Site speed & Core Web Vitals revenue benchmark
Is your website fast enough to convert, and what is each second costing? We measure your real-user Core Web Vitals and page speed, benchmark them against the HTTP Archive 2025 web, and translate the gap into conversions using Deloitte's speed research.
What the market looks like.
| KPI | Market reference | Top performers / note | Source |
|---|---|---|---|
| Core Web Vitals pass rate — mobile | 48% of mobile websites (2025) | 32% in 2021 | 1 |
| Core Web Vitals pass rate — desktop | 56% of desktop websites (2025) | 41% in 2021 | 1 |
| Retail mobile conversions per 0.1 s faster | +8.4% | — | 2 |
| Travel mobile conversions per 0.1 s faster | +10.1% | — | 2 |
| Lead-gen information-page bounce rate per 0.1 s faster | 8.3% better | — | 2 |
Speed is both a ranking signal and a conversion lever. Half the mobile web still fails Core Web Vitals1, which makes a fast site a competitive advantage, especially on mobile, where most traffic now lands.
Is this benchmark for you?
Industries
Retail & e-commerce · Hospitality · Professional services · SaaS & tech
Buyer
Sales & marketing · CIO, CTO & IT
Company profile
Any company whose website sells or generates leads
Good first benchmark if you are new to AI
You need it if…
- Pages feel slow on mobile
- Search rankings or ad quality scores are slipping
- Mobile converts far below desktop
What we measure. What you get.
What we measure
- Real-user LCP, INP and CLS on key templates
- Mobile vs desktop gap
- Heaviest scripts, images and third-party tags
- Speed on revenue pages: home, product, checkout, forms
- Conversion rate by speed band from your analytics
What you receive
- Your Core Web Vitals vs the 2025 web
- Revenue at stake from a speed lift, modelled on your traffic
- Prioritised fix list by effort and impact
- Monitoring set-up so speed stays fixed
Site Speed Benchmark 2026.
The full study behind this page: Core Web Vitals pass rates from 2021 to 2025, each measure in turn, platforms, page weight and third parties, what speed is worth with the evidence graded, a cost and payback model, ten playbooks and eleven worksheets.
What’s included
Tools you can use straight away
- Speed baseline
- Third-party tag inventory
- Main image checklist
- Interaction log
- Layout shift checklist
- Revenue at stake
- Project cost estimate
- Test plan
- Performance budget
- Questions for vendors and agencies
- Twelve-month tracker
- Scenario tables: payback by lift
- Six-month roadmap
About this edition: built from published research and NEWMIND models. It does not yet include our own survey; Edition 2 adds survey data from decision-makers, free for Edition 1 buyers.
Contents
- 01About this editionp. 5
- 02Executive summaryp. 7
- 03The three Core Web Vitalsp. 9
- 04How a site passes: the 75th percentile rulep. 10
- 05Pass rates 2021 to 2025: better, but most sites still fail on phonesp. 11
- 06Pass rates by popularity and page typep. 13
- 07Supporting measures: server response and first paintp. 15
- 08Loading (LCP): where phones fall behindp. 16
- 09What the LCP element is: mostly imagesp. 17
- 10Responsiveness (INP): fine on desktop, weaker on phonesp. 18
- 11Main-thread blocking: phones are doing more workp. 19
- 12Layout stability (CLS): the most improved measurep. 20
- 13Instant back and forward: the cache most sites blockp. 21
- 14E-commerce platforms: the choice of stack shows in the scoresp. 22
- 15Content management systems: integrated platforms are more consistentp. 24
- 16Page weight: almost three times the 2015 pagep. 25
- 17Third parties: on nine pages in tenp. 26
- 18Delivery: CDNs, caching and protocolsp. 27
- 19Fonts, resource hints and animationp. 28
- 20Field data and lab data: which number countsp. 29
- 21Speed and Google Search: a factor, not a guaranteep. 30
- 22Part A in one pagep. 31
- 23The most-quoted study: Milliseconds Make Millionsp. 32
- 24What 0.1 seconds was linked to, by verticalp. 33
- 25Where in the journey speed showed upp. 34
- 26The study's limits, in its own wordsp. 35
- 27Company case studies: what single sites reportedp. 36
- 28Loading cases: Vodafone, Renault and The Economic Timesp. 37
- 29Responsiveness case: redBus and INPp. 38
- 30Grading the evidencep. 39
- 31Measuring what speed is worth on your own sitep. 40
- 32How the NEWMIND model worksp. 41
- 33Revenue at stake: traffic and liftp. 42
- 34Revenue at stake: order value and conversion ratep. 43
- 35Loaded hourly ratesp. 44
- 36What speed work costsp. 45
- 37The lift that pays for the work in a yearp. 46
- 38Months to pay backp. 47
- 39Can you prove it? Test size and lengthp. 48
- 40Three worked scenariosp. 49
- 41Which input matters mostp. 50
- 42Playbooks: from evidence to actionp. 51
- 43Playbook 1: the main imagep. 52
- 44Playbook 2: server response and cachingp. 54
- 45Playbook 3: responsiveness (INP)p. 56
- 46Playbook 4: layout stability (CLS)p. 58
- 47Playbook 5: third-party tagsp. 59
- 48Playbook 6: fonts and early hintsp. 61
- 49Playbook 7: page weightp. 62
- 50Playbook 8: back, forward and next pagep. 63
- 51Playbook 9: keeping it fastp. 65
- 52Playbook 10: platform and agency decisionsp. 66
- 53A six-month roadmapp. 68
- 54Worksheet 1: speed baselinep. 69
- 55Worksheet 2: third-party tag inventoryp. 70
- 56Worksheet 3: main image checklistp. 71
- 57Worksheet 4: interaction logp. 72
- 58Worksheet 5: layout shift checklistp. 73
- 59Worksheet 6: revenue at stakep. 74
- 60Worksheet 7: project cost estimatep. 75
- 61Worksheet 8: test planp. 76
- 62Worksheet 9: performance budgetp. 77
- 63Worksheet 10: questions for vendors and agenciesp. 78
- 64Worksheet 11: twelve-month trackerp. 79
- 65Sources by typep. 80
- 66Method and limitsp. 84
- 67Glossaryp. 86
- 68Sourcesp. 87
Read the first 17 pages free.
Cover, contents, the executive summary, the three Core Web Vitals, the 75th percentile rule and pass rates from 2021 to 2025 and by site popularity, with their sources. Tell us where to send it and download it straight away.
Turn this benchmark into your custom workflow.
The benchmark shows where site speed & core web vitals stands against the market. NEWMIND PARTNERS designs, builds and runs the workflow that closes your gap, built around the systems you already use.
Feasibility audit
We measure your workflow on your data (read-only), then compare every solution on the market, off-the-shelf or custom.
Design
The workflow that fits your systems, suppliers and approval rules, with a human checkpoint on every decision that matters.
Build and run
Senior engineers build it, connect it to the tools you already use and run it in production with an audit trail.
Prove it
Baseline on day one, then the change in cost, hours and cash, reported every month.
Start with a feasibility audit
Find out what your workflow really costs, which solution fits best, and the business case to back it. No commitment.
Four steps. No disruption.
Intake call
30 minutes with a senior engineer to agree scope, KPIs and the data we need.
Data, read-only
Exports or read-only access to the systems involved. Nothing is changed.
Analysis
Your numbers placed against the published market references, KPI by KPI.
Readout
Your position, the cost gap in dollars and a costed roadmap to close it.
Questions or a custom benchmark?
Ask the team about the site speed & core web vitals report, or request a custom benchmark on your own data.
- Your Core Web Vitals vs the 2025 web
- Revenue at stake from a speed lift, modelled on your traffic
- Prioritised fix list by effort and impact
Sources
- 1HTTP Archive, Web Almanac 2025 — Performance chapter. almanac.httparchive.org/en/2025/performance
- 2Deloitte, Milliseconds Make Millions (37 brand sites, 30M sessions). deloitte.com/ie/en/services/consulting/research/milliseconds-make-millions.html
