One Metric to Rule Them All? The North Star Debate That Shapes Every Product Team

Advocates: Itamar Gilad (Gmail, YouTube, Microsoft), Sri Batchu (Ramp, Instacart)

Itamar Gilad -- former product lead at Gmail, YouTube, and Microsoft, now a product coach and author of Evidence-Guided -- builds his framework on the value exchange loop. Every organization delivers value to the market and captures value back. Measuring only one side is dangerous.

"The North Star metric measures how much value we create for the market. For example, let's take WhatsApp. WhatsApp for a very long time measured messages sent because every message sent is a little incremental of value for the sender, the receiver."

Itamar Gilad

▶ 00:27:02

Gilad makes a provocative claim: Lenny's own well-known survey of 40+ companies' NSMs actually captured Top KPIs (business metrics), not true North Star Metrics (value delivery metrics). What companies reported as "our most important metric" was their revenue or growth number, not a measure of user value.

"I would argue that what they told you is what is the most important metrics we measure? What is the number one metric for us? But it's not what I call the North Star metric."

Itamar Gilad

▶ 00:27:02

Each metric decomposes into a metric tree. Teams align around sub-metrics that map to their domain, and the tree enables cross-team prioritization and impact estimation. When Team B sees that Team A's project would move the tree more, they can choose to collaborate. The tree also informs team topology -- organizing around goals rather than software architecture.

Sri Batchu -- who led growth at Ramp (fastest SaaS company to $100M ARR), Instacart, and Opendoor -- adds operational sophistication. He uses two NSMs: one for volume/growth (monthly active orderers at Instacart) and one for efficiency (payback period at Ramp). The volume metric motivates; the efficiency metric prevents growth at any cost.

"I like having two. One is something around volume and growth and you want that to be, A, very motivating and intuitive for people to understand and also, B, something that the growth teams can directly impact, right?"

Sri Batchu

▶ 00:32:43

Batchu's real innovation is the translation layer. Sub-teams at Instacart owned specific metrics (app load time, search queries, cart-to-checkout conversion) that were translated into MAO-equivalent impact through regression analysis. This made cross-prioritization possible: compare MAO-per-engineer across teams to allocate resources. He is pragmatic about imprecision -- "it's like 70/30, 80/20, it's a guide" -- and updates translation factors every six months.

Advocate: Bill Carr (Amazon)

Bill Carr -- who spent 15 years at Amazon starting from its fifth year, launched Amazon Music and Prime Video, and co-authored Working Backwards -- reframes the entire debate. Instead of asking which output to measure, Amazon asked which inputs they could control.

The insight emerged from painful experience. In Amazon's early public years, when revenue growth slowed, teams would "run around like chickens with their heads cut off" trying to hit quarterly numbers with a month left. Bezos and the S-Team realized that output metrics are lagging indicators of past decisions -- they tell you nothing about what to do today.

"We then figured out ways to measure them creating a set of input metrics. And so then when we would develop our operating plans and review our business each week and set our goals, we were hyperfocused on those inputs and the input metrics."

Bill Carr

▶ 00:59:03

Carr's test for a good input metric: (1) Do you control it -- can you apply resources to make it better or worse? (2) Does it touch customers? (3) Can you measure it reliably? Amazon tracked selection breadth, delivery speed, page load times, and price competitiveness -- all directly controllable by teams. They even learned the hard way that their selection measurement was wrong for several years and had to be refined.

"So you know you're doing input metrics. If it is, do you control it? Meaning can you apply resources to make this thing better or worse? Does it touch customers?"

Bill Carr

▶ 01:03:30

Bezos and the S-Team reviewed these inputs weekly -- not the outputs. The operating faith was that if Amazon served customers well through controllable inputs, revenue and market share would follow. Carr is honest that this was an "article of faith," not a provable theorem. But it worked.

Advocate: Kayvon Beykpour (Twitter)

Kayvon Beykpour -- former Head of Product at Twitter, who co-founded Periscope and led Twitter's consumer product through its ranked timeline transition -- provides the strongest warning against single-metric orthodoxy. At Twitter, DAU was the north star. The ranked timeline drove DAU growth effectively. And that success calcified the organization's resistance to anything that might hurt DAU in the short term.

"If the thing that matters is driving DAU quarter over quarter, then that doesn't leave enough room for nuance to accommodate new speculative bets that might hurt one metric, but over time have other consequences that are positive and beneficial, like enabling an entire new vector of content creation and conversation on the platform."

Kayvon Beykpour

▶ 00:55:31

Twitter Spaces needed a prominent bar at the top of the app to succeed. That bar moved tweets down, moved ads down, hurt DAU, and hurt revenue. A metric-obsessed organization saw this as a cost, not an investment. Worse, the DAU focus created customer-hostile behaviors: teams left bad experiences in place because they were "good for metrics in aggregate."

The fix, when it came, was not to abandon metrics but to carve out explicit exceptions. Community Notes was structured like an internal startup, exempt from OKRs entirely. But Beykpour acknowledges this does not scale.