Bartłomiej Lenard — Experience Design Lead at EY
Design decision making - is it possible to manage the product by design?
Design & Business
🇬🇧 English
Design & Business
Always Measure Your Assumptions – Growth Design in Theory and Practice
About
How to guarantee your product doesn’t just deliver value to its users but also creates value for your business? How to make the right calls confidently and learn from inevitable failures? How to make sure your growth direction will improve the product long-term? During the lecture, we’ll explore a tried-and-tested approach to creating a stable and continuous product development strategy, one which allows you to balance out sweeping changes and moderate adjustments. We’ll examine how to blend design, marketing, programming, and data analytics to rapidly generate and validate ideas that directly grow your business and help inform future decisions.
Watch the full talk
Watch this WaysConf session, then continue with related talks or explore the current programme.
From the recording
Talk in brief
A growth-design framework for measuring assumptions instead of confusing correlation with causation. Quantitative data shows where behavior changes; qualitative research helps explain why, but both have limitations. The talk describes experiment plans that specify what, where, when, who, and projected impact, then connects them with continuous discovery rather than one-way departmental handoffs. Metrics should act as goals and learning signals, supported by a growth equation that links engagement and business outcomes. Measurement becomes useful when it guides the next product decision.
Key takeaways
- 01
Separate correlation from causal evidence
Two measures can move together without one producing the other, so growth teams need hypotheses and controlled tests.
Watch from 7:02 - 02
Combine behavioral scale with qualitative depth
Analytics reveals patterns while research explains expectations and thought processes, and each method compensates for the other’s gaps.
Watch from 12:10 - 03
Specify the complete experiment decision
A plan should state the change, journey location, duration, audience cohort, and expected effect before implementation.
Watch from 17:25 - 04
Replace handoffs with continuous discovery
Ongoing cross-functional learning prevents insights from stopping at departmental boundaries or disappearing after one project.
Watch from 22:37 - 05
Use growth equations to connect metrics
Engagement and bottom-line measures can form a model that shows which product behaviors are plausible levers for growth.
Watch from 28:09
Read edited transcript highlights
These concise notes were edited from automatic captions and checked against the talk structure. They are not a verbatim transcript.
Movement together does not prove a mechanism
Two metrics can correlate for reasons unrelated to the product story a team wants to tell. Before acting, state the mechanism: which change should affect which behavior, for whom, and why. An experiment or stronger observational design can then test that relationship. Without this step, a dashboard can encourage confident investment in coincidence.
Watch from 7:02Quantitative and qualitative evidence answer different questions
Behavioral data can show the size, location, and frequency of a pattern across many customers. Interviews and observation reveal expectations, interpretation, and context. Neither source is complete: metrics can hide motivation, while small samples and social responses can distort qualitative findings. A stronger decision deliberately combines their distinct contributions.
Watch from 12:10An experiment needs a decision-ready specification
Define what product change will be implemented, where in the journey it appears, when and for how long it runs, which cohort receives it, and what effect is expected. Writing these choices before launch exposes ambiguity and prevents the success criterion from being rewritten after results appear. It also makes operational dependencies easier to review.
Watch from 17:25Discovery should continue across functional boundaries
A linear process passes research to design, design to engineering, and completed work to maintenance. Each handoff can strip away context and leave no owner for the next question. Continuous discovery keeps product, design, and engineering close to evidence, enabling them to update the opportunity and solution together as results arrive.
Watch from 22:37Metrics become useful inside a growth model
A single number rarely explains sustainable growth. A simple equation can connect acquisition, activation, engagement, retention, monetization, or other relevant behaviors to the bottom line. The model is not permanent truth; it makes assumptions visible. Teams can test which factor is actually constrained and avoid optimizing an easy metric with weak product impact.
Watch from 28:09
Explore WaysConf 2026
