The data behind every second of load time, and what a sub-two-second site is actually worth to your revenue.
Every second is a line item on the P&L
When a page slows down, a measurable share of visitors leave before it finishes loading, and they take their intent with them. The drop-off is not linear and it is not small, which is what makes speed a revenue question rather than an engineering preference.
Framed that way, a performance budget stops looking like a constraint and starts looking like a forecast. You are deciding, in advance, how much revenue you are willing to leave on the table.
A sub-two-second site is a commercial target
Under two seconds is not an arbitrary badge of honour. It is the threshold where a site stops feeling like software and starts feeling like a place, and the behaviour of buyers shifts accordingly.
Hitting it reliably is less about a single heroic optimisation and more about a budget that is defended on every release, so the gains you fought for do not quietly erode over the next six months.
The wins are boring, which is why they last
The durable improvements are unglamorous: right-sized images, fewer third-party scripts, content that arrives before the decoration. None of it makes a good conference talk, and all of it survives the next redesign.
We favour the boring wins precisely because they compound. A site that is fast by construction stays fast, while a site that is fast by heroics regresses the moment attention moves elsewhere.