Upgrade to Pro — share decks privately, control downloads, hide ads and more …

7 Levels of a Web Performance Journey

7 Levels of a Web Performance Journey

The path to understanding the data around Web Performance, its value and business decisions surrounding this topic can be long. It is hard to see the whole journey when you and your organization are at any particular step, but it can be helpful to understand where to go next.

Eric Goldstein, Sergey Chernyshev and Alexander Chernyshev joined forces to describe and visualize this path to help make it more understandable and give you directions for your trip through various levels of this complex game.

Avatar for Sergey Chernyshev

Sergey Chernyshev

September 21, 2025
Tweet

More Decks by Sergey Chernyshev

Other Decks in Technology

Transcript

  1. 7 Levels of a Web Performance Journey How Focused Performance

    Metrics Can Improve User Experience Eric Goldstein & Sergey Chernyshev & Alexander Chernyshev NY UX Speed and Web Performance Meetup September 18, 2025
  2. About Us Eric Goldstein Sergey Chernyshev Alexander Chernyshev Principal Engineer

    @ Babbel Speed Engineer @ Cloudflare Web Developer & Designer
  3. Engineering focus: Speed Getting Up to Speed Business focus: Revenue

    Combine Focus to Create Performance Culture
  4. What Matters to My Business? Total Traffic Conversion Rate Order

    Size Level 1 Conversion Rate origins: performance, presentation quality, SEO, availability, etc.
  5. Examples of Perf/Revenue Improvements • ALDO – an online shoe

    retailer – found that fast rendering brought over 4x more revenue than slow rendering for mobile users and almost 7x more for desktop1 • Ray-Ban, Oakley, & Sunglass Hut – online sunglasses retailers – saw a direct, inverse relationship between Largest Contentful Paint and Cost-per-Click via Google Ads. In other words, as LCP improved, ads got cheaper, or as LCP got worse, ads got more expensive. They even established an equation between the two: Δ𝗖𝗣𝗖 = 𝗸 × Δ𝗟𝗖𝗣 + 𝗾 (𝗸 is the coefficient that measures the correlation between the two values and 𝗾 represents external variables like Ad quality and competition)2 • Swappie – a refurbished phone retailer – improved load time by 23%, LCP 55%, CLS 91%, and FID 90% resulting in a 42% increase in mobile revenue and a 10 percentage point increase in relative mobile conversion rate3 1) https:/ /wpostats.com/2020/05/27/aldos-revenue-rendering/ 2) https:/ /www.linkedin.com/posts/daniele-merola-2a531839_webperformance-performancemarketing-ux-activity-7338097683078774784-KYM9/ 3) https:/ /wpostats.com/2021/09/16/swappie-revenue-increase/ Level 2
  6. How Fast is My Site? 3.4 seconds Poor Needs Improvement

    Good (the time it takes to do one important thing) Level 3
  7. Plot percentile over time Mar 14 Mar 15 Mar 16

    Mar 17 Mar 18 Mar 19 Mar 20 Level 5
  8. 3 percentiles over time P99 P75 P50 Mar 14 Mar

    15 Mar 16 Mar 17 Mar 18 Mar 19 Mar 20 Level 5