Php allowed my team to move amazingly fast and deliver what the business needed very quickly. In e-commerce and dealing with Amazon specificity, rules and methods change monthly and you always have to be able to quickly pivot to remain competitive and I think php was a big factor in allowing this to happen.
It has no compile time, the development cycle is crazy fast, the stdlib is good for the web, every problem is solved so they’re all googleable, it has a good package manager now, great frameworks, good testing stuff, there’s a lot of positives atm
And because you don't need build tools you can deploy it as easily years later.
Also while developing you can just hit F5 or rerun a PHP script in your console compared to waiting for a build.
Python devs don't necessarily know web development, and Ruby is pretty old-school (I don't mean that in a bad way, it's just not what they teach these days and not as applicable outside of web development).
Yes, it's pretty decent. Very flexible and always allows you to drop down to lower levels when you need to. knex.js is a port of Laravel's query builder from PHP to JS. But Laravel also has an active record style ORM system on top of the query builder.
Now... this was naive ASP - all the session state was kept in memory. There probably would have been a way to not do that, but no one where I worked knew how to do that (or even thought in those terms - it was just "set a sticky bit on the load balancer!".
The PHP code was slower than the ASP, but because the session state all went to disk, the memory usage never got too high. We still ended up using the sticky bit on the load balancer, but there was far less concurrent memory usage (IIRC servers had either 256m or 512m). We could serve more - slightly more slowly - via PHP.
The upside was that it was a known user-based - contractor/dealers buying from the company - so we didn't have to deal with "well, we might get 40,000 browsing visitors tomorrow because of a TV ad". The user base didn't fluctuate wildly, so it was easier to manage against.
By comparison, in the same company, another team of people were working on a Java web app. It would often take them 15-20 minutes to recompile/redeploy something to a test server to try out an idea to get feedback on something, whereas with the PHP, I could put code on a test server for feedback in a few seconds. (could usually do that with the ASP too). Yes, no doubt, the Java people were "doing something wrong" and someone can say they'd have known how to do it "right" in the year 2000, but... it was what it was.
Note: I am not saying they should use C# and Java - just that they do. I was a Python dev for 15 years before doing this, and I was very surprised to see how dominant those still are in companies over 10 years old and over $10M in revenue. And the bigger the company, the more likely it is be on C# or Java. I think every one I've seen going for north of $100M was at at least mostly on them (i.e. they might have a few services in Go or Node or something too, or some model engine in a niche language). Once a company is into the land of bankers selling it to other bankers, a lot of decisions get made for "safe" reasons, like ease of hiring good developers, enterprise infrastructure support, well audited security, and so on. And from the perspective of the investment firms, that frankly makes a lot of sense. By the numbers, most successful tech companies are not particularly exciting places to work anymore.
If the F in FAANG doesn't qualify, then I don't know what does.