What the evidence will carry
Current revenue: $170,000 a month
Worth closing the priced part of the gap
A band across three randomised results from three unrelated organisations, not a single number. Each bound names the experiment it comes from.
Your gap, split at the point the evidence stops
- Improvement you asked for
- 1.6 s
- Priced
- 0.5 s
- Left unpriced
- 1.1 s
Why part of it is not priced
A straight line is validated to about half a second. The same company measured the shape directly at five delay levels in 2009. Revenue per user fell 1.2% at 500 ms, 2.8% at 1000 ms and 4.3% at 2000 ms. So the marginal cost of 100 ms roughly doubles up to a second, then halves. Their own conclusion was that delays under half a second impact business metrics. So this prices the first half second and declines to price the rest.
What one tenth of a second is worth, both answers
30.6x apart
These are the same quantity from the two best sources in the field, at your numbers. The lower one is a controlled experiment: 10% of Bing users were deliberately slowed by 100 ms and another 10% by 250, for two weeks. The higher one is a correlational study of 37 brands whose speed moved on its own, and it is the side every other calculator reports.
The band above uses the randomised side. That choice is the whole difference between this tool and the rest, and the method page sets out the four reasons for it.
What this cannot tell you
Read these before the number leaves this page.
- Every causal source measured server delay. This tool reads LCP, because LCP is what field data exposes. That substitution is the largest single weakness in the exercise.
- Bing's figure is from web search, Zalando's from fashion retail, Vodafone's from a telecoms landing page. None of them is your site.
- The elasticity is local. The same authors write that if a site's performance or its audience changes, the impact may be different.
- Not every millisecond counts the same. A 250ms delay to below-the-fold elements at Bing produced no detectable effect across nearly 20 million users.
- A linear approximation is validated to roughly half a second. Past that the curve bends, and the doses Bing published show the second second costing about half what the first one did. Anything beyond that half second is reported as distance, never as money.
- Etsy ran the same kind of experiment and found a 200ms delay did not matter for its users. The effect is real but not universal.
Take the business case with you
A PDF of about three pages with the figures above, how they were calculated, what they cannot tell you, and every source with its design and DOI. Written to be forwarded to someone who was not here.
One email with the PDF attached. The result above stays open whether or not you fill this in.
Where these numbers come from
The three experiments the band is built from. All three assigned speed to users, which is what lets a figure claim that making a site faster would gain you something.
Tier A, causal
Vodafone: A 31% improvement in LCP increased sales by 8%
Vodafone, published by Google on web.dev, 2021
Randomised controlled experiment
web.devTier A, causal
Online Controlled Experiments at Large Scale
Kohavi, Deng, Frasca, Walker, Xu, Pohlmann (Microsoft), 2013
Randomised controlled experiment, peer reviewed
doi:10.1145/2487575.2488217Tier A, causal
Loading Time Matters
Zalando engineering, 2018
Randomised controlled experiment
engineering.zalando.comTier B, observational
Milliseconds Make Millions
Deloitte with Google, 2020
Observational
www.thinkwithgoogle.com
If the figure is large enough to act on, the next question is which layer the time is in. That is read from your own field data rather than from a score.
How the work itself runsLet’s Create an Amazing Project Together!
Questions people ask about this
How much does page speed actually affect conversion rate?
The randomised experiments measure roughly 0.3% to 0.7% of revenue per 100 ms. The observational study the rest of the market quotes measures about 18% for the same step. This tool uses the randomised side, which is why its figure is around thirty times smaller than every other speed calculator's. The method page sets both out with their sources.
Why is this number so much lower than other calculators?
Two reasons. It prices one tenth of a second from controlled experiments rather than from a correlation, and it refuses to compound that figure across seconds nobody measured. Compounding the market's coefficient over a 1.6 second improvement claims a shop can gain fourteen times its own revenue, and that is where the large numbers come from.
Where do I find my LCP?
Put your domain into the lookup on this page. It reads the Chrome User Experience Report for your origin: real visitors, phone first, at the 75th percentile. That is the same number Search Console grades you on. If your site has too little traffic to appear in that dataset, the free audit gives you a lab measurement instead.
Why does it refuse to price more than half a second?
Because that is where the evidence stops. The experimenters validated a straight line to about half a second, and their own longer tests came in under that line. Past it the curve bends and nobody has published by how much. The rest of your gap is reported in seconds rather than in money.
Does anything I type get sent anywhere?
No. The arithmetic runs in your browser. The one request this page can make is the optional field-data lookup, and only when you ask for it. An email address is needed only if you want the PDF sent to you.