Designing for the ideal (or for the <98%) is fine. As long as the experience is gracefully degraded for the rest.
Profit is often at the margin.
Taking the conversation slightly outside the original context: if I go to a restaurant, should I have a phone and an app ready so that I can order food? If I go to the gym, should I have a gym app ready so that I can sign in? I don't like having to do that. But that's just another instance of this same attitude.
And I don't give a shit about your site/content/whatever. If you don't work with Firefox or my old Mac browser, your whatever isn't worth my time. For "content" sites this is insanely true, even for "news".
As of 2024 at one of my clients we were still supporting IE8 and as of 2026 I still have significant traffic at some clients from IE9 and 11 or ancient firefox/chrome versions.
The reason is quite simple when you analyze the data: it's concentrated between 8.30 and 5.30 pm.
Those are people sitting at their desk in a bank or some different office. They cannot install other browsers, they cannot update them. Their perfectly working computers (for their job) may not even support newest browsers at all.
Losing 2-6% of the office hours traffic of those well paid-stable job individuals has an outsized impact on revenue and margins that cannot be estimated by naive data analysis.
In other sectors many users are B2B2C retailers in machinery or carpentry using the same computer they bought 15 years ago and they need to provide a quote to the customer in front of them. Single orders can easily be 5 or even 6 figures.
Small numbers in many sectors not only matter they have an outsized impact and a compounding effect long term.
But being forced to use ancient Chrome/Edge versions? You are exposing your users to half a dozen sandbox escapes, and there's no big blocker from upgrading. Is their IT department asleep at the wheel?
It's been 15 years since IE9. Where do you draw the line?
Often you simply don't offer the feature. E.g 3d rendered previews may not be available but product configuration and cart keeps working on a shop selling custom showers (you fallback to dynamic static images).
In real estate a page displaying fancy maps with price statistics by area/neighborhood might be unavailable, but the core business of listings and search does.
Fifteen years! Unless it's a government agency what's the point even in doing business with a company that uses 15 y.o. browser? They will pay you in silver coins according to 2011 prices.
And they mostly check your website when they are bored at work. Not when they leave it and have kids, hobbies or a household to care for.
In travel sector users predominantly navigate in office hours from their work devices. You go meet them where they are. 4% of 6 million daily users is 240'000 potential customers. Converting 3% of them means millions at the end of the year.
Maybe some like airbnb have (or at least used to have) a unique catalogue and they can play a different game and afford to lose some money.
Most e-commerces play differently, at different scales and enjoy different moats and different shareholders/owners expectations.