The following is my opinion of what you should never do when redesigning a series of successful websites (or even a single website), and is based on recent changes to the Gawker network of sites.
Gawker Media, owner of popular sites like Gizmodo, Lifehacker, Jalopnik, recently began to introduce a new layout among their network of sites. Now if you’ve never used any of the Gawker sites before, you might not have a problem with their redesign. In fact you may even like it. However, for those of us that have been visiting their sites for years, this new layout is an atrocity. A pure abomination. According to Gizmodo’s won announcement, “It just looks a little different now. And works so much better.” I beg to differ.
Over the years Gawker has tweaked their various sites and changed layouts to keep things fresh and presumably improve the user’s experience. That’s a good thing. Essentially all of their sites are blogs, make no mistake about it, and each site has (or had) a fairly typical blog layout. A series of posts on a page, with paginated links at the bottom. Their sites get a lot of traffic and you might (mistakenly) think that over time they’ve learned what works best for them and honed their layout based on user data and feedback. If only that were the case. It appears as if Gawker has thrown both common sense and logic to the wind and simply decided to spitball new layout ideas, calling on some of the worst elements from around the web and incorporating them into their new layout. When Gawker last redesigned their network of sites about a year or two ago, they at least allowed the user to pick a layout that was somewhat similar to the previous design, rather than ramming a new layout down the user’s throat unilaterally, without any way to revert. Don’t get me wrong, I don’t hate change. I only hate it when it’s an obvious step backwards.
My biggest pet peeve with the new design is the infinite scroll. Oh how I detest the infinite scroll. The way infinite scrolls typically work is that a page loads with a specific amount of content. As you scroll down towards the bottom of the page, an AJAX request is sent to the web server to load more content, which is then tacked on to the bottom of the page. Scroll further and the process repeats. I hate this. And for the sake of fairness, it’s not just the Gawker websites that do it. Even some of the big boys like Google do this. I’m not sure who originated the idea, and while the concept is nice, the execution is always flawed. My biggest gripe about the infinite scroll is the jump that kicks in when new content is appended to the existing page. I cannot stand the jump. Now on tablet devices you might not notice this, but on desktops it’s unavoidable. As you drag the browser’s scroll bar downward, you’ll hit some unseen magic point that begins the AJAX loading of new content. Once that content is sent and appended to the page, God only knows what you’re in for. Typically the scroll bar will change size, since the page itself has now changed size, and it may even jump out from under your mouse. Next, the page itself usually jumps. Up or down, your guess is as good as mine. You might be reading a post and innocently scroll down to read more, when suddenly you’re at an entirely new spot in the page and the story you were reading is now several hundred pixels above (or below) where you are now. The idea of an infinite scroll makes sense, which is to try and seamlessly extend the page as you scroll down automatically, avoiding have to click through pagination links to get to more content. However, the jarring jumping and dynamic changing of position is unsettling to anyone, not to mention annoying. Let me give you a quick example of an analogous situation. Imagine you were reading a book on your Kindle, and instead of tapping the screen or pressing a button to go to the next page you had to scroll down, only to have the sentence you were on suddenly vanish, and you were at an entirely new paragraph that didn’t continue from where you just left off. Fail. Bottom line, always give the user the option of having pagination links over infinite scroll.
Related to the infinite scroll nightmare is printing. Yes, while we may not print as much as we used to, we still do print. And does infinite scroll help things? Of course not! As you might expect, should you try to print a page that has continually appended content to itself, you’re going to get one firggin’ huge printout with everything that was on the screen. Unless of course you like having to figure out what specific page the content you were interested in is on, and then you just have to go into print preview mode to figure out what page or pages your content now exists on. Fail. Worse yet, apparently at Gawker no one cares about printing as the layout is a complete disaster on paper. Oh, and don’t even think of trying to link to a part of the page in the hopes of returning there in the future. There are no bookmark anchors or changes to the URL that would help you return to where you left off.
The next thing I don’t understand at all about Gawker’s redesign choice is their use of ridiculously large images, seemingly placed randomly throughout the site. Bandwidth concerns aside, why would you need to randomly make the image associated with a story take up the full width of the page? Normally making an image so large would mean there’s emphasis on the story; however with Gawker it appears as if this is totally random. Most stories have normal sized images, while every so often, an image is ginormous, and takes up usable space. From a consistency standpoint this is bad. From a usability standpoint this is bad. From a semantic standpoint this is bad.
TL;DR – Don’t implement infinite scroll and randomly huge images on your site. Stick with what works.