There may be many reason to prefer a fixed layout instead of a responsive or a fluid one. This answer (in my opinion) can not be exhaustive unless it's applied to a specific case: I do not think there is a best solution that should always be used.
You're asking about fluid layout (built using percentage of width) but there are also: responsive layout (built using a fluid grid and media queries), adaptive layout (fixed but with media queries) and elastic layout (fluid until a breakpoint).
Of course you may want to improve user experience at higher resolutions (without jumping into responsive layout, which is also a compromise). That's good but with a fluid layout you have too little control over page appearance and final result may be a too big compromise (because it has to look good both for low-res and high-res displays). In this case you should strongly consider elastic layout (which is BTW what Amazon does, in this moment).
MORE AVAILABLE SPACE MEANS YOU NEED TI REDESIGN YOUR USER EXPERIENCE
Sorry to be so loud but it's important. More, it is fundamental. Stretch few images and randomly zoom your UI won't make your site better. You do not need same UX just doubling products you can see on each page. If you have more space then you should use it better, dropping compromises and decisions you took for smaller screens.
Extended discussion...
E-commerce platforms rely on rich imagery, and fluid layouts allow usage of bigger images.
Of course to have bigger images is pretty but bigger images are also heavier to download and to store. I don't even mention all the code to serve images of required size (and cache them, also don't forget you can't completely rely on browsers because window size may vary).
In this moment Amazon is serving me images at 126×135. It's OK with a fixed width on my 1900×1200 but to increase image size on 4K monitors its resolution has to double (at least). Bigger images are undoubtedly a benefit but they use bandwidth and storage, this may not be a problem for our portfolio site but for Amazon (and the others) it may be an issue to consider. I do not have up-to-date data but:
- In my homepage it presents ~50 products.
- Each image is in average 4 KB and if no image is cached it means I downloaded 200 KB of data (with HTTP 1.1 with more than 50 server connections).
- Amazon has more than 20 million users per day, it means that - just for home page - and for worst case we're using ~4 GB (server-side bandwidth! do not consider storage). Now let's double this and apply to each product image in whole site. You may need to change your hardware only because of this.
Do you want to use 256×256 previews? Now let's consider client-side download time (loudly again):
First of all web-site must be FAST. Give me 128×128 previews but do it as fast as possible. Doubling image size you will also almost double loading time.
From comments: no, you should not scale low-res images at client side. It doesn't matter how easy it is. It will look ugly and unprofessional.
E-commerce websites have a ton of products to showcase, and full width will give them more room.
If you use fluid layout then more room means more white space around each product box. Some may help but too much is useless and counterproductive. If you're talking about a responsive layout then you're right: more space is better but also in this case you would set a limit. Even if your screen is 3840×2160 I doubt you want to fill it with 128×128 images (a decent size for a preview), it means 400~500 products on the same page, for sure too many unless user is scrolling an image gallery.
E-commerce websites seldom have long lines of text, and have space partitioned into sections.
It depends which product you sell! Product description on Amazon (for example) is often very long and a fluid layout (alone) is not best option to handle them. Note that you can handle long lines with a well-done responsive layout.
960px is oh-so-old...
Don't assume this. Statistics are misleading because you should narrow population to your target audience but, generally speaking, most common resolution is now (end of 2015) 1366×768. 1920×1080 is just slightly more common than 1024×768.
...and most of the users have a 16:9 monitors now.
Screen resolution and display aspect ratio are related but separate things. For example a brand new product (Apple iPadPro, released one month ago) has a resolution 2732×2048 and its display aspect ratio is 4:3.
Fluid designs provide a clean and immersive experience.
Also fixed width layouts, it depends how you make it and the type of content you have to present. An adaptive layout (fixed but with media queries) can be often a nice and easy solution.
Grid systems have become very strong and easy to use now, so fluid designs are no longer a pain to create.
I absolutely agree but I'd stress on one word: create.
Some of these websites came into existence when 960 was the norm, and they have grown too big and...
It may be true but we can just speculate.
They are skeptical that it may drive orthodox users away? (but what about the next-gen user?)
I don't think so but it's just my own opinion. Anyway consider that you can't simply make your orthodox users unhappy waiting for the next generation unless you're planning to sell nothing for one or two years. Progressive changes are often the key, unless you don't have users but fans (I do not mention anyone).
They feel that major change in layout will cause loyal users to re-learn navigation from scratch...
It may be true but again progressive changes will help to smoothly upgrade. I saw some governative web-sites with both layouts (fixed and fluid, at that time) as options. It's a pretty common approach: think Microsoft with Windows XP (and it's gummy UI) o Facebook and transition to diary (just to mention two).
They want to play safe and stick with the tried and tested.
I absolutely and strongly agree. To test fluid layout can be 5X times more expansive and long because you have to test at least most common resolutions and aspect ratios plus edge cases (remember that fluid is with percentage). Responsive layout can be even worse (for testing) because layout will change (not just widths).
I think a responsive (not a fluid) layout is pretty often a good COMPROMISE solution if:
- If your target different devices;
- If your content requires it.
If you answers yes to both ifs then remember you have to count much (much much much) more testing and a more careful design. It means more money both for first release and for future updates. Bigger sites will carefully balance true needing and costs (think about Stack Exchange, a partially responsive layout may help?). Also note that to make a good responsive layout from XGA to 4K (or even 8K) is all but not easy.
