Lighthouse Users Program
A lighthouse users program is a staged early-customer expansion model: begin with 10 named customers, expand to 100, then 1,000, with explicit success criteria gating each transition. Used by Atlassian’s Point A incubation teams — described by Tanguy Crusson in the context of building Jira Product Discovery.
Structure
Band 1 — 10 customers. Known by name. The team interacts with them individually. Engineers attend customer calls directly; this is how engineers develop the intuition Crusson calls ‘product engineers’ — people who understand user problems rather than just implementation. Success criteria are defined before the band begins, not after.
Band 2 — 100 customers. The team expands only after meeting Band 1 criteria. At this scale, the product encounters a meaningfully diverse user base while still allowing the team to track each customer’s experience.
Band 3 — 1,000 customers. Pre-beta scale. Expansion from Band 2 to Band 3 requires Band 2 criteria to be met. At this point the product is approaching the scale needed to evaluate retention and expansion economics.
The safety funnel
Running alongside the bands is a hard cap on total user numbers — a ‘safety funnel’. The logic: a bad first experience with a new product is hard to reclaim. Users who encounter a broken or confusing product form a durable negative association and are difficult to re-engage. They are also expensive to acquire a second time.
Better to have 10 delighted customers than 100 neutral ones. The safety funnel enforces this preference operationally rather than leaving it as a principle.
Why named customers matter
Working with 10 known customers by name produces qualitatively different feedback than working with 100 anonymous users. Named customers:
- Can be reached directly when a question arises mid-build.
- Are known well enough to distinguish their specific context from general signal.
- Become advocates if the product succeeds; their word-of-mouth carries credibility.
Crusson’s practice: engineers attended customer calls from the earliest stage. This made customer problems concrete for the people building the solution and shortened the feedback loop between observation and implementation.