Site Speed Conversion Rate Impact (2026)

TL;DR: Site speed conversion rate impact is a three-rung ladder, not one “8.4% per 0.1s” law. Deloitte’s 2019 mobile study finds retail +8.4% conversion and +9.2% AOV when four load metrics improve 0.1s. Google’s Core Web Vitals pass bar is LCP 2.5s, INP 200ms, CLS 0.1. Yottaa’s 2025 index shows conversion still rising when LCP moves from 2.5s to 1.3s (1.50% → 2.25%).

Introduction

Merchants Google site speed conversion rate impact and get a recycled tenth-of-a-second slogan. “Deloitte proved 8.4%.” Then the same deck slaps that number onto Largest Contentful Paint, Interaction to Next Paint, and a Shopify theme app. Those are not the same meters. One is a 2019 mobile correlation on four deprecated or proxy metrics. The others are 2026 field vitals.

Site speed conversion rate impact matters when you decide whether to pay for image CDN, app-script diets, and INP work—or keep shipping a “passed CWV” homepage that still converts like a 4-second PDP. This page sits next to shopify conversion rate benchmarks, ecommerce bounce rate benchmarks, product page conversion rate benchmarks, and add to cart rate benchmarks by niche.

• Deloitte’s Milliseconds Make Millions study (commissioned by Google, data from Fifty-Five) analyzed 37 retail, travel, luxury, and lead-generation sites across Europe and the US over four weeks at the end of 2019—about 30 million user sessions overall, and 20.5 million in the 0.1s-improved subset (Deloitte PDF; web.dev).
• Retail saw +8.4% conversions and +9.2% average order value on that 0.1s improvement; travel saw +10.1% conversions and +1.9% AOV (Deloitte).
• Google’s current Core Web Vitals “good” bars are LCP within 2.5 seconds, INP 200 milliseconds or less, and CLS 0.1 or less, judged at the 75th percentile (web.dev).
• Yottaa’s Web Performance Index as of September 23, 2025 finds LCP 2.5s → 1.3s conversion 1.50% → 2.25% (about 50% relative) and INP 200ms → 50ms conversion 3.91% → 4.44% (about 13.5% relative) (Yottaa).
• There is no public dataset that turns Deloitte’s 8.4% into a universal “percent conversion per second of LCP,” and no public dataset for feat. LCP mix. Refuse those mashups.

What Is Site Speed Conversion Rate Impact

Site speed conversion rate impact is the measured change in purchase (or funnel) completion when pages become faster to load, respond, and stay visually stable—compared with a slower control on the same journey.

It is not one KPI. Operators mash three jobs. First is a historical correlation: what happened when mobile load metrics improved by 0.1s on audited brand sites in 2019. Second is a search-and-UX pass bar: Google’s Core Web Vitals thresholds. Third is a beyond-pass conversion curve: whether conversion keeps rising after you already “pass.” Deloitte owns the first with FMP, estimated input latency, observed load, and TTFB. Google owns the second with LCP, INP, and CLS. Yottaa’s 2025 index is the cleanest public third rung we found this run (Deloitte PDF; web.dev; Yottaa).

This article owns the speed-to-money ladder. Session-purchase medians live in Shopify conversion rate benchmarks. Bounce as an experience meter lives in ecommerce bounce rate benchmarks. Do not grade a Lighthouse lab score against Littledata’s 1.4% session CVR.

Why Site Speed Conversion Rate Impact Matters

Finance will not fund a CDN on vibes. You need labeled rungs before you promise “8.4% if we shave 100ms off LCP.”

• The 0.1s study is real and dated. Deloitte isolates speed on 37 brands with no UX redesigns during the window. Retail conversion +8.4% and AOV +9.2% are the board numbers. They are mobile, four-metric, 2019 (Deloitte).
• Passing CWV is not a conversion ceiling. Google’s good bars (LCP 2.5s, INP 200ms, CLS 0.1) are 75th-percentile field targets, not revenue optima (web.dev).
• Beyond pass still prints conversion. Yottaa reports LCP 2.5s → 1.3s moving conversion from 1.50% to 2.25%, and INP 200ms → 50ms from 3.91% to 4.44% (Yottaa).
• Load and bounce move together. On that index, 2s load sits at 2.05% conversion and 45.63% bounce; 3s sits at 1.59% conversion and 57.81% bounce (Yottaa). Pair with your bounce ladder.
• CLS is a visibility job more than a CVR job in that panel. Conversion stays 1.99% at CLS 0.10 and 0.15. Fail CLS and you still risk search visibility (Yottaa; web.dev).

How Site Speed Conversion Rate Impact Works

Score speed with three labeled rungs. Quote Deloitte’s 0.1s correlation, Google’s pass bar, and Yottaa’s beyond-pass conversion separately. Do not write “every second costs 7%” unless you have that panel.

What Deloitte’s 0.1s finding actually measures

Google commissioned Deloitte Ireland; Fifty-Five collected the data. 37 sites qualified after analytics and technical audits. The window was four weeks at the end of 2019. The PDF counts about 30 million sessions in the study and 20.5 million in the slice where each of four metrics improved 0.1s: First Meaningful Paint (now deprecated—web.dev says use LCP), estimated input latency, observed load, and max server latency / TTFB. No UX redesigns ran during the window (Deloitte PDF; web.dev).

That design answers: when mobile speed improved a tenth of a second across those metrics, did funnel and revenue metrics move? It does not answer: if my LCP is 2.6s, will cutting 100ms add 8.4% Shopify conversion?

Vertical (Deloitte 0.1s mobile)	Conversion	Other labeled metric
Retail	+8.4%	AOV +9.2%; PLP→PDP +3.2%; PDP→ATC +9.1%
Travel	+10.1%	AOV +1.9%
Luxury	+3.6%	Page views per session +8.0%; PDP→ATC +40.1% (wishlist caveat)
Lead generation	Conversion decreased (PDF)	Form first-step→submit +21.6%; mobile bounce −8.3%

Source: Deloitte / Google / Fifty-Five, Milliseconds Make Millions. https://www.deloitte.com/content/dam/assets-zone2/ie/en/docs/services/consulting/2023/Milliseconds_Make_Millions_report.pdf.

Luxury’s +40.1% add-to-basket step is real and misleading if you treat basket as purchase. The PDF notes luxury shoppers often use add-to-basket as a wishlist. Lead generation’s form lift sits next to falling conversion—offline meetings still ate the sale. That honesty is more citable than a fake all-vertical 8%.

What “passing Core Web Vitals” actually is

Google’s current Core Web Vitals are LCP (loading), INP (interactivity), and CLS (visual stability). Good experience: LCP within 2.5s, INP 200ms or less, CLS 0.1 or less, at the 75th percentile of field loads, split mobile vs desktop. A page passes only if all three hit those targets. INP is the stable interactivity metric; FID was retired in 2024 (web.dev).

Pass is a floor for Search Console and UX. It is not Deloitte’s 8.4%. It is not Yottaa’s 1.50% vs 2.25%. Lab Lighthouse scores are not field CrUX. Measure field, then spend.

What Yottaa’s beyond-pass index adds

Yottaa’s October 1, 2025 post cites its Web Performance Index as of September 23, 2025 and says the index updates daily. Treat the figures as a dated vendor panel, not a census of the web.

On that snapshot: cutting LCP from Google’s 2.5s pass to 1.3s lifts conversion from 1.50% to 2.25%—about a 50% relative move. Cutting INP from 200ms to 50ms lifts conversion from 3.91% to 4.44%—about 13.5% relative. A 2s load converts at 2.05% with 45.63% bounce; a 3s load converts at 1.59% with 57.81% bounce. CLS 0.10 vs 0.15 conversion is flat at 1.99% (Yottaa).

Do not average Yottaa’s 1.50% with Littledata’s Shopify 1.4%. Different panels, different clocks. Use Yottaa to argue beyond pass; use Shopify CVR to grade session purchase.

Speed Beyond Pass Ladder (comparison table)

Ladder rung	What it measures	Published figure	Source
0.1s correlation	Revenue/funnel when four 2019 mobile metrics improve 0.1s	Retail +8.4% CVR / +9.2% AOV	Deloitte
Pass bar	Google “good” field thresholds	LCP 2.5s / INP 200ms / CLS 0.1 (p75)	web.dev
Beyond-pass LCP	Conversion at pass vs faster LCP	1.50% → 2.25% (2.5s → 1.3s)	Yottaa WPI 2025-09-23
Beyond-pass INP	Conversion at pass vs faster INP	3.91% → 4.44% (200ms → 50ms)	Yottaa
Load vs bounce	2s vs 3s page load	CVR 2.05% vs 1.59%; bounce 45.63% vs 57.81%	Yottaa
Universal % per second of LCP	One law pooling 2019 + 2025 meters	no public dataset	—

How this differs from bounce and Shopify CVR

Bounce can fall while purchase stays flat. Deloitte’s lead-gen row already shows bounce improving while conversion fell. Grade bounce on the ecommerce bounce page. Grade money on orders ÷ sessions with a named panel. Speed work that only moves bounce is still a UX win—just not an 8.4% retail conversion win.

How to Improve Site Speed Conversion Without Fake 8.4% Dreams

1. Name the meter in the ticket. Write Deloitte-style load correlation, CWV pass, or beyond-pass LCP/INP. Do not mix them mid-sprint (Deloitte PDF).
2. Pass field CWV first. Hit LCP 2.5s, INP 200ms, CLS 0.1 at p75 on mobile and desktop in CrUX / Search Console (web.dev).
3. Then spend past the pass bar. Treat Yottaa’s LCP 2.5s → 1.3s (1.50% → 2.25%) as the argument for remaining image and TTFB work (Yottaa).
4. Fix INP on tap-to-cart, not only hero LCP. 200ms → 50ms is a 13.5% relative conversion move in that index—the milliseconds that hide in app scripts (Yottaa).
5. Watch bounce and purchase on the same cohort. A 3s load’s 57.81% bounce vs 2s at 45.63% is a leak; confirm orders still move (Yottaa; bounce benchmarks).
6. A/B speed with conversion as the decision metric. Refuse “every second costs 20%” slides. There is no public dataset that lawfully pools Deloitte 8.4% onto 2026 LCP seconds.

Frequently Asked Questions

Q: What is site speed conversion rate impact in 2026? A: Use the Speed Beyond Pass Ladder. Deloitte’s 2019 mobile study ties a 0.1s improvement to retail +8.4% conversion and +9.2% AOV. Google’s pass bar is LCP 2.5s, INP 200ms, CLS 0.1. Yottaa’s 2025 index shows conversion still rising beyond those bars.

Q: Does passing Core Web Vitals increase conversion? A: Passing is a field UX and Search Console baseline, not a guaranteed conversion lift. Yottaa’s panel finds additional conversion when LCP improves from 2.5s to 1.3s (1.50% → 2.25%) and INP from 200ms to 50ms (3.91% → 4.44%). Pass first, then keep going.

Q: Is Deloitte’s 8.4% the same as 8.4% per 0.1s of LCP? A: No. The 8.4% is retail conversion on a 2019 mobile subset where four metrics—including deprecated First Meaningful Paint—each improved 0.1s. LCP did not exist as a Core Web Vital in that design. Do not paste it onto a 2026 Lighthouse ticket.

Q: What load time is associated with higher ecommerce conversion? A: In Yottaa’s September 23, 2025 snapshot, 2s load converts at 2.05% versus 1.59% at 3s, with bounce 45.63% vs 57.81%. That is a vendor index, not a Shopify median. Grade your store against your own baseline and Shopify CVR.

Q: Can I average Deloitte, Yottaa, and Shopify 1.4% into one speed benchmark? A: No. There is no public dataset that cleanly pools those denominators into one “percent per second.” Quote each rung with its publisher and year, or you will invent a fake industry law.

Conclusion

Site speed conversion rate impact rewards merchants who label the ladder. Deloitte’s 0.1s mobile correlation is still the best public tenth-of-a-second evidence—retail +8.4% conversion, +9.2% AOV. Google’s LCP/INP/CLS bars are the pass. Yottaa’s 2025 index is the beyond-pass conversion curve. The “8.4% forever, on every metric” slide was never a 2026 LCP law.

If you want distribution that already converts attention into affiliate-sold storefronts, list on feat..