Summary
Bill begins by recounting his experience at Netflix, where he led multiple simultaneous experiments and learned that 95% of UI code was thrown away, which shifted his mindset to designing for throw-away ability and viewing the UI layer as an experimentation layer. At PayPal, he confronted legacy technology stacks and a resistant culture, but with top-down support, he initiated cultural and technological transformations focused on rapid, user-centered iteration. Bill highlights the power of co-locating engineers, designers, and product folks to maintain weekly user studies and iterate checkout experiences, significantly improving customer satisfaction and business results. He introduces the concept of shearing layers in software stacks to separate stable, secure core layers from volatile experimentation layers, allowing calculated risk-taking. Bill stresses the need to shift organizational culture from delivery-centric to learning-centric, enabling teams to be empowered with continuous customer feedback and collaboration. He discusses democratizing engineering by leveraging internal open source (GitHub) and creating KrakenJS, an open-sourced Node.js framework, which increased developer engagement and reduced onboarding overhead. Bill critiques Agile’s lack of strategic direction and describes how integrating lean UX practices alongside Agile gives it a “brain” by grounding backlogs in real customer problems. He concludes with three essential lessons: achieve shared understanding via a common vocabulary, foster deep collaboration across disciplines, and maintain continuous customer feedback to keep teams aligned with user needs. Throughout, Bill emphasizes embracing failure as incremental learning and the importance of engineering infrastructures designed expressly for learning and change.
Key Insights
-
•
Netflix’s experimentation approach led to discarding around 95% of UI code annually, requiring design for throw-away ability.
-
•
Engineering’s highest value is enabling learning, not just building durable systems.
-
•
At PayPal, slow legacy stacks forced months-long changes that impeded innovation and led to cultural resistance.
-
•
Introducing shearing layers in software stacks helps separate stable secure components from volatile experimental layers.
-
•
Rapid user testing with co-located teams catalyzes accelerated learning and better product outcomes.
-
•
Democratizing innovation through internal open source practices like GitHub and KrakenJS enabled wider participation and faster adoption.
-
•
Agile alone lacks strategic direction; combining it with lean UX grounds backlog creation in real user problems.
-
•
Smaller, autonomous 'two-pizza' sized teams promote faster, higher-quality development than large monolithic teams.
-
•
Continuous customer feedback and embedding user context within teams prevents disconnected, misguided solutions.
-
•
Failure should be incremental and small to enable learning and avoid costly big mistakes.
Notable Quotes
"I started thinking of the UI layer as the experimentation layer."
"Engineering’s number one goal should be to enable learning."
"Organizations contain antibodies — people who resist change but often have good intentions."
"We locked ourselves away for 12 to 18 weeks with customers every week — the rhythm was incredibly powerful."
"Most organizations have a culture of delivery, but they need to move to a culture of learning."
"Great teams, if given enough context, will do smart things."
"We democratize the code base internally to create an intersourcing model, like open source inside the company."
"Agile needs a brain — without user context, it’s just a machine cranking out code."
"Failure is inevitable, but failing in small increments is the key difference."
"Continuous customer feedback soaks your teams with the context of the customer and problem set."
Or choose a question:
More Videos
"If design didn’t move the needle, we wouldn’t be having this conversation."
Standardizing Product Merits for Leaders, Designers, and Everyone
June 15, 2018
"There’s no such thing as one-size-fits-all methods; what works in one part of the map doesn’t work in another."
Simon WardleyMaps and Topographical Intelligence (Videoconference)
January 31, 2019
"Bias is a tendency, feeling or opinion for or against something without reason or evidence."
Sandra CamachoCreating More Bias-Proof Designs
January 22, 2025
"Appealing to stakeholders’ best interests helps build trust and rapport."
Darian DavisLessons from a Toxic Work Relationship
January 8, 2024
"Timely engagement helps stakeholders make faster, more informed decisions that lead to better business and product outcomes."
Fisayo Osilaja[Demo] The AI edge: From researcher to strategist
June 4, 2024
"Temporary, quick, and sometimes rough artifacts have a lifespan just long enough for the conversation they enable."
Uday GajendarThe Wicked Craft of Enterprise UX
May 13, 2015
"We formed a small core task force empowered to make decisions to avoid delays from formal reviews."
Davis Neable Guy SegalHow to Drive a Design Project When you Don’t Have a Design Team
June 10, 2021
"We stopped talking about patterns and consistency and started talking about scalability and speed to connect with stakeholders."
Eniola OluwoleLessons From the DesignOps Journey of the World's Largest Travel Site
October 24, 2019
"Design producers love their creative tools and want to keep using them without multiple manual reporting steps."
Aurobinda Pradhan Shashank DeshpandeIntroduction to Collaborative DesignOps using Cubyts
September 9, 2022