  1. Who are the people who use your product repeatedly, with no prodding? No ads, no emails, no notifications.
  2. How often do they come back daily/weekly/monthly?
  3. And critically, do they share your product with friends?

One easy way to figure out who your best users are is by plotting retention rate against usage. That is, what % of users came back in week 2 vs. how often they used your product in week 1. Does higher usage correlate with stickiness?

Next you want to figure out what the threshold is for your best users. Look for an inflection point. For example, maybe 80% of people who used your product 3 times a week or more came back in week 2.

Dig into the data of your best users, and try to figure out what things they did that might have made them more likely to stick around in week 2.

But don't just look at dashboards. Talk to them. It's easy to believe that, with all the data we have access to today, you can just answer these questions on your own. In my experience you will save yourself a ton of time, and usually come across more surprising insights, if you just talk to 5 users every week. You don't need to survey 100 users and get statsig results, just talk to 5 people and any big things will jump out. Ask them why they signed up in the first place, what was easy/hard about their first few uses, and why they keep coming back.

What you want to do is guide brand new users all the way to being your best users, not just through the first few steps of your funnel. Don't optimize your onboarding into a super slick one-click wonder and then drop users into your product with no further guidance. Teach them what you're about. Encourage them to play around as they get more comfortable. Take your time.

Don't keep filling a leaky bucket. Figure out what your key metric is, talk to your best users, fix your bucket. Then you can pour gas on the acquisition fire.