In this chapter, you will learn:
About the two parts of a North Star Metric, the name and the definition
How to define your North Star’s Inputs
How to test your Inputs
The Dance of Uncertainty
As teams converge on a North Star Metric, there is often a dance of uncertainty. Teams bounce among “Yes, that metric idea makes sense,” “But how would we measure that exactly?” and “Oh wait, I think what we really want is something else entirely.” As the idea solidifies, the discussion slowly turns to how they could define the metric.
Don’t get paralyzed by thinking that your North Star needs to be perfect. While it should be precisely defined, it will likely still change and evolve as you learn more. The key is to build a North Star that’s directionally accurate, and then adapt as you learn. Early stage startups (or companies new to the framework) might find themselves modifying their North Star Metric every six to 12 months. More established organizations shouldn’t need to change their North Star Metric that frequently. Every 12 to 36 months is more typical.
The Name and Definition
While the name and definition are important for metric and Inputs alike, pay particular attention to the name and definition of the metric. The name of your North Star Metric should be engaging and descriptive, something pithy and punchy that inspires the organization. The definition should be precise and clear, explaining exactly how you’ll measure it.
. “So make it dead simple to talk about, repeat it over and over, and generally simplify it to the point where a lot of your growth product roadmap is focused on moving the metric up.”
(some new reality you want to enable with your product). For a first pass, brainstorm the actual behaviors that might align with the concept behind your North Star Metric and Inputs. For example, if your product is designed to encourage regular meditation, you might observe prolonged streaks of daily meditation sessions.
When converging on your North Star Metric and Inputs, consider the following questions:
How will changes in this metric impact your decision-making? What decisions are you hoping to inform? If you have historical data, “try on” the various changes in the metric over the months or years.
What doesn’t this metric tell you? Are any of those questions or uncertainties more important to address?
Assume that all product development activities stopped. Would the metric increase? For how long? Why?
,” Web Performance Consultant Andy Davies argues that, “Your analytics are skewed. They’re biased towards visitors who’ll tolerate the experience you’re delivering.” How does your chosen metric account for this?
Will you eventually be able to generate this metric continuously or frequently with minimal effort?
How is the metric affected by seasonality, day-of-the-week, and day-of-the-month effects?
Is it possible (and helpful) to compare this metric across various account and user cohorts? Ideally, there is value in viewing it in isolation and across cohorts.
What will be your signal to revisit this metric? Assuming you revisit it periodically, what tests will you run to determine that it is still useful?
Some teams get stuck believing that only a rigorous, statistically refined model will suffice for their North Star Metric, or that not having one points to an irreparable flaw in their business model. Not true! With most products—especially if you’re thinking about long-term impact—the relationship between the metric and long-term impact are rarely that clear cut. Asymmetric opportunities often lurk where there is uncertainty.For example, Amplitude’s North Star Metric, Weekly Learning Users (WLUs), is defined as the count of active Amplitude users who have shared a learning that is consumed by at least two other people in the previous seven days (see “North Star in Action: Amplitude’s North Star Metric and Inputs”). You might wonder how we came up with exactly two or more users. Did we just pull that out of thin air? Or did we have a bullet-proof case for those precise numbers? The reality is that these numbers aren’t magic. A learning consumed by two other people is great; three is even better; one is better than zero. We’re more confident about some things and less confident about others.
The important thing is to identify thresholds that are both aspirational and achievable. Once you’ve done that, don’t get too worried about whether you’ve selected the perfect threshold. The important thing isn’t the exact number, but the combination of customer behaviors that our North Star represents. It’s consistent with our strategy, and we’re open about our uncertainty.
Getting Inputs Right
Dana Levine, Amplitude product manager, explains why Inputs are so important. “A big trap is when teams attempt to move the North Star Metric directly, rather than trying to move Inputs that roll up to the North Star.” He continues, “A good North Star Metric is the result of a combination of things that a team can do something about. Identifying those things, working on them, and seeing how the North Star Metric responds is far better than trying to game the North Star directly.”
Just like the North Star Metric, the Inputs should have both a name and a definition.
The following fill-in-the-blank template can help you determine the Inputs to your metric:
Mind-mapping is a helpful technique for identifying Inputs and metrics. Start by noting your North Star Metric, or even just a candidate for it, on a whiteboard. Then collaboratively note relationships and concepts in clusters.
Your mind map will undoubtedly start out somewhat messy, like this.
Then, refine the map by combining related concepts, eliminating extraneous information, and reinforcing the relationships between elements, like this:
Spend time editing and clarifying the concepts in your mind map so that the metric is at the heart of it and the contributing factors are all neatly named and defined at a high level, like this:
Once you have your Inputs defined at a high level, you’ll find it much easier to tackle the challenge of defining an exact metrics for each Input.
Testing Your Inputs and North Star
On the other hand, if the Inputs are too specific and prescriptive, you may struggle to identify innovative solutions to address them. For example, an Input like “Satisfied Customers” may be too broad. An Input like “Positive Reviews on Social Media” may be too narrow.
Here are a couple ways to test these Inputs:
Input Test 1: The Greenfield Test
For this test, forget about current roadmaps, missions, or other restrictions. Instead, focus on generating new ideas. Pose this question to the team:
You can use this format:
If the team quickly runs out of ideas, then you might need higher level Inputs. If they are swimming in ideas, many of them overly broad, then you should get more specific.
Input Test 2: The Roadmap Check
For this test, do consider your current roadmap or work in progress. Make a list of current initiatives and discuss how they could influence the Inputs you have chosen. In many cases, you’ll see a clear link. Sometimes, the link is implicit and contains a host of assumptions worth chatting about. Have these discussions! In some cases, you can’t see the link between the roadmap item and the Input. This is a good signal that your set of inputs is missing some factor of your North Star—or that you’re working on something that isn’t actually valuable.
This process of defining a metric and its Inputs often goes back and forth, and you may not go in this exact order. That’s okay! This isn’t a linear process. First, you converge on the metric, but then defining it becomes difficult. So you go back to the metric and adapt it a bit. Then you add Inputs and realize that you’ve gotten too specific or too broad, so you refine the metric again. This is all perfectly normal and even healthy.
Sharing Your Work
You’ve got a trove of material now: background research and data, decisions about the game you’ve played, notes from your workshop, and, most notably, your North Star Metric and its Inputs. To clarify the decisions you’ve made and discussions you’ve had so far, share your work with colleagues.
A simple way to share is to present your North Star work at an all-company meeting. But beyond that, we suggest you socialize the North Star repeatedly, in all sorts of forums, both formally and informally—and encourage your co-workers to share it as well.
Organizing the work you’ve done, presenting it to colleagues, and discussing the North Star across your company will help you prepare for your next challenge: putting your North Star into action.
Take Time to Celebrate
At this point, you and your team have defined a North Star, including both a metric and its Inputs. Congratulations! This is a good time to do something to thank your co-workers for their contributions and publicly celebrate your progress. We recommend cheese.
Chapter in Review
After your workshop, precisely define the North Star Metric and Inputs.
The metric should meet certain criteria; review the checklist to confirm you’ve selected a good metric.
The metric and Inputs should each include a name and a definition.
Inputs are the three to five factors that produce the North Star. The North Star acts like an equation or formula that is produced by these Inputs.
Be sure to share your North Star with colleagues and celebrate your successes.