Functional Threshold Power (FTP) is a tried and tested method of giving you an estimate of your 1 hour threshold power, from a 20 minute test. Rather than going into the details, here's a short video from GCN to help explain...
However I'm here to try to explain why you shouldn't place too much value on FTP, both yours and especially other people’s.
Your FTP is your FTP, its primary focus is to make sure the numbers you are training to are accurate and you are getting the best out of your sessions. That's it. It has some secondary uses, like providing a benchmark of your current fitness, although this only works if you use exactly the same equipment for every test. It can also be used for your pacing strategies, although the longer the race the less reliable it can be. It is not to be used for top trumps, willy waving or seeing how you stack up against anyone else. Let's look into why...
It's an estimate, the standard test is 20 minutes long and is being used to estimate paces for longer efforts - and that's the first potential problem. I can tell you from experience, having done both 20 minute and 1 hour versions of the test, there can be huge variations between the numbers and I've seen differences of up to 10%.
The equipment you use also has a massive impact. Different kit gives different readings; in my own data I've seen about 10% across different kit, but I've read articles that quote higher variations. If you're taking your weight to calculate power to weight (the key to Zwift's algorithms) you can also have huge variations here, e.g. weight fluctuations throughout the day (I've had almost 5% at different times of day) but also between devices. This is one of the key reasons why your numbers are your numbers - do your best to keep how you measure them consistent but they can't be reliably compared to other people's unless you're using the same kit.
Was the test honest and relevant? What I mean by that is that we're triathletes, so we're interested in time trialling. I know that I can hold a much higher number 'climbing' than in the TT position so if I do an FTP test sat up rather than in the TT position I reckon my number is about 10% higher. But if I tried to use my 'sat up FTP' number to pace my time trial, I'm going to get a shock.
There's plenty more reasons for variations, not only in your own data but especially if you try and compare it to others. So don't, it's not helpful. Even looking at the examples above you could have 10% in power measurement and test protocol and 5% in weight. For a 240w FTP in an 80kg rider (3w/kg) on other devices and tests that could be as low as 192w and 84kg (2.3w/kg) or as high as 298w and 76kg (3.9w/kg). I'm over simplifying and being slightly creative but you get the idea.
As I said your FTP can be used as a benchmark of your fitness, and for pacing but be cautious here too...
It's generally a pretty good benchmark of your fitness, but if your FTP goes down even though you've been training like a beast, don't be dejected. An FTP test is hard, both physically and mentally, and they're pretty easy to f*** up royally - fatigue, sleep, pacing and nutrition all have an impact. You're trying to extract everything out of yourself and when you're on the limit it only takes a few 1 percenters to be off and your maximum potential will quickly start to fall. It's not the end of the world, use that number for your training, if it gets too easy, time to re-test.
For pacing and performance there are also considerable variables. Just because you can hold a decent number for 20 minutes, it doesn't actually mean you can hold 95% of a decent number for an hour. It's a fair assumption, but it might not be the case (as above). That level of accuracy only decreases with the length of event because of a multitude of other variables. Again from experience, my best race performances have not coincided with my highest FTP numbers, not by a long shot. High FTP does not automatically mean fast, especially when you've got an indoor FTP for an outside event - aerodynamics, nutrition, equipment, weight, pacing, tactics... the list goes on, and these dictate how quick you're going to go. I think I'm right in saying the fastest man on a bicycle in our club doesn't know or regularly test his FTP.
In summary, aside from setting the right power for your sessions in the environment you set it, your FTP isn't as important as you might think - and it should be retested every 6 weeks to even be useful in that context. It doesn't matter if it goes down and there's always someone with a bigger one so don't bother looking, you don't know what sort of tape measure they have.