KEYNOTE: Why Customers Can’t or Won’t Use Your Technically-Right Data Products
summary
Claim: technically sound analytics fail when outputs don’t map to human goals. Evidence: surveys show >80% adoption problems; a cathedral fire detection UI displays sensor codes not "fire" and leads to misdirected response; a monitoring product redesign that surfaces anomalies across 1,321 VMs with a 72 hour sparkline cuts tool use and speeds fixes; an artist income portal misleads users until counts (not amounts) are shown. Prescription: routine user exposure, design for adoption, teams owning outcomes — technically right, effectively wrong — users will pay to use it if they have to.
AI-written summary · about
chapters (13)
- 0:00Keynote introduction
- 1:15Customers rejecting data products
- 3:23Notre Dame fire detection failure
- 6:58Human-centered data design
- 13:11Three design principles
- 27:39Good vs bad data design
- 30:29Design principles in practice
- 44:33Data product resources
- 45:35Recap and call to action
- 51:30Q&A leadership & design
- 55:46Q&A design pitfalls
- 1:00:19Q&A ML & real-time
- 1:02:27Closing remarks
related talks
40mBuilding Data Cultures
52mScaling Data Teams: Three Pitfalls and How To Avoid Them
24mA Practical Approach To Unlock Value From Data and Analytics
28mWhat Data Analysts Wish Application Developers Knew
23mUsing Data to Bridge Product and Engineering Teams
37mMaximizing Impact of Analytics Teams
31mWhy Most Data Projects Fail and How to Avoid It
22mBridging Business and Tech for Data Science Success