When to leave WordPress isn't decided in a comparison table. I understood it through a three-line email to our server provider, the day before a booking launch.
It was a Monday in August, mid-afternoon. The next morning at 9:00, a guesthouse whose website we look after was opening its booking dates for the season.
We expected somewhere between 100 and 200 people trying to book at once, from nine o'clock until there was practically nothing left.
So I did what ten years of WordPress had trained me to do.
I emailed our server provider: we're worried we won't have enough RAM. Do you recommend upgrading anything?
More machine. The old reflex.
The reply came in under half an hour. It wasn't the one I expected.
Roughly: we can see the site is no longer a WordPress, it's now a Laravel app and it runs very well. No loading problems, each page loads what it needs and nothing more. On the server side, everything's fine.
Then they asked how many people, for how long, which pages. I told them. They closed with something like: at that level, the server and the code should cope without problems. If anything did choke, it would more likely be the code or the database tables than the machine, and even that would be rare.
Nobody suggested upgrading anything.
What struck me, and embarrassed me a little, wasn't their answer. It was my question.
The site had already moved on. I hadn't. I asked the way you ask about a WordPress before a big day: will the box hold? That's the question I'd carried around for a decade, and it came out on its own.
In my experience, a WordPress site loads a lot on every visit: the theme, the plugins, what that page needs and plenty it doesn't. When a site only shows things, it barely matters. When a site has to work, taking bookings, taking payments, holding 200 people pressing the same button, it starts to matter a lot. And the usual fix is to pay for a bigger server so the inside doesn't show.
If every problem with your website is solved by paying for more server, the server isn't the problem.
So if you ask me when to leave WordPress, these are the signs I'd look at today:
It depends on what your website is. I mean that, I'm not just being polite.
If your site is a homepage, four pages and a blog, stay on WordPress. I said it last week in WordPress has run out of excuses and I stand by it: for publishing content it's still a great tool. A decent theme, fewer plugins and hosting that isn't the cheapest you could find will fix most things.
If your site is the place where your business happens, the maths changes. A custom build used to be a luxury. With AI development tools it's much less of one. But, as I wrote a while ago, building software is getting cheaper and judgement isn't. Knowing what to build still costs exactly what it used to.
Then you probably don't need to leave it. Slow is a maintenance problem.
What I'm describing is a different kind of problem: a problem of shape. You're asking a tool built for publishing to behave like a booking system.
You notice the difference in the question you ask the day before something important. On a WordPress, you ask whether the machine will hold. On an application built for the job, the question becomes whether the code is good. That's a better question, because it's one you can actually answer.
That afternoon I asked the question of someone who still lived in WordPress. The website had already moved out. I hadn't, not entirely.
That's the last step in leaving WordPress, and nobody invoices you for it: changing your reflexes.
P.S. If, the day before your most important day, you find yourself emailing your host to ask for more memory, keep that email. It's the cheapest audit of your website you'll ever get.