How Chrome keeps users happy: six truths and a lie.

Monica Dinculescu
750 million people use Chrome. That's blind people, old people, young people, people who hate change, and sometimes, your cat.

How do we make sure that those people are having a good experience? Science! Mostly.

So, how do you build a usable product? What are user metrics and how do you make sure that your changes actually improve the product experience? How do you deal with user feedback, and what can we do about accessibility?

These are some of the things I've learned from a year of working on the Chrome UI.

Transcript

1. USERS HAPPY (six TRUTHS AND A LIE) HOW CHROME KEEPS
3. HELLO, I’M MONICA! @NOTWALDORF
4. CHROME, RIGHT?
5. SO MANY USERS 750 MILLION-ISH
6. SO MANY USERS 1% IS STILL A LOT
7. SO MANY USERS KEEP THEM HAPPY
8. DON’T BE A JERK USERS ALREADY HATE THINGS TRUTH #1
9. A LOT OF THINGS USERS HATE
10. CHANGE USERS HATE
11. BLOCKS OF TEXT USERS HATE
12. DECISIONS USERS HATE
13. DECISIONS USERS HATE PICK GOOD DEFAULTS!
14. QUIRKY THINGS USERS ❤
15. QUIRKY THINGS USERS ❤
16. NEAT THINGS USERS ❤
17. NEAT THINGS USERS ❤
18. “TO THE GOOGLE ENGINEER THAT ADDED A SPEAKER ICON […] SOMEDAY I HOPE TO BUY YOU A STEAK DINNER. OR A NOBEL PRIZE.” - @MATHOWIE
19. “TO THE GOOGLE ENGINEER THAT ADDED A SPEAKER ICON […] SOMEDAY I HOPE TO BUY YOU A STEAK DINNER. OR A NOBEL PRIZE.” - @MATHOWIE
20. USER FEEDBACK HOW TO LISTEN TO USERS LIE #1
21. FEEDBACK IS GOOD
22. FEEDBACK IS GOOD GENERALLY
23. FEEDBACK IS GOOD YOU != YOUR USERS
24. FEEDBACK IS GOOD CAN’T EDUCATE USERS
25. PEOPLE CARE ABOUT CHEESE
26. FEEDBACK SUCKS BECAUSE THE INTERNET
27. FEEDBACK SUCKS STILL MAYBE RELEVANT
28. FEEDBACK SUCKS DOESN’T MEAN YOU SUCK
29. FEEDBACK SUCKS YOU != YOUR PRODUCT
30. FEATURE FLAGS SAFETY NETS FOR EVERYONE! TRUTH #2
31. PICK AND CHOOSE TURN FEATURES ON/OFF
33. THINGS LIKE
34. EXPERIMENTAL JS THINGS LIKE
35. EXPERIMENTAL JS THINGS LIKE FAST TAB CLOSE
36. EXPERIMENTAL JS THINGS LIKE FAST TAB CLOSE PREFETCH RESULTS
37. EARLY FEEDBACK DOES THE FEATURE SUCK?
38. SPOILS SURPRISES BECAUSE OPEN SOURCE
39. EASY ⌘Z UNDO IS BEST DO
40. 㽈 ̇ ̇ 㽈 ̇ THIS LEADS US TO…
42. USER METRICS SCIENCE == BEST TRUTH #3
43. USER FEEDBACK WHAT USERS THINK THEY DO
44. USER METRICS WHAT USERS THINK THEY DO
45. COLLECT DATA ABOUT EVERYTHING
46. OPT-IN ANONYMOUS DATA
47. OPT-IN
48. SAMPLE DATA
49. CLICKS SAMPLE DATA
50. CLICKS SAMPLE DATA PAGE LOAD TIME
51. CLICKS SAMPLE DATA PAGE LOAD TIME NUMBER OF ERRORS
52. HISTOGRAMS
53. ANALYZE DATA DO THE SCIENCE!!!
54. SPOILERS: NO DIFFERENCE
55. SPOILERS: SO MUCH LESS MALWARE!
56. SPOILERS: SO MUCH LESS MALWARE!
57. ALWAYS MEASURE PROVE THAT YOU NEED IT
58. NEVER ENABLE WITHOUT USER TESTING
59. DESIGNERS ❤ USER {INTERFACE, EXPERIENCE} TRUTH #4
60. UX IS HARD
61. GOOD UX IS HARD ACTUALLY
62. GOOD UX IS HARD LOTS OF USER TESTING!
63. OMG PIXELS
64. OMG PIXELS DEVS USING GIMP
65. OMG PIXELS CONSISTENT UI == BEST UI
66. INTERNATIONALIZATION C’EST BEAUCOUP DE LETTRES TRUTH #5
67. 55.7% ENGLISH (W3TECHS, 2013)
68. 44.3% OTHER
69. CODE IS MAGIC! DON’T INLINE STRINGS
70. RIGHT-TO-LEFT
73. ACCESSIBILITY IT’S NOT ALL ABOUT YOU TRUTH #6
74. THE WORLD
75. COLOUR BLIND THE WORLD 10%
76. COLOUR BLIND THE WORLD LOW VISION 4% 10%
77. COLOUR BLIND THE WORLD LOW VISION 4% 10% BLIND 1% (WHO, 2013)
78. COLOUR BLIND THE WORLD IN CHROME LOW VISION 30 M 75 M BLIND 7.5 M
79. CHROME CARES ABOUT
80. KEYBOARD ONLY CHROME CARES ABOUT
81. LOW VISION CHROME CARES ABOUT
82. LOW VISION CHROME CARES ABOUT
85. ASSISTIVE TECH CHROME CARES ABOUT
86. <div id="percent-loaded" role="progressbar" aria-valuenow="75" aria-valuemin="0" aria-valuemax="100" />
87. THE MORAL? KNOW YOUR USERS OK, SO
88. THE MORAL? THINK LIKE YOUR USERS OK, SO
89. THE MORAL? SCIENCE IS YOUR FRIEND OK, SO
90. ❤ @NOTWALDORF