Speed is revenue: how much each extra second costs you in sales
A 0.1-second improvement raised retail conversion by 8.4% in a Google study with Deloitte. Speed isn't a technical detail — it's a direct revenue lever. See the math and where the seconds hide in your store.
- The real numbers on how much speed moves conversion and order value — not opinion, studies with tens of millions of sessions.
- How to translate "1 extra second" into lost revenue per month in your store.
- Where the seconds hide: TTFB, LCP, JavaScript and third-party tags.
- Why trusting Lighthouse alone is misleading — and what to measure in the field.
There's a Google study with Deloitte that measured 30 million mobile sessions and found a number that should be taped to every e-commerce team's monitor: a mere 0.1-second improvement in load time raised conversion by 8.4% in retail and average order value by 9.2%.[1] It's not that 0.1 second changes everything — it's what 0.1 second reveals: speed is a direct revenue lever, not a technical detail.
The math nobody does: seconds turn into money
"My site is a bit slow, but it sells" is the most expensive sentence in e-commerce. The problem is that the sale lost to slowness shows up nowhere — nobody files a ticket saying "I gave up because it took too long".
Let's put a number on it. A store with 100,000 visits/month, a R$ 200 average order and a 1.5% conversion rate brings in about R$ 300,000/month. If the home and product pages are at ~4 seconds today and you get them to ~2 seconds — the range where e-commerce conversion is highest per Portent[2] — a conservative 15% lift in conversion (1.5% → ~1.7%) means:
Before: 100,000 visits × 1.50% × R$ 200 = R$ 300,000/mo
After: 100,000 visits × 1.72% × R$ 200 = R$ 345,000/mo
─────────────────
Monthly gain from speed alone: R$ 45,000/mo
Annual gain: R$ 540,000/yrHalf a million reais a year that was being left on the table — without more traffic, without cutting price, without a new campaign. That's why performance competes for budget with paid media, not with "IT improvements".
Why the studies all hit the same note
Skepticism is healthy — one study can be luck. But when independent studies, with huge samples and different methods, reach the same conclusion, it becomes a practical law:
| Study | Sample | Core finding |
|---|---|---|
| Deloitte + Google (2020) | 30M mobile sessions | 0.1s faster → +8.4% retail conversion |
| Portent (2022) | 94M pageviews, 10 stores | Peak conversion at 0–2s; drops sharply after 5s |
| Google / SOASTA | Millions of mobile pages | From 1s to 3s, bounce probability rises ~32%[3] |
| Amazon (classic) | Internal tests | Every 100ms of latency cost ~1% of sales[4] |
Most people overestimate how long they'll wait. In a lab they say "3 seconds, fine". In real behavior, measured across millions of sessions, they've already started leaving.
Synthesis of the Think with Google performance reports
Where the seconds hide in your store
"Slow site" is a layperson's diagnosis. Performance splits into layers, and each has a typical culprit:
- TTFB (server): time to first byte. Above ~800ms it's broken caching, an overloaded database, or weak hosting. It's the #1 bottleneck in mid-sized stores — see how to diagnose it in Magento 2 and WooCommerce.
- LCP (largest visible element): almost always the banner image or the product gallery. It's the Core Web Vitals metric that moves conversion the most — there are 7 techniques in order of impact just for it.
- JavaScript and hydration: heavy theme, page builder, giant bundle. The page "appears" but doesn't respond to taps — that's a bad INP.
- Third-party tags: pixel, chat, heatmap, A/B testing tool. Every external script is a request to a server you don't control. Three or four "little tags" add up to seconds.
How to measure the real impact (without fooling yourself)
Scoring 100 on Lighthouse and still losing sales to slowness is more common than it seems. Lighthouse measures a single simulated session, on a Google server, with no cart, no login, and none of your real customer's shaky 4G.
To understand the financial impact you need field data: what real users actually experience. The two sources that matter:
- CrUX (Chrome User Experience Report): the real Core Web Vitals from Chrome, at the 75th percentile. It's what Google uses for SEO and what reflects the actual customer — understand the difference between LCP, CLS and INP in practice.
- RUM (Real User Monitoring): measures every real session, in real time, segmentable by device and region. See when to use each in synthetic monitoring vs RUM.
Mobile is where it hurts most
Most e-commerce traffic in Brazil is mobile, on slower networks and more modest devices than the developer's laptop. It's no coincidence that the Deloitte study measured precisely mobile sessions. If you only test on the office Wi-Fi, on a top-of-the-line desktop, you're measuring the experience of the people who least represent your customer. Test with 4G throttling and an entry-level device — that's where the sale slips away.
The honest summary
Speed is the conversion optimization with the best cost-benefit there is: you don't pay per click, you don't give a discount, you don't build a campaign. You just stop scaring away the people who already decided to buy. Treat the seconds on your home and product pages for what they are — a revenue line.
References
- Deloitte Digital, Google & fifty-five. Milliseconds Make Millions, 2020 — 30 million mobile sessions. web.dev/case-studies/milliseconds-make-millions
- Portent. Site Speed is (Still) Impacting Your Conversion Rate — 94 million pageviews across 10 stores. portent.com
- Google / SOASTA. The State of Online Retail Performance and Find out how you stack up to new industry benchmarks for mobile page speed, Think with Google. thinkwithgoogle.com
- Akamai and the classic Amazon/Google accounts on latency and revenue. akamai.com
- web.dev (Google). Why does speed matter? and Core Web Vitals documentation. web.dev/why-speed-matters
Frequently asked questions
- How much does one extra second of load time cost in sales?
- It depends on the store, but the strongest reference is Deloitte's study with Google: every 0.1 second of improvement raised retail conversion by 8.4% and average order value by 9.2%. Portent shows e-commerce conversion peaks between 0 and 2 seconds and drops sharply after 5. In revenue terms, a second is often worth a double-digit slice of conversion.
- Does a 100 Lighthouse score mean my store is fast?
- Not necessarily. Lighthouse measures a single simulated session, on a Google server, with no cart or login and on an ideal network. What affects SEO and conversion is field data (CrUX, 75th percentile) and RUM — the experience of real users, on their phone and their network. You can score 100 in the lab and still lose sales in the field.
- Should I prioritize LCP or TTFB?
- Start with whatever is worse in the field. A high TTFB (above ~800ms) blocks everything else and points to caching, server or database — fix it first. With the server responding fast, LCP then depends on the main image (hero or gallery), which is where most of the problem lives on a product page.
- Does speed affect SEO or only conversion?
- Both. Core Web Vitals is a Google ranking signal, so poor speed can push positions down. But in e-commerce, the conversion impact is usually even bigger and faster than the SEO one — you lose the sale right away, before you even lose a ranking position.
Keep reading
Monitor all of this automatically
Especialista Loja Virtual runs real browser checks on your site every few minutes, alerts on Discord, Slack or email and shows a screenshot of the incident. Start free.