What a North Star Metric is

The North Star Metric is the idea of a single navigational reference point for a product team.
Instead of dozens of metrics (CTR, time on site, pageviews, revenue, NPS — all at once), pick
one that:

  1. Reflects real value to the user, not just a business number.
  2. Acts as an early indicator of long-term growth — it moves before revenue does.
  3. Is measurable and interpreted the same way by everyone on the team.

The classic examples: Airbnb — nights booked; Spotify — time spent listening; Slack — messages sent
per month.

What an NSM looks like for e-commerce personalization

The job of a personalization platform is to increase the value of every visit. The NSM candidates:

Metric What it captures Weakness
RPV (revenue per visitor) Revenue, conversion and AOV in one number Sensitive to seasonality
Repeat purchases / 90 days Loyalty and the value of personalization Long feedback cycle
Revenue attributed to recommendations The direct contribution of personalization Depends on the attribution model
Share of logged-in users with 2+ orders The quality of the engaged audience Harder to measure

Tip: an NSM must not be easy to game. If the team can improve the metric in a way that creates
no value for the user — showing more popups to drive clicks, for example — it is a bad NSM.

How the NSM connects to A/B tests

The NSM sets the prioritisation criterion: run the experiments that could move it. This prevents
local optimisation — the situation where a test improves a micro-metric (button CTR) but leaves
the main goal untouched, or makes it worse.

Good practice: state the hypothesis explicitly on every A/B test record — “this test moves the NSM
through mechanism X”. If no such explanation exists, the test’s priority is worth revisiting.

NSM vs vanity metrics

Vanity metrics are numbers that look good in a report but do not correlate with growth:
– Total traffic (with no regard for quality)
– Pageviews
– Registered users (with no activity behind them)

An NSM should hurt when it falls and feel good when it rises — because it is directly tied to
how useful the product is to its users.