Funnel & CRO
Turn slow load times into the sales you are quietly losing.
- Category
- Funnel & CRO
- Signup
- None
- Your data
- Stays in your browser
- Also in
- Claude, via MCP
How much revenue does a slow website lose?
Estimate it as visitors × conversion rate × value per conversion × the conversion lift per second saved. Published studies vary: Portent found B2B sites loading in 1 second converted 3 times better than at 5 seconds, and Deloitte found a 0.1-second mobile speed gain lifted retail conversions 8.4%. Validate the per-second rate with your own analytics.
Cutting load time from 4.5s to 2.0s could recover about $42,000/month, $504,000 a year. That’s the revenue case to hand engineering for a Core Web Vitals sprint.
The per-second penalty is an industry average, it’s steepest on mobile and for the first few seconds. Validate against your own analytics by load-time cohort.
About this calculator
Every extra second of load time is a small tax on conversion, well-documented but rarely translated into a number engineering leadership will act on. This calculator turns the gap between your current and target load time into a monthly revenue figure, using an assumed per-second conversion penalty, so a Core Web Vitals sprint has a business case attached, not just a Lighthouse score.
How to use it
- Enter monthly visitors and your current conversion rate for the page or flow you're assessing.
- Enter current load time and a realistic target load time, the gap between them is what gets modeled.
- Enter the value of one conversion and a conversion-loss-per-second rate (7% is a commonly cited industry average, adjust if you have your own data).
- Read seconds recovered, projected conversion rate, extra conversions per month, and the monthly and annual revenue that recovers.
Methodology
Seconds recovered is simply current load time minus target load time, floored at zero.
The uplift factor is 1 + (penalty per second ÷ 100) × seconds recovered, applied multiplicatively to your current conversion rate to get a projected new rate.
Extra conversions per month is visitors × (new conversion rate − current conversion rate). Monthly revenue recovered is extra conversions × value per conversion, and the annual figure is simply ×12.
The per-second penalty is a blended industry figure, not measured from your traffic, and it isn't linear in reality: the first couple of seconds after a slow load tend to cost more than the marginal second on an already-fast page. Treat this as a planning estimate to justify the work, then validate the actual lift against your own analytics once the fix ships.
The page speed revenue model in this calculator
Seconds recovered = max(0, Current load time − Target load time) Uplift factor = 1 + (Conversion loss per second ÷ 100) × Seconds recovered Projected conversion rate = Current conversion rate × Uplift factor Extra conversions / month = Monthly visitors × (Projected rate − Current rate) Revenue recovered / month = Extra conversions × Value per conversion; per year = × 12
- Current / target load time
- Seconds to load for the page or flow being assessed. Largest Contentful Paint (LCP) at the 75th percentile of real visits is the most useful measure.
- Conversion loss per second
- Assumed relative drop in conversion for each extra second. The tool default is 7%; it is an assumption, not a measured constant.
- Value per conversion
- Average order value, or lead value (deal value × lead-to-customer rate) for lead generation.
- Monthly visitors
- Visitors who reach the page being sped up, not total site traffic.
The uplift is linear and relative: each second saved raises the conversion rate by the same percentage of today's rate, with no compounding and no ceiling. Real data is not linear. In Portent's 2022 study, most of the change happened in the first few seconds. The often-quoted 7% figure comes from Akamai's 2017 retail report, which said a 100-millisecond delay can hurt conversion rates by 7% (eCommerceNews), so it is not a per-second rate. Treat the penalty input as a scenario, and run the model at a low and a high value.
Worked example: an Indian D2C store cutting mobile load time
A D2C skincare brand gets 2,00,000 mobile visitors a month to product pages that load in 4.5 seconds. The target is 2.5 seconds, the edge of Google's "good" LCP range. Conversion is 1.8% and average order value, after cash-on-delivery returns, is ₹1,500.
- Seconds recovered = 4.5 − 2.5 = 2
- At the 7% default: uplift factor = 1 + 0.07 × 2 = 1.14; projected conversion = 1.8% × 1.14 = 2.052%
- Extra orders = 2,00,000 × (2.052% − 1.8%) = 504 a month
- Revenue recovered = 504 × ₹1,500 = ₹7,56,000 a month, about ₹90.7 lakh a year
- At a cautious 3% per second: uplift 1.06, conversion 1.908%, 216 extra orders, ₹3,24,000 a month (₹38.9 lakh a year)
The business case is somewhere between ₹38.9 lakh and ₹90.7 lakh a year depending on the assumed penalty, and the low end is the one to defend in a budget meeting. If the speed work costs less than a few months of the cautious figure, it is worth doing; then measure the real lift by comparing conversion for fast and slow sessions before and after release. Use delivered order value, not placed order value, for Indian D2C, as the RTO-adjusted ROAS calculator explains.
Average conversion rate by page load time (2022)
| Load time | B2B lead generation sites | B2C ecommerce sites |
|---|---|---|
| 1 second | almost 40% | 3.05% |
| 2 seconds | 34% | 1.68% |
| 3 seconds | 29% | 1.12% |
Effect of a 0.1-second mobile speed improvement (2019 data)
| Sector | Conversions | Average order value |
|---|---|---|
| Retail | +8.4% | +9.2% |
| Travel | +10.1% | +1.9% |
What is a good page load time?
Use Google's Core Web Vitals thresholds as the target and your own data as the business case:
- LCP of 2.5 seconds or less is "good" and above 4.0 seconds is "poor", measured at the 75th percentile of page loads, split by mobile and desktop (web.dev).
- Speed gains do show up in revenue in controlled tests: in Vodafone's A/B test, a 31% LCP improvement led to 8% more sales (web.dev case study).
- The largest losses are on slow mobile pages, which is where to start.
A decision rule: if your 75th-percentile LCP is above 4 seconds on mobile, fix it before spending more on traffic, because every paid visitor pays the speed tax. If it is already under 2.5 seconds, further speed work usually has a smaller return than offer, copy and form changes; see the conversion funnel calculator and CRO consulting.
Page speed in India vs the US
Indian ecommerce and lead generation traffic is heavily mobile, often on mid-range Android phones and variable networks, so test on a throttled mid-range device rather than a developer laptop. Heavy tag stacks are a common cause of slow pages: several pixels, chat widgets and checkout scripts from providers such as Shiprocket or GoKwik all load on the same page. Moving tracking server-side can remove some of that weight; see Shopify server-side tracking. Value per conversion should be net of cash-on-delivery returns, or the revenue case will be overstated.
US desktop share is higher for B2B, so compare mobile and desktop separately; a B2B site can pass on desktop while failing on mobile. Paid search landing pages matter most, because slow pages waste media spend and can affect ad quality. Consent banners and tag managers often add delay; a GA4 audit usually finds tags that can be removed or deferred. Deloitte's study included US brands, but its effects came from improvements across the whole journey, not one page.
Common page speed ROI mistakes
- Treating the 7% Akamai figure as a per-second rate; it referred to a 100-millisecond delay in retail data.
- Using total site traffic as visitors when only some pages will get faster.
- Measuring speed in a lab test on a fast connection instead of real-user LCP at the 75th percentile.
- Assuming the uplift is linear and unlimited, which overstates gains from pages that are already fast.
- Using placed order value for cash-on-delivery stores, where returns reduce real revenue.
- Shipping the speed fix without measuring conversion by load-time segment before and after.
Sources
- Site Speed is (Still) Impacting Your Conversion Rate, Portent. Accessed 2026-10-18.
- Milliseconds Make Millions, Deloitte Ireland (commissioned by Google). Accessed 2026-10-18.
- Milliseconds make millions, web.dev (Google). Accessed 2026-10-18.
- Vodafone: A 31% improvement in LCP increased sales by 8%, web.dev (Google). Accessed 2026-10-18.
- Largest Contentful Paint (LCP), web.dev (Google). Accessed 2026-10-18.
- Millisecond delays can hurt conversion rates, online transactions: Akamai report, eCommerceNews Australia (reporting Akamai's 2017 State of Online Retail Performance report). Accessed 2026-10-18.
FAQ
There is no universal figure. Published studies (for example Portent's 2022 analysis of conversion rate by load time) show conversion falling as load time rises, most steeply in the first few seconds, so the effect is not linear. The often-quoted 7% from Akamai's 2017 report is per 100 milliseconds of delay, not per second. Use a lift measured on your own site where possible, and treat this tool's linear model as a rough estimate.
Largest Contentful Paint is usually the best proxy since it's closest to when a visitor perceives the page as usable, and it's what Core Web Vitals and Google's ranking signals track. Time-to-interactive matters more for pages with heavy above-the-fold interactivity.
Segment your analytics by load-time cohort before and after the change, and compare conversion rate for visitors who experienced fast vs. slow loads in the same period. That gives you an actual, not modeled, per-second relationship for your traffic.
Core Web Vitals, including LCP, are part of Google's page experience signals, but relevance and content quality matter far more. The stronger case for speed is conversion: faster pages convert more of the traffic you already have, which is what this calculator measures.
LCP measures when the largest visible element, usually the main image or headline, has rendered. Full page load includes everything else, including scripts that load after the page looks ready. LCP is closer to what a visitor experiences as "loaded", so it is the better input for this tool.
Split sessions by load time in analytics and compare conversion rates for fast and slow groups before and after release, over the same weeks to avoid seasonal effects. Where possible, run an A/B test like Vodafone's, with half the traffic on the faster page.
Large uncompressed images, too many third-party scripts (analytics, ad pixels, chat, heatmaps), render-blocking fonts and slow server response. Removing unused tags and compressing hero images often gives the largest gains for the least work.
Numbers looking off?
If the model says one thing and the pipeline says another, that gap is usually where the revenue is. A 30-minute call finds it.
It starts with a 30-minute call. Pick a time below.