The eleven second booking page
What a slow page actually costs, measured in customers rather than seconds.
The worst number I measure on small business websites is not on the homepage. It is on the booking page, because that is the page that carries the money and it is usually the page carrying the most software.
Here is how it happens. The site itself is fine. Then a booking widget arrives, and it brings a calendar library, an analytics bundle, two fonts of its own and a chat bubble nobody asked for. Each piece was added on a different day by a different person and every one of them made sense at the time. Nobody ever loaded the finished page on a phone, on mobile data, in a car park.
I did. Eleven seconds.
The research on what that costs has been public for years. Google measured that the probability of a visitor bouncing rises 32 percent as load time goes from one second to three. Deloitte measured that improving speed by a tenth of a second moved real conversion numbers at real retailers. None of this is exotic. A slow page is a shop with a jammed door, and people do not wait at jammed doors when the shop next door opens with one tap.
The fix is rarely clever. It is subtraction. Measure what the page ships, name what each piece is for, and remove everything that cannot answer. Then set a budget so the weight cannot creep back, because it always tries to creep back.
Two things matter when you check your own booking page. First, test it on a phone over a cellular connection, not on the office wifi, because your customers are not on the office wifi. Second, time the moment the largest thing on the screen appears, which is the number Google calls LCP. Under two and a half seconds is acceptable. Under one second is where it stops being a topic.
This is the kind of thing I fix.
Ashley Mbah
Sources