How Fast Is Fast Enough? [PerfNow 2025]

It sounds like a simple question, but if you've been working in web performance long enough, you know the answer is far from straightforward. In this keynote, I ask questions — ranging from neuroscience to business metrics — that will get you thinking about how you create, measure, and report the performance goals for your website.

Transcript

1. How Fast Is Fast Enough? @tammyeverts #PerfNow
2. ¯\_(ツ)_/¯
3. Fast is magical. Catherine Jue, Co-founder at Kernel catherinejue.com/fast
4. Fast is magical. Fast eliminates cognitive friction. Raycast surfacing the right application before you finish typing feels like an extension of your mind. Superhuman's sub-100ms rule—plus their focus on keyboard shortcuts—changed the email game in a way that no one's been able to replicate, let alone beat. I recently used Mercury to pay another business that uses Mercury, and its instant settle felt surprising in a world where bank transfers usually take days. Catherine Jue, Co-founder at Kernel catherinejue.com/fast
5. Fast is magical. Fast eliminates cognitive friction. Raycast surfacing the right application before you finish typing feels like an extension of your mind. Superhuman's sub-100ms rule—plus their focus on keyboard shortcuts—changed the email game in a way that no one's been able to replicate, let alone beat. I recently used Mercury to pay another business that uses Mercury, and its instant settle felt surprising in a world where bank transfers usually take days. No one praises these tools explicitly for their speed. Catherine Jue, Co-founder at Kernel catherinejue.com/fast
6. Fast is magical. Fast eliminates cognitive friction. Raycast surfacing the right application before you finish typing feels like an extension of your mind. Superhuman's sub-100ms rule—plus their focus on keyboard shortcuts—changed the email game in a way that no one's been able to replicate, let alone beat. I recently used Mercury to pay another business that uses Mercury, and its instant settle felt surprising in a world where bank transfers usually take days. No one praises these tools explicitly for their speed. They just feel magical. Catherine Jue, Co-founder at Kernel catherinejue.com/fast
7. Fast is pragmatic.
8. Fast is pragmatic. You should make your site fast enough to improve business metrics like bounce rate and conversion rate.
9. Fast is pragmatic. You should make your site fast enough to improve business metrics like bounce rate and conversion rate. “If there’s no ROI, there’s no point in making pages faster.”
10. Pragmatic vs Magical
11. Pragmagical?
12. 12 @tammyeverts.com /tammyeverts
14. pragmatic magical
15. pragmatic magical
16. Who are you trying to please?
17. Who are you trying to please? Your boss?
18. Who are you trying to please? Your boss? Google?
19. Who are you trying to please? Your boss? Google? Your users?
20. Your boss*
21. Your boss* *the Very Important Person who sent you here
22. “We’re so cool, we don’t need to worry about site speed.”
23. Luckily, we can test that assumption!
25. 1. Slow site = fewer conversions
26. 1. Slow site = fewer conversions 2. Slow site = fewer return visits
27. 1.1 second Largest Contentful Paint (LCP) 6.2% conversion rate speedcurve.com/blog/site-speed-business-correlation/
28. speedcurve.com/blog/web-performance-plateau/
29. speedcurve.com/blog/site-speed-business-correlation/ 1.1 second is your GOAL 2.8 seconds is your THRESHOLD
30. 1. There’s no “one size fits all” correlation
31. 1. There’s no “one size fits all” correlation 2. If your metrics stay on the performance plateau, your business metrics don’t change
32. Google
33. speedcurve.com/web-performance-guide/get-started-with-core-web-vitals/
34. “We highly recommend site owners achieve good Core Web Vitals for success with Search and to ensure a great user experience generally. This, along with other page experience aspects, aligns with what our core ranking systems seek to reward.” developers.google.com/search/docs/appearance/core-web-vitals
36. Google thresholds ≠ Your business thresholds
37. contentsquare.com/blog/interaction-to-next-paint/
38. 200ms is “good” INP for Google
39. 200ms is “good” INP for Google speedcurve.com/blog/INP-user-experience-correlation/
40. speedcurve.com/blog/INP-user-experience-correlation/ 200ms is “good” INP for Google 50ms 150ms ??? 150ms
42. Your users
44. Metrics ≠ Feelings
45. People spending a long time on a web page or clicking on many links and visiting many pages can indicate that they're engaged… Kara Pernice President & CEO of Nielsen Norman Group
46. People spending a long time on a web page or clicking on many links and visiting many pages can indicate that they're engaged… Kara Pernice President & CEO of Nielsen Norman Group or totally lost.
47. People spending a long time on a web page or clicking on many links and visiting many pages can indicate that they're engaged… Kara Pernice President & CEO, Nielsen Norman Group A usability test might show us why the user is engaged or having trouble. or totally lost.
48. People spending a long time on a web page or clicking on many links and visiting many pages can indicate that they're engaged… Kara Pernice President & CEO of Nielsen Norman Group A usability test might show us why the user is engaged or having trouble. This is why I think of usability tests and site analytics as a perfect pairing for user research . or totally lost.
49. Engagement tasks vs Productivity tasks
50. Engagement tasks vs Productivity tasks Open ended, absorbing, potentially lengthy
51. Engagement tasks vs Productivity tasks Open ended, absorbing, potentially lengthy Higher tolerance for waiting
52. Engagement tasks vs Productivity tasks Open ended, absorbing, potentially lengthy Higher tolerance for waiting e.g., researching a trip
54. Engagement tasks vs Productivity tasks Fast completion of specific outputs Open ended, absorbing, potentially lengthy Higher tolerance for waiting e.g., researching a trip
55. Engagement tasks vs Productivity tasks Fast completion of specific outputs Lower tolerance for waiting Open ended, absorbing, potentially lengthy Higher tolerance for waiting e.g., researching a trip
56. Engagement tasks vs Productivity tasks Fast completion of specific outputs Lower tolerance for waiting e.g., administering your trip Open ended, absorbing, potentially lengthy Higher tolerance for waiting e.g., researching a trip
58. Same person, metrics, clicks, time on site…
59. Same person, metrics, clicks, time on site… Different user experience
60. One slow step can make the whole journey feel slow Sander van Surksum Founder of Iron/Out
61. Creating flow
62. “…a state in which people are so involved in an activity that nothing else seems to matter; Mihály Csíkszentmihályi Flow: The Psychology of Optimal Experience (1990)
63. “…a state in which people are so involved in an activity that nothing else seems to matter; the experience is so enjoyable that people will continue to do it even at great cost, for the sheer sake of doing it” Mihály Csíkszentmihályi Flow: The Psychology of Optimal Experience (1990)
64. Creating flow
65. Faking flow
66. 66 “We want you to be able to flick from one page to another as quickly as you can flick a page on a book. Urs Hölzle Google Fellow
67. 67 “We want you to be able to flick from one page to another as quickly as you can flick a page on a book. So, we’re really aiming very, very high here… at something like 100 milliseconds .” Urs Hölzle Google Fellow
69. pragmatic magical
70. pragmatic unique to your site magical 100ms
71. Pragmagical
72. Know who you’re trying to please
73. Know who you’re trying to please Know why users are on your site (engagement or productivity?)
74. Know who you’re trying to please Know why users are on your site (engagement or productivity?) Identify the right metrics and thresholds
75. Know who you’re trying to please Know why users are on your site (engagement or productivity?) Identify the right metrics and thresholds Meet your targets and mitigate regressions ASAP
76. Know who you’re trying to please Know why users are on your site (engagement or productivity?) Identify the right metrics and thresholds Meet your targets and mitigate regressions ASAP Never stop monitoring
77. Know who you’re trying to please Know why users are on your site (engagement or productivity?) Identify the right metrics and thresholds Meet your targets and mitigate regressions ASAP Never stop monitoring Do usability testing to avoid false assumptions
78. Know who you’re trying to please Know why users are on your site (engagement or productivity?) Identify the right metrics and thresholds Meet your targets and mitigate regressions ASAP Never stop monitoring Do usability testing to avoid false assumptions And most important…
79. 79 Always err on the side of creating an experience that isn’t just tolerable…
80. 80 Always err on the side of creating an experience that isn’t just tolerable… It’s delightful.
81. Thank you! :) @tammyeverts.com /tammyeverts Help desk @ 1:10 Performance budget demo @ 14:25