Tanguy Crusson on Incubating New Products, the Point A Framework, and Protecting Ugly Babies
Tanguy Crusson, Group Product Manager at Atlassian and creator of Jira Product Discovery, in conversation with Lenny Rachitsky. The episode covers HipChat’s failure and its lessons, the Point A incubation framework, the psychology of protecting new bets inside large companies, acquisition integration pitfalls, and the Jira Product Discovery journey from solo research to 8,000 customers.
Key ideas
- “Don’t eat your own bullshit.” The synthesis of Atlassian’s “open company, no bullshit” value and the practice of dogfooding: the most dangerous assumption is that what made you successful before will make you successful again, because you have direct experience confirming the playbook. HipChat assumed the Atlassian bottom-up adoption engine would work on business users the same way it worked on developer users. It did not.
- Competitive myopia kills positioned products. Once HipChat started fast-following Slack feature by feature, it abandoned what made it successful — depth of developer tooling integrations. The competitor’s features are the visible tip of an iceberg resting on 12 months of user research and strategic decision-making you cannot see. Reacting to features means you are always fighting last year’s war.
- Point A: four stages for new bets. Wonder (prove there is a market worth entering and articulate why now), Explore (validate solutions with customers, primarily in Figma, before writing code), Make (alpha, then beta, then generally available), Impact (monitor and grow the business). Every Atlassian team knows these stages; a team in Explore is not expected to have a production architecture review.
- Failure is the most likely outcome — frame it that way deliberately. Announcing internally that your new bet has perhaps a 70% chance of not existing in six months is not defeatism; it is a firebreak against the organisational processes that will otherwise consume the team. Once the bet is framed as possibly temporary, the rest of the company stops insisting on production-quality processes for something that may be paused.
- Why now is the question most new bets cannot answer. Great work can sit on a shelf indefinitely unless the case for urgency is compelling. Tanguy’s Jira for IT operations pitch was enthusiastically received and then parked for a year because no one could articulate why Atlassian had to enter that market now, not next year. The team that eventually executed it made the urgency argument explicitly.
HipChat: Three Failure Modes
HipChat was built on three interacting errors.
Competitive myopia. When Slack began dominating mindshare, HipChat reacted feature by feature — implementing what was visible on Slack’s surface without understanding the user research and strategy underneath. The result was a product that had abandoned its core strength (developer tooling depth) without meaningfully closing the gap on the consumer-grade experience that made Slack sticky for business users.
Segment misjudgement. Atlassian’s adoption engine worked through developer users who championed products upward into IT and management. The assumption was that the same engine would route through developer users into business users for messaging. Business users had different criteria — emoji, polish, modern UI, consumer-grade UX — that HipChat had not invested in. The assumption was never validated before the bet was made.
The rewrite trap. Atlassian’s engineering depth made a simultaneous platform rewrite attractive: rebuild HipChat on microservices so that the chat editor component could be reused across all Atlassian products. Rebuilding the plane mid-flight while fighting a market leader and validating a new user segment was too much. The code did survive — 70% of it is still in the Atlassian platform today — but for the HipChat bet specifically, the rewrite was a local optimum that cost the team time it could not afford.
Point A and Jira Product Discovery
Point A was Atlassian’s institutional response to the HipChat failure: a formal incubator to rebuild the new-product muscle. It offered psychological safety (teams borrow people from other departments; no one’s job is at risk if the bet fails), access to company resources (research, corporate development, analyst relationships), and a shared vocabulary (Wonder/Explore/Make/Impact) that let teams communicate their stage to the rest of the organisation without negotiating what that meant each time.
Jira Product Discovery was one of roughly 100 pitches in the first cohort; three products made it through all four stages. The journey took four years from solo research to general availability. The Explore stage ran almost entirely in Figma, testing solution concepts with potential customers before a line of production code was written. The gate review from Explore to Make involved a six-page document reviewed silently before discussion — borrowed from Amazon’s practice.
The outcome: 8,000 customers, strong CSAT, one of the fastest-growing products in Atlassian history.