Reaching Gen Z: A Product-Discovery Story
I found a segment the product was quietly losing, Gen Z, and turned it into a 3% lift in daily retention by deciding what to build before a line of code was written. Leading a cross-functional team of 7 and partnering with teams worldwide, I sized the opportunity in the data, validated it with real users, and made the prioritization calls that separated signal from noise. This is the "what to build" side of product, shown end to end.
The opportunity hiding in the cohort data
The platform had a healthy core audience, but growth in that segment was flattening. The next chapter of growth was not going to come from squeezing the same users harder; it was going to come from a new segment. Gen Z stood out: they were arriving on the platform, but they were not staying. They showed up in acquisition and then dropped out of the engagement funnel faster than any other cohort.
That gap was the opening. Gen Z expectations, mobile-first, fast, authentic, and low-friction, were not the assumptions the product had originally been built around. The question was not "should we care about Gen Z?" The data already said yes. The question was "what specifically do we build for them, and how do we know it will work before we spend a quarter of engineering on it?"
Lead discovery, and decide what to build
My mandate was to own discovery for the Gen Z expansion: turn a fuzzy "reach a new segment" goal into a validated, prioritized set of things worth building. Concretely, that meant three responsibilities held at once.
Find the truth
Validate the opportunity with real Gen Z users before committing the roadmap, so we built from evidence rather than assumption.
Make the calls
Decide what made the cut and what did not, and be able to defend why, against impact, effort, and confidence.
Align the team
Lead a cross-functional team of 7 and international partners behind one product vision and a single roadmap.
Framing the opportunity as testable hypotheses
I started where the signal was strongest: the engagement metrics. By segmenting the funnel by cohort I could see exactly where Gen Z users dropped relative to the core audience, and pair each drop with the qualitative complaints we already had. Rather than jump to features, I translated the gaps into problem hypotheses we could actually test, so discovery had a clear pass/fail bar.
Onboarding friction
Gen Z abandons before first value because onboarding is too long and asks for too much up front.
Time-to-value
The first session does not reach a rewarding moment fast enough to form a habit, so they do not return the next day.
Tone & relevance
The content and voice read as built for an older audience, so Gen Z does not see themselves in the product.
Validating with real users, not opinions
I designed the flows, mockups, and wireframes in Figma, then put them in front of real Gen Z users through UserTesting. Moderated and unmoderated sessions told me where people got stuck, what language landed, and which of my hypotheses survived contact with an actual user. I mapped what I heard against the journey so the team could see friction in context, not as a list of disconnected complaints.
Discover
High intent
Arrives curious, mobile, low patience for setup.
Onboard ⚠
Friction spike
Too many steps before anything useful happens.
First value ⚠
Too slow
Rewarding moment arrives late, habit never forms.
Habit
Rarely reached
Few return on day two without an earlier payoff.
Theme · Speed
"I gave up before I figured out what it did." Repeated abandonment during multi-step setup.
Insight: cut onboarding to the shortest path to a first win.
Theme · Payoff
Users wanted a visible reward in the first session, not after configuration.
Insight: move a rewarding moment to the first session.
Theme · Voice
Tone read as corporate; copy did not feel made for them.
Insight: rework tone, but test it before assuming what lands.
What to build, and what to leave on the table
IllustrativeDiscovery produces more ideas than any team can ship. The real product work was deciding what not to build. I scored each candidate against impact on the validated friction, effort, and how much confidence the research gave me, then made an explicit Build, Defer, or Cut call on each. The tone rework was the clearest trade-off: users complained about it, but it scored lower on impact-per-effort than the onboarding and time-to-value fixes, so I deferred it rather than let it dilute the focused bet.
| Opportunity | Impact | Effort | Confidence | Decision |
|---|---|---|---|---|
| Streamline onboarding to first win | High | Med | High | Build |
| Pull a rewarding moment into session one | High | Med | High | Build |
| Rework tone & copy for Gen Z | Med | High | Med | Defer |
| Net-new social feature set | Unproven | High | Low | Cut |
The call: concentrate the quarter on the two highest-confidence, highest-impact fixes the research had earned, defer the tone rework until those shipped, and cut the speculative social bet that no evidence supported yet. A focused roadmap the team believed in beat a broad one nobody could defend.
Letting A/B tests settle the design debates
IllustrativeOnce the two bets were live, I did not assume the first design was the right one. I A/B tested the onboarding flow, a shorter guided path against the longer original, and let the engagement data decide instead of arguing it in a review. The winning variant became the default, and the result fed straight back into the optimizations we kept making.
Variant A · original
ControlLonger multi-step setup. Baseline activation.
Variant B · streamlined
trophyWinnerShorter path to first win. Higher activation and day-two return.
A validated bet, a measurable lift, a global rollout
+3%
daily user retention
7
cross-functional team led
Global
rollout with intl. teams
Verified outcome: the shipped onboarding and time-to-value changes contributed to a 3% increase in daily user retention. The engagement metrics I surfaced kept informing optimizations after launch, and I partnered with international teams to drive the product vision and roll the work out globally.
This discovery work was presented at the company All Hands as an example of evidence-led product decision-making.
What I'd do differently
- check_circleInstrument the segment before discovery, not during. I rebuilt some cohort views mid-discovery. Standing up clean Gen Z engagement dashboards up front would have shortened the time from question to evidence.
- check_circleRevisit the deferred tone work with its own test. Deferring the copy rework was the right call against impact-per-effort, but it deserved a fast, scoped experiment afterward rather than sitting in the backlog by default.
- check_circleBring international partners into discovery earlier. Aligning on the vision globally went smoothly, but pulling those teams into the research phase, not just the rollout, would have surfaced regional nuances sooner.