North Star Seduction: Why Startups Are Shunning Metrics-Led Product Strategies
“A North Star Metric is meant to keep a whole company focused on one number that stands in for customer value. The problem is that the easiest number to measure, engagement, often has little to do with the value you actually create, so optimising for it quietly pulls the team away from customers.”
Medium is the cautionary tale. Its North Star was "Total Time Reading", so it chased engagement: a recommendation widget that lured readers away from writers, then clickbait, then ads. Top writers left. Substack is the counterexample: it has no usage-based metric and instead ranks and solves writers' highest-priority problems, and it's winning because of it. The lesson isn't to throw out analytics. It's that engagement data only measures what people do in your product, not what they think and need. Building a customer-centric company means measuring that second thing too.
“We’ve overwhelmingly used our wealth to make the world cheaper instead of more beautiful, more functional instead of more meaningful.”
Substack is a black sheep in Silicon Valley.
Unlike almost every major startup, Substack doesn't care about a usage-based North Star Metric. They don't obsess over how many newsletters someone subscribes to, how many they read a week, or how often a writer publishes. They just prioritise Gross MRR.
That begs the question: why is a company focused on maximising revenue the outlier?
Nathan Baschez, Substack's former Head of Product, explains:
“The point [of Substack] isn’t just to make money — it’s to change the systems that human attention flows through… The Substack model [isn’t] just a business strategy, [it’s] a political philosophy.”
To see how this philosophy runs through Substack and its approach to product, start by comparing it with its ideological opposite: Medium.
The allure of engagement maximisation
Medium launched in 2012 with a mission to be the easiest way for anyone to publish their writing on the Internet.
It triggered a new wave of blogging by removing a bunch of technical barriers. At the centre was its recommendation engine: unlike traditional blogs, you didn't need an existing audience to get started. If your writing was interesting enough to grab attention, Medium would help more readers discover your stuff.
But to work out which articles were most engaging, Medium put an algorithmic widget on the side of every article recommending other posts. That was strike one: luring readers away from the content and writers that drove Medium's traffic.
To compete for these discovery slots, writers leaned harder into clickbait. By 2015, 7 of the 9 most recommended articles on Medium were clickbait listicles. Medium had become the dumping ground for "easy" content, not "great" content. Strike two: shallow content became the winning format, turning Medium into a distraction for readers, not a source of rich information.
Then Medium added advertising in mid-2015. Instead of pushing clickbait people thought they wanted, Medium now promoted content readers didn't want to see at all. Strike three: monetisation cut value for both readers and writers.
Almost as fast as they'd adopted it, top writers fled to personal blogs free of distraction and misaligned incentives. Medium built the easiest way for anyone to publish online, but forfeited its aim of being the best product for anyone, readers or writers.
What drove this string of poor strategic decisions? Medium's North Star Metric.
Medium's North Star Metric, and Substack's alternative
Like most startups, Medium's employees all worked to grow the company's North Star Metric: "Total Time Reading".
A 2013 article from Medium's then Product Lead Pete Davies shows pretty clearly how the team read this North Star:
“When a user engages with your platform, you have their attention. And attention is the precious commodity of the super-connected era. I think of competing for users’ attention as a zero-sum game. Thanks to hardware innovation, there is barely a moment left in the waking day that hasn’t been claimed… At Medium, we optimize for the time that people spend reading.”
To drive ad views and content discovery, Medium focused on maximising time on-site. But "Total Reading Time" has nothing to do with building the best blogging tool for writers. As so often, the metric was disconnected from customer value.
Substack has no usage-based metric because it isn't trying to maximise engagement like Medium. Instead, it focuses on solving the highest-priority problems writers hit when publishing online. Founders Chris Best and Hamish McKenzie make this writer-centric view SUPER clear in their 2017 post on the company's vision:
“Twitter makes money from your attention, so they need to compel your attention. Sometimes that leads to good things, like connecting you to people and ideas that matter. But it also means that the addiction, abuse, and outrage that thrive on Twitter and other social platforms may be impossible to eradicate. So what’s left to do? You can change the rules. That’s why we started Substack: when readers pay writers directly, it’s a whole new game.”
Substack's founders understood the big problems online writers face. They started with the biggest, the underlying monetisation model for publishing on the Internet, then moved down the "problem stack rank" over time, solving pain points like list ownership, recommendations, and distribution.
Substack is winning because it cares most about solving customer problems, and it didn't need a North Star Metric to do it: making online writers' lives easier brought in more great writers, which brought more readers, which fed revenue growth.
This doesn't seem hard to grasp, so why do so many companies fall into optimising for engagement over impact? Because engagement has become stupidly easy to measure.
Our collective addiction to engagement data
Everybody knows quantitative data beats opinions and anecdotes at any startup. Using data to drive decisions leads to better commercial outcomes and better cross-team collaboration. Data is the undisputed mother tongue, and language of power, in tech.
What most people don't realise, though, is that all of today's most popular product analytics tools trace back to advertising businesses.
The founders of Amplitude, Mixpanel and Heap all worked in social media right before launching their startups. Google Analytics, used by tens of millions of websites, is obviously owned by the largest ad-revenue company in the world.
These products were built to measure engagement because that's what advertising businesses depended on. They've basically spent the past few decades and countless hundreds of millions making that data collection as simple and smooth as possible.
So say you want to build a contrarian company like Substack. Instead of maximising user engagement, you'll focus on solving high-priority problems and increasing customer impact. Do you really think you'll pull that off with nothing but qualitative quotes and storyboards? Even on the founding team, that won't hold over time.
If we want to build customer-centric startups, we need new data that quantifies our customers' highest-priority problems and biggest unmet needs. That's where People Analytics comes in.
From product analytics to people analytics
“We shape our buildings; thereafter they shape us.”
Now, I'm not saying scrap product analytics. Measuring onboarding bottlenecks, activation rate, and feature adoption is essential for building a great product. But that's all product optimisation, not product strategy.
Product strategy is about marrying our company objectives to our opportunities to impact customers. For that, we need more than data on what people do in our products: we need data on what they think and feel when they pick our product to solve a problem. Only with that data can we build genuinely data-driven, customer-centric companies.
This is the mission we're working on at OpinionX. We see a future where product teams can measure, with real data, what matters most to their customers.
And we're making progress on it. Thousands of teams already measure customer priorities with OpinionX. Methods like Customer Problem Stack Ranking and The Discovery Sandwich are in use by teams around the world today.
But the part that gets me most excited isn't even what we've built so far.
I can't help thinking about what's possible when you combine stack-ranked data with a company's Customer Data Platform. Teams will have needs-based segments that power their entire marketing and tech stack.
Imagine powering your email campaigns, product personalisation, or proactive education with real data on what each customer cares about most. It'll be the first time real needs-based segmentation is possible at scale.
That's why I'm excited about OpinionX. Most software products today are just trying to siphon a fraction more of your time, attention and energy. But there's another future where our products help us accomplish more in less time. A world where people are the top priority, not the products.
Frequently asked questions
What is a North Star Metric? A North Star Metric is the single number a company rallies its whole team around, meant to capture the core value the product delivers to customers. Many startups pick a usage-based one, like time spent or actions taken.
Why do some startups avoid North Star Metrics? Because the easiest metric to measure, engagement, often has little to do with real customer value. Optimising for it can quietly steer a team toward addictive design and away from solving customer problems, as happened at Medium.
What went wrong with Medium's North Star Metric? Medium's North Star was "Total Time Reading". Chasing it led to a recommendation widget that lured readers off writers' pages, then clickbait, then ads, and top writers left the platform.
What should product teams measure instead of engagement? What customers actually think, need and prioritise, not just what they do inside the product. That means pairing product analytics with people analytics: data on your customers' highest-priority problems.
Daniel Kyne is the Founder and CEO of OpinionX, the platform for advanced market research surveys. Thousands of teams use OpinionX to measure what matters most to their customers, giving them real data to inform their big product decisions. Create unlimited free surveys at app.opinionx.co.
If you enjoyed this post, consider subscribing to The Full-Stack Researcher, our newsletter which shares actionable user research strategies with founders and product teams around the world.