Calculator

To turn "the site looks dated" into a number somebody can approve or refuse. A redesign is usually argued on taste and paid for out of a budget that answers to arithmetic, which is why the argument goes badly.

The useful output is not a return figure. It is the break-even: how much the conversion rate, the order value or the support load would have to move for the project to wash its face. Once that is on the page, the conversation stops being about taste and starts being about whether that movement is realistic.

Start from your own numbers, not from a benchmark. Traffic, conversion rate, average order value, and the full cost of the project including the internal time nobody invoices for.

Then solve backwards for the break-even rather than forwards from a hoped-for uplift, because a forward estimate is just your optimism with a decimal point. Run it at three levels: the movement you expect, half of it, and none. The middle column is the one to plan against. Put a date on when the change would have to land, since a return that arrives two years late is a different decision.

An uplift percentage gets borrowed from a case study about somebody else. Those figures are real for the site they came from and tell you nothing about yours, and a borrowed number is the fastest way to a business case that cannot be defended in the room.

The second failure is counting the build and forgetting the rest: content, migration, training, the weeks after launch when things are worse before they are better. The third is having no plan to measure afterwards. If nobody agrees now how the result will be read, the project will be judged on whether people liked the look of it, which is where you came in.