Sooner or later a developer or an agency tells you the site is slow. The awkward part, for a marketer, is the question that follows: “so how much are we losing?” Search around and you will find plenty of sentences claiming that a second of delay costs a particular percentage of conversions — but every one of those numbers comes from somebody else’s site. Copy a figure from a different sector, a different device mix and a different traffic source into your own report, and the first time someone challenges it the rest of your numbers lose their credibility too.
Two things are more useful. Knowing by what route speed reaches revenue, and knowing how to confirm each route with your own data.
Route 1 — the people who leave before it renders
While the first screen is still blank, the visitor is looking at nothing. If they leave at that point they leave without a trace anywhere in your site. In analytics it shows up only as sessions with no engagement, which means a single bounce figure cannot tell you whether the content was wrong or the page was late.
Look instead at metrics split by entrance page and by device. If engaged time on mobile is markedly shorter than on desktop, you have grounds to suspect loading before you blame the copy. Paid landing pages are the clearest case: that traffic is mostly mobile, mostly first-time, and therefore arriving with an empty cache.
Route 2 — search engines wait as well
A crawler is also something making requests to your server. When responses are slow, fewer pages get fetched in the same window, and on a large site that shows up as how quickly new articles reach the index. Search Console reports average response time in its crawl statistics, so open that screen rather than guessing.
On top of that, Google has publicly stated that real-user page experience metrics are among its ranking signals. It is not the factor that decides a ranking on its own, but it is the sort of signal that separates documents of otherwise similar quality.
Route 3 — where the ad budget shifts
Search advertising platforms fold landing page experience into their quality assessment. Whether the same bid buys you a better position, or the same position costs you more, is decided partly there. Your ad account exposes a landing page assessment, so you can see how the platform currently reads your page instead of inferring it.
Producing your own numbers
The rules are simple. Same URL, same device profile, same time of day, three runs, take the median. A single run swings with network conditions, so it never goes into a report. Then measure once before the work and once after — only two measurements under identical conditions let you claim that anything changed.
Always place the speed figure next to conversion data. Pick one paid landing page, compare the conversion rate over a few weeks before against the same span afterwards, and you can state what speed is worth on your site without quoting anyone. How that discipline runs in practice is published on our process page, and the rest of the performance writing sits in the performance archive.
Next part
Having decided to measure, you have to open a tool. The next part is how to read a Lighthouse report — and to give away the conclusion, the score is only a summary. The work lives in the itemised list underneath it.