There is a dirty secret in product management. Most features that get shipped do nothing. Absolutely nothing. These features might sound can look great, sound fancy, and get a lot of praise internally. But when it comes to impacting the business, they do F all. They do nothing, not because they aren’t designed well, but because they don’t drive your business model in any meaningful way. The only way to impact your business is to have sustained focus on the things that matter. Without defining what matters, its easy for teams to scatter changes around the product, get lost in rabbit holes or micro optimizations that are a waste of time. This quote from Alice in Wonderland captures it perfectly: Alice: “Would you tell me, please, which way I ought to go from here?” Alice: The Cat: “That depends a good deal on where you want to get to.” The Cat: Alice: “I don’t much care where—” Alice: The Cat: “Then it doesn’t matter which way you go.” The Cat: Don’t be Alice. Direction matters. What is a North Star Metric? All the great software development companies solve this problem by defining North Star Metric. We did this at Uber and Codecademy. Meta does it, Google does it, Slack does it and you should do it too. A North Star Metric is a way of quantifying the combined user and business value that your product produces. Users come to your product to solve a problem. Your business makes money somehow. You need to find a way of measuring the leading indicator of both. Facebook - “Daily Active Users” - FB is designed to be a daily use product so this measures successful user engagement. FB makes money on ads, so user have to show up to see the ads. Uber - “Trips Completed” - They’re a marketplace, so they make money when a ride or delivery happens. “Completed” is key here because it drives value to both sides of the marketplace and captures liquidity + quality. AirBnB - “Nights Booked” - Also a marketplace so this measures value to both the host and the guest. Their revenue up if this number goes up. Slack - “Messages Sent per Team” - B2B, so revenue driven by accounts. The more active the account, the higher the retention. They also charge per seat, so the more messages, likely the more people in that account, the higher the revenue. Notion - “Documents Created” - Notion has both solo users and team users. In team environments, the real value is the creation of a document for other to read. The more documents, the more valuable the workspace, the higher the retention. Facebook - “Daily Active Users” - FB is designed to be a daily use product so this measures successful user engagement. FB makes money on ads, so user have to show up to see the ads. Uber - “Trips Completed” - They’re a marketplace, so they make money when a ride or delivery happens. “Completed” is key here because it drives value to both sides of the marketplace and captures liquidity + quality. AirBnB - “Nights Booked” - Also a marketplace so this measures value to both the host and the guest. Their revenue up if this number goes up. Slack - “Messages Sent per Team” - B2B, so revenue driven by accounts. The more active the account, the higher the retention. They also charge per seat, so the more messages, likely the more people in that account, the higher the revenue. Notion - “Documents Created” - Notion has both solo users and team users. In team environments, the real value is the creation of a document for other to read. The more documents, the more valuable the workspace, the higher the retention. No metric is perfect, but all of these orient product development towards growing the business and helping their users. Why Use a North Star Metric and Not MRR? Product development, if not measured carefully can be insanely wasteful. We learned this painful lesson at Codecademy. It was roughly in 2018 and we hired our first data scientist (shout out to Hillary) who correctly calculated our retention numbers for the first time. I remember staring at our fancy new dashboards and seeing that the last 2 years of our product development did absolutely nothing to improve our usage retention. Nothing. A flat line. We probably spent millions of dollars on engineers, PMs, designers, office space, snacks etc to get no movement in our retention, which is the main thing that matters. This taught me a valuable lesson. Good product development isn’t just shipping things. It also isn’t just shipping things that your users like. It getting the highest ROI for the dollars you put into developing your product. North star metrics give you a currently to denominate the value of your work in. Without a clear and thoughtfully chosen North Star Metric, you can’t quantify this ROI. Just optimizing around MRR can by very misleading as MRR is a lagging indicator. If you remember the concept of the growth ceiling from a few posts ago, your MRR might still be growing just from past momentum and not your current work. growth ceiling MRR can still go up even if your current features aren’t working. It can also go down when you’re shipping things that work. If you follow this as a signal, its really easy to go astray. How to Set a North Star Metric There are 2 basic choices you have to make here. What are you going to measure?  How are you going to measure it? What are you going to measure? How are you going to measure it? For “what to measure”, you should try to capture “one complete unit of value” from the user’s perspective. Typically value exists at a few levels and its tricky to find the right altitude. For example at DuoLingo, this could be the completion of a question, pack of questions, collection of packs, etc. The first time you do this, pick something that you can easily measure that signals user value and you can get more specific with time. For “How to measure it”, there are 3 traditionally ways of doing it. “Count of Something” - e.g Uber’s total completed trips per month  “Average of Something” - e.g Slack’s average number of message per team.  “X per Y” - e.g Facebook’s 7 friends in 10 days metric for new users. “Count of Something” - e.g Uber’s total completed trips per month “Average of Something” - e.g Slack’s average number of message per team. “X per Y” - e.g Facebook’s 7 friends in 10 days metric for new users. There are no perfect metrics and all of these involve tradeoffs. “Count of Something” - This number is sensitive to new user acquisition. So it might go up, just because you have more users not because you are improving the product. That might still be a good thing, but be aware of that.  “Average of Something” - This is great at quantifying user level health but it also hides outliers/power users. This is important in certain models like content creation, marketplaces, or anything involving usage based pricing.  “X per Y” - Setting thresholds are really powerful, but you need a lot of data to know you’re setting this threshold correctly. Optimize for the wrong level and you can go off course. “Count of Something” - This number is sensitive to new user acquisition. So it might go up, just because you have more users not because you are improving the product. That might still be a good thing, but be aware of that. “Average of Something” - This is great at quantifying user level health but it also hides outliers/power users. This is important in certain models like content creation, marketplaces, or anything involving usage based pricing. “X per Y” - Setting thresholds are really powerful, but you need a lot of data to know you’re setting this threshold correctly. Optimize for the wrong level and you can go off course. Advance Topic: North Star Metrics Drive Strategy The details of how you define this metric will also impact what strategy you take. Airbnb’s metric of “Nights Booked” has only a loose relationship to profit. They could drive much more profit if that was factored into their metric. Not focusing on profit means the teams can focus on expanding geographically as all nights booked anywhere in the world are equal under this metric, which means they can take more marketshare. If they set a metric of: “Profit per Night Booked” - this would likely favor optimizing the cities where they already have density.  “Avg Nights Booked Per User” - They would focus on high frequency travelers and addressing their specific needs. “Profit per Night Booked” - this would likely favor optimizing the cities where they already have density. “Avg Nights Booked Per User” - They would focus on high frequency travelers and addressing their specific needs. When I was at UberEats, we actively optimized for profit. To do that we shut down countries that we didn’t think would be profitable (Italy, Hong Kong, Argentina). Italy Hong Kong Argentina We definitely lost marketshare doing this but it was clearly the right thing to do for the business. So What Do You Do With This Information? North Star Metrics are a simple framework but deceptively hard to get right. Go from Broad to Narrow - Start with a metric that has broad coverage and narrow your focus with time. Make Sure Its Easy To Explain - your team can't optimize for it if they don't understand how it's calculated. Simple is better. Stay focused - in my experience the life cycle of a North Star metric is somewhere between 1-3 years. If you are struggle to move you north start metric, it might not be the metrics fault, you might be working on the wrong things. Go from Broad to Narrow - Start with a metric that has broad coverage and narrow your focus with time. Make Sure Its Easy To Explain - your team can't optimize for it if they don't understand how it's calculated. Simple is better. Stay focused - in my experience the life cycle of a North Star metric is somewhere between 1-3 years. If you are struggle to move you north start metric, it might not be the metrics fault, you might be working on the wrong things. Additionally, here are some good gut checks: Can everyone in the company impact this metric in some form?  If you are seeing this metric improve, does your MRR eventually go up?  If you see this metric improve, is your usage retention improving?  Is this based on a user action of some kind? E.g not just landing in your product. Can everyone in the company impact this metric in some form? If you are seeing this metric improve, does your MRR eventually go up? If you see this metric improve, is your usage retention improving? Is this based on a user action of some kind? E.g not just landing in your product. As in most things, simple is typically better. Good luck out there.

Slack

The Persona Process Is Broken—Here’s the Faster One

Newsletter

Read My Stories

The Dirty Secret of Shipping

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Annual Plans as a Churn-reduction Lever

Software Package or Specific Development

Top Benefits of Incorporating Machine Learning Into Business Processes

What Do Nobel Prize and Your Business Have in Common? Natural Experiments

Rethinking Data Leadership: Beyond Dashboards and Towards True Business Impact

Beyond Prediction: Econometric Data Science for Measuring True Business Impact

Annual Plans as a Churn-reduction Lever

Software Package or Specific Development

Top Benefits of Incorporating Machine Learning Into Business Processes

What Do Nobel Prize and Your Business Have in Common? Natural Experiments

Rethinking Data Leadership: Beyond Dashboards and Towards True Business Impact

Beyond Prediction: Econometric Data Science for Measuring True Business Impact

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps