CHAPTERS
Why product discovery is #1 leverage for PMs (and how AI changes the game)
Aakash frames discovery as the highest-leverage PM activity and sets up Tanguy Crusson’s perspective from leading Jira Product Discovery. Tanguy signals there’s no perfect “ideal” process, but there is a repeatable system that works at both feature and product levels.
The 4-stage discovery vocabulary: Wonder → Explore → Make → Impact
Tanguy introduces a four-stage model used across Atlassian to describe where an initiative truly is. The stages create shared language so stakeholders don’t assume an idea is further along than it is and so teams know what kind of help is being requested.
Wonder stage in practice: deep problem exploration with real customer footage
Wonder focuses on unpacking a fuzzy problem area through a small number of rich conversations rather than broad surveys. The output is designed to make the pain visceral: short, curated videos of customers describing their challenges in their own words.
Running better user interviews: avoid leading, embrace silence, get researcher training
Tanguy explains how PMs commonly bias interviews by steering users toward preconceived answers. He recommends getting coached by a professional researcher and adopting tactics like open-ended prompts, no options, and minimal interruptions to capture genuine insight.
Turning messy interviews into a compelling story: clip highlights, don’t over-write docs
The “document” matters far less than the narrative created by selected video snippets. Tanguy describes how to know when to stop interviewing (convergence) and how to assemble clips that ladder up to a clear, shared understanding of the problem.
Tooling and systematizing customer access: Dovetail, Loom, Pendo, community, Slack
Discovery scales when customer contact is frictionless and habitual. Tanguy outlines a stack and operating rhythm that makes recruiting, interviewing, clipping, and tagging insights nearly automatic—then ties it to sustained CSAT at large scale.
Explore stage: validate solutions with prototypes before writing code
Explore is about fast solution testing with the same users who articulated the pain in Wonder. Tanguy shows how Jira Product Discovery itself started with extremely lo-fi prototypes (even a single slide) and progressed only when customers clearly signaled urgency.
From ‘nice idea’ to ‘must-have’: how customer reactions guided the JPD concept
Tanguy contrasts weak validation (an inbox-like feedback concept that got lukewarm responses) with strong validation (idea-centric prioritization tied to goals/impact that triggered immediate demand). The lesson is to prototype multiple concepts and let users pull you toward the right one.
Make stage strategy: the ‘safety funnel’ and staged rollout (10 → 100 → 1,000)
In Make, the team commits to building but stays disciplined about who gets access when. Using a “safety funnel” mindset, they minimize bad early experiences by gradually expanding access as confidence and user satisfaction rise.
Launch timing and distribution: waitlists, messaging tests, and validating demand
Tanguy challenges simplistic advice like “just ship fast” by reframing around the biggest risk. For JPD, they validated market pull and distribution early via a landing page and waitlist, then progressively onboarded segments based on readiness and appetite.
Free vs paid: when feedback is real without a credit card
They discuss the risk of relying on free usage as a proxy for willingness to pay. Tanguy argues it depends on the market and switching costs—especially in B2B workflow tools where adoption effort itself is a meaningful commitment—while still emphasizing the need to validate monetization.
Keeping discovery alive during Make: spikes, shaping V0, and rapid internal iteration
Tanguy explains how discovery continues through delivery by pairing validated prototypes with technical spikes and iterative shaping. The team resets designs with engineers involved, defines a minimal V0, and uses tight feedback loops to reach customers quickly without waiting for “done.”
Fast feedback loops in code: branch previews, internal testing, and customer co-building
The team shortens iteration cycles by letting PMs preview work directly from feature branches. They use internal dogfooding, then co-develop with a small set of customers via Slack/Zoom, widening access only after learning and polishing.
Where to connect and closing remarks
Tanguy shares how listeners can reach him and the team, especially through the Jira Product Discovery community. Aakash closes with pointers to the full podcast, linked artifacts, and ways to support the channel.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome