How to validate your site speed optimization efforts

Optimizing site speed isn't merely about implementing various techniques; it's about validating the effectiveness of these optimizations and ensuring they yield tangible results for your users and your business.
In this talk, Aymen will delve into the crucial aspect of validating your web performance optimization efforts.
He will equip you with practical strategies, tools and actionable insights to assess and validate the contribution of what you deployed as optimization to your user’s experience.
The main takeways from Aymen's talk are:
- Don't use Google Lighthouse to validate site speed optimizations : It's an emulation and doesn't represent real users
- Equip your project with a Real User Monitoring tool and learn how to use it
- Monitor the good metrics and always segment RUM data to extract value
- Be sure to make atomic, detailed, and specific web performance user stories

Aymen, is an international SEO and Web Performance Consultant and founder of Speetals.

Transcript

1. How to validate your site speed optimization efforts Loukil Aymen SPEETALS Speakerdeck.com/aymenloukil @aymenloukil @LoukilAymen
2. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Site speed matters for Users SEO Conversions Brand reputation Inclusiveness Accessibility Costs
3. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Site speed matters for Users SEO Conversions Brand reputation Inclusiveness Accessibility Costs Business
4. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals It costs money. 1500$ /day Web Performance Consultant.
5. @LoukilAymen #BrightonSEO Speetals 1$ every 30 seconds
6. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Did it even work ? Effort and investment Impact
7. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals No impact evidence Effort and investment Impact
8. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Projects can easily go wrong 💰Migration 😱 TTFB X 2
9. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Main reasons Inadequate tools/data
10. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Main reasons Inadequate tools/data Inadequate metrics
11. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Main reasons Inadequate tools/data Inadequate metrics Unclear expectations
12. @LoukilAymen #BrightonSEO Speetals Let’s make Web Performance more predictable and impact proven
13. @LoukilAymen #BrightonSEO Speetals How to validate site speed efforts?
14. @LoukilAymen #BrightonSEO Speetals Use Google Lighthouse! Jonathon Burton
15. @LoukilAymen #BrightonSEO Speetals Check my 2023 Brighton’s talk https://www.youtube.com/@AymenLoukil
16. @LoukilAymen #BrightonSEO Speetals Set up right tools, Measure, know you current performance and get a history Set a baseline The Lazy Artist Gallery
17. @LoukilAymen #BrightonSEO Speetals You need real users data using your product..
18. @LoukilAymen #BrightonSEO Speetals You need real users data browsing your website…
19. @LoukilAymen #BrightonSEO Speetals Real User Monitoring RUM
20. @LoukilAymen #BrightonSEO Speetals RUM is like weather stations
21. @LoukilAymen #BrightonSEO Speetals Weather station Real-time weather conditions Rain, wind, sun… Monitor, alert, and predict Image: RitaE Pixabay tailulai
22. @LoukilAymen #BrightonSEO Speetals RUM Real-time users data and interactions
23. @LoukilAymen #BrightonSEO Speetals RUM Real-time users data and interactions Fast, slow, laggish experiences
24. @LoukilAymen #BrightonSEO Speetals RUM Real-time users data and interactions Fast, slow, laggish experiences Prioritize, monitor, alert, validate
25. @LoukilAymen #BrightonSEO Speetals Clear view of your users experience Main benefits Masood Aslami
26. @LoukilAymen #BrightonSEO Speetals Optimize and see the impact Main benefits Clear view of your users experience Masood Aslami
27. @LoukilAymen #BrightonSEO Speetals RUM data is the cream of the crop Somben Chea
28. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals
29. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals It’s site speed not the NASA RUM tool should be simple to use
30. @LoukilAymen #BrightonSEO Speetals How to take benefit from RUM data?
31. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Split and break down https://www.stiernholm.com/en/blog/how-to-divide-a-large-task-into-smaller-pieces
32. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Break down, get value Ella Olsson 2$ 17$ 5$
33. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Segment the data Device...
34. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Segment the data Device, Connection type...
35. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Segment the data Device, Connection type, Location...
36. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Segment the data Device, Connection type, Location, Navigation type...
37. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Segment the data Device, Connection type, Location, Navigation type, CPU, …
38. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Segment the data Device, Connection type, Location, Navigation type, CPU, Device memory…
39. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Segment the data Device, Connection type, Location, Navigation type, CPU, Device memory, Page type…
40. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Different locations
41. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals X2 TTFB CLS fails
42. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Different connection types 4G 3G
43. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Perfect LCP VS Poor LCP 4G 3G
44. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals https://web.dev/case-studies/vitals-business-impact
45. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals
46. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals AVG depth of 1m
47. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Would you feel safe to cross the lake walking? AVG depth of 1m
48. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Web Performance data is asymmetric, so don’t chase averages.
49. @LoukilAymen #BrightonSEO Speetals Averages are miƨlɘɒbinϱ George Becker
50. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals You do need PERCENTILES
51. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Scan your audience 25th percentile 75th percentile 95th percentile
52. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals 5% of the users got an LCP > 3.3s 75% of the users got an LCP <= 1.3s
53. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Inadequate tools/data Inadequate metrics Unclear expectations Main reasons
54. @LoukilAymen #BrightonSEO Speetals Optimization need to be small, isolated, and atomic
55. @LoukilAymen #BrightonSEO Speetals Atomic optimizations, easier validation Markus Spiske
56. @LoukilAymen #BrightonSEO Speetals Add “Fetchpriority” high to hero image
57. @LoukilAymen #BrightonSEO Speetals Add “Fetchpriority” high to hero image Metric: LCP Page type: PDP Device: Mobile, Desktop Location: Belgium
58. @LoukilAymen #BrightonSEO Speetals Description : The main product image should have the fetchpriority="high" attribute set in its HTML tag. The browser should prioritize downloading the main product image over other less critical resources. Add “Fetchpriority” high to hero image Metric: LCP Page type: PDP Device: Mobile, Desktop Location: Belgium
59. @LoukilAymen #BrightonSEO Speetals [..Description..] How can we measure the impact ? RUM data, PDP segment. LCP percentiles and distribution. Exclude Firefox (doesn’t support) Add “Fetchpriority” high to hero image Metric: LCP Page type: PDP Device: Mobile, Desktop Location: Belgium
60. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Site speed validation Takeaways Atomic optimizations Get RUM & learn to use Adequate metrics
61. @LoukilAymen #BrightonSEO Speetals @LoukilAymen #BrightonSEO Speetals Hey, I’m Aymen International Web Performance Consultant speetals.com