Creating an inaccessible website

Recently I found a post which explores some of the limitations of automated website accessibility checking tools by building a site that scores a 100% on Google’s accessibility checker, while being completely inaccessible.

That post explores the very real limitations of automated tools. Tools like that can be useful, and help us by checking some of the low hanging fruit. But they never do a great job. For instance, just to highlight one obvious example mentioned in the post, here’s some examples of alt text on a logo which links to the homepage which would pass an automated check, but are really terribly inaccessible:

  • alt=”logo”
  • alt=”logo.png”
  • alt=”Picture”
  • alt=”The Lane Community College Logo”
  • alt=”Lane Community College”

You know what some good alt text would be? “Visit the LCC homepage” – it’s a linked logo, so provide the purpose of the link.

When it comes to automated accessibility tools, use them to give you an idea of the overall accessibility of a page, but always double-check some items yourself – particularly ones that are difficult to for a computer to check, like link purpose.

Looking for more on alt text? Check out a decision tree for alt text

Welcome back!

I know, I know, we’re in week 4, and I’m finally putting out my first post of the year. But it’s 2020. It’s that kind of year.

Last week the web team attended the 2020 HighEdWeb annual conference. This is one of my favorite conferences, and it was even better this year because it was absolutely free online (though in central time, which made for some early mornings). Here’s some gems:

Plain Language Matters.

It’s an accessibility issue. It’s an equity issue. People skim online. If we make that hard for them, they leave. Take a look at our previous plain language post.

Resource pages & emails don’t work.

One of the groups at Miami University did a lot of extensive testing. People don’t self serve. If you have a lot of links and resources that you want to put out there, consider a drip campaign to slowly provide those links at a pace people can digest. Use social to highlight different resources at different times. Include titles that highlight the problem the resource solves (“Looking for a tutor late on a Sunday night before class? Look no further!”). Track what you do to see what works (and reach out if you’d like help setting that up!). Rather than a resource page, consider a blog, where those resource links can be provided in context, and do some content marketing for you. We’ve also covered resource pages in the past.

Stop posting flyers and event posters online.

Especially now, when we’re not going to see them in person. Flyers and posters are designed for print, and don’t translate well to digital. If you’re considering putting a flyer or poster online or on a digital sign, reach out – we’ll connect you with some graphic design resources, and help design for the medium you’ll be promoting on.

The college website is for prospective students. And those students know when they’re being marketed to.

Carlton did a great session where they detailed extensive user testing before their homepage redesign. They found things like:

  • prospective students (particularly Gen Z) think the entire website is for them. Even the section clearly labeled “Alumni”. But your homepage should be for them before any other audience: most other audiences search for something, then land on some other page. Prospective students are the most likely, by far, to land on the homepage.
  • they know when pictures are staged. They want to see people in place: shots that show what students actually do some place on your campus, and how that sets you apart. Person under a tree reading a book? Clearly staged. Dining hall shot? Every college has a dining hall. Candid shot of a class outside? Student learning to machine something? Much better.
  • Carousels don’t work. I think one possible exception is a photo gallery, but that’s tricky.
  • The large hero image on a program’s website sets the tone and creates a greater impression than all the text there.
  • From one of their slides: “Students want #nofilter, but we’re giving them #fellowkids”

FAQs don’t work.

While we may think splitting our content up into questions is easier for the student, it actually makes things harder to understand. Read your entire FAQ page, make groups out of the content and write a header for each one, and then rewrite the content in each group to paragraph form. It’ll work better for everyone. Here’s a page with the slides, a sample FAQ with real life before and after examples, and some other resources for why we should stop using FAQ pages. You can also review our previous post on FAQs.

2020 Site Archive now available

While I know I’d promised to put most blog content on pause, I wanted to make a quick post about the availability of our new site archive.

Every once in a while, we create a snapshot of the entire www.lanecc.edu website. There’s a couple limitations to that. We can’t capture:

  • Files that aren’t linked from a page somewhere
  • Pages that aren’t linked from another page
  • Pages that are password protected
  • Some of the dynamic content (for instance, all the forms are disabled)
  • All of the changes that have happened since the last archive.
  • corrections to broken links (there were about a dozen internal links that are broken)

This type of snapshot, created using httrack, creates a standalone site that isn’t dependent on a content management system, meaning it’s very resistant to security problems, and is the only real option for creating a semi-permanent archive. I say semi-permanent, because while I intend to keep this archive up as long as possible, it’s possible, if not likely, that changes in browser rendering  will eventually make this archive difficult to view as intended.

This year’s archive will allow us to remove a bunch of old  content from the website, to further prepare for our new website. It also sets us up for splitting all the employee content out of the website: the current Lane website is monolithic, but under the new site, we’ll have audience specific websites.

Check out the 2020 Site Archive

Looking for an older archive? (you should! There’s some gems in there)

 

 

Helpful Plain Language Resources

Today’s helpful resource is the Plainlanguage.gov’s web language guidelines. While the whole thing is worth at least a skim, we’ve already covered a lot of the content in depth (see, for instance, the content redevelopment series). But I want to just briefly dive into the section labeled Follow Web Standards. There are four items that get their own subsection:

Avoid FAQs: We’ve talked about this before, but it’s worth revisiting. FAQ pages tend to be disorganized and hard to process. Try to eliminate them where you can.

Write effective links: We’ve also mentioned this in the context of accessibility, but it also makes a lot of sense from a usability perspective. Clear links are easier to use.

Repurpose print materials for the web: I think there’s some text here that’s worth quoting:

Don’t cut and paste the text of print documents to create web content. People are more likely to leave your webpage, potentially costing you time and money, because they will not take the time to find what they are looking for.

Print writing is different from web writing.

If you’ve created print materials, you’re going to need to rework them if you want them to be effective on the web. Make sure you’re speaking directly to the page visitor, and using conversational, but clear language. The purpose of print content is different than the purpose of web content.

Avoid PDF overhead: Here’s another quick quote:

The Nielsen Norman Group has done multiple studies on PDFs and has consistently found that users hate them and avoid reading them at all costs.

That should speak for itself. Avoid posting PDFs unless there’s no other option, or unless you need a document to print a certain way.

That’s it for this week, and also this year! Summer starts next week, so I’ll be taking a long break from blogging. Lots more to come this fall, along with lots of detail on the new website!

Website Redesign Check-in

It’s been a while since our last redesign post, but don’t think we haven’t made some progress. We’ve been averaging about one video call per week, and are getting closer and closer to development work.  Some of the things we’ve done:

Developed Batch 1 Designs

Our first batch of pages included the homepage, a career community page, and a program page. Our assumption is that these are some of the first pages most prospective students are going to look for, so we wanted to dive right into them. Our homepage is definitely going to shift direction, and be focused very narrowly on prospective student.

User Tested Batch 1

To be certain that we were on track with design and the information architecture, we did some intensive testing with some real prospective students. Users were asked to perform specific tasks, with people watching exactly what they did and seeing where they struggled.

Finalized Batch 1 Designs

We made some changes to the designs to address issues uncovered in user testing. Some were easy to address, but one has been a particular thorn.

Lane has a lot of different offerings, and people are confused by them. We have degrees, 2 year certificates, 1 year certificates, less than 1 year certificates, career pathways certificates, and non-credit credentials. There’s even more variety within the certificates. Some are financial aid eligible, some are not. Some are stackable with a degree, some are independent. Some are stackable, but you choose between multiple options. Some are technically stackable, but are marketed to a different market segment than the degree. We’ve gone several rounds with trying to balance standardization of design (to reduce confusion) with the flexibility to accommodate all our programs (to stay accurate). We’ve landed on a layout we think will work, and we hope to test it again, but it’ll be difficult to know if it’s worked for all programs until well after launch.

Reviewed Batch 2 Mockups

Our batch 2 pages included some Registration and Tuition related pages. While we’re pretty happy with the design of these pages, they’ve helped highlight a problem for us: our internal organization doesn’t always match how people think about us. For example, consider how students pay for college. We have a lot of departments that deal with money: a Financial Aid office, a scholarship and student employment office, a veterans benefits office, a bursar, and several people that work with sponsored accounts. There’s probably more. There’s really good reason for splitting them apart, and each requires a ton of very specific expertise. But if I have a question, and I’m not sure which of those areas can answer my question, who do I call?

Started Batch 3

Our Batch 3 designs relate to the application sorter and steps to enroll pages. We’ve done quite a few versions of these since our last redesign in 2013. For instance, our sorter page swapped from being person type oriented to goal oriented. Yet, despite all those changes, our sorter continues to be one of the least liked pages on the site. Our new design is going to try to leverage some of that experience, and include information that can help you navigate either way, while simultaneously emphasizing the most commonly used enrollment pathways.

Content planner

Our greatest amount of work has been the content planner. This maps content on our current website to the new website, and identifies where the gaps are. We’ve got a bunch of folders and empty documents set up in google docs right now where we’ve been starting to develop new content and rewrite some old content. There were more than a hundred pages which we need to keep, but which don’t have an obvious home in the new website, and I’ve been slowly making my way through. Some of the rewritten content will be launched before we launch, while most of the new stuff won’t be launched until the whole site is ready.

Meanwhile, as we continue our review of every page on the website, Lori’s been aggressively working on some of the recommended page merging and deletion. Thank you to the dozens of you that have helped us delete old pages!

What’s next?

After we finalize batch 3 this week, we’re hoping to do another round of prospective student testing. Very soon development will start, and while the site is being built, we’ll continue our work on content.

One of our big challenges will be photography. Normally for a website redesign you’d schedule a couple of professional photo sessions on campus, but due to COVID-19, that’s tricky. Before launch, it’s unlikely campus will look quite as busy as it would normally, we won’t see groups of people together, and the people we do see may be wearing masks. I’ve been trying to make it out to campus once in a while to get some photos, but there’s only so many empty shots of campus we can use. If you have any amazing photos – ideally where everyone in the picture has signed a photo release – and you’d be willing to let us use them, send them our way!

One of the campus turkeys in front of the center building
One of the photos from my weekend adventures shooting photos on campus

The structure of a backlink request

We’ve talked about why we get backlink requests, but what do they look like? So glad you asked. Let’s talk about my dog.

Five years ago I tried to take my dog hammock camping. She got cold overnight, and I had to put her in the car to get warm. When I wrote up the trip on my personal website, I linked to a dog sleeping bag manufacturer, since I na?vely thought my dog would sleep in one and keep warm (spoiler alert: she will not).

Picture of my adorable dog
Dog tax. Picture from the trip. She was even wearing a jacket!

A couple weeks ago, I got a backlink request from a website that wrote up a guide to buying a dog sleeping bag. This request fit the form of a standard backlink request template so well that I based a fill in the blank backlink generator request on it:

Having trouble? View this embedded page on its own.

This isn’t the exact email I got, but it’s close enough to get an idea of the structure of a request.  They try to include a little information about you, which they can look up online (like your job title). They include something vaguely complimentary (like calling your post “amazing”). Then they explain how their site would fit perfectly in with your content before wishing you well.

I think the process goes like this. Someone writes a vague template, based on a script available on some SEO website. Then they search for some term like “College scholarships” or “Student Resources”, and go through the top 50 or 100 pages, emailing the site owner with an email customized just enough that it won’t feel generic, unless you get a lot of backlink requests.

So how do you respond?

Consider not responding at all. There are ways to automate gathering contact information (using a WHOIS record or various screen scraping techniques), so it’s entirely possible that your request was sent by a script, instead of a person. Don’t feel bad about ignoring a computer.

Definitely don’t click on the link they sent you. If someone sends you an unsolicited link, never ever ever* click on it. If you’re really curious, at least Google the url first. See what you can learn about the page. Don’t just trust the text of a url – there’s ways to spoof that too. 

You can also just forward those requests to us. Believe me, we get a lot of them. Our general rule is to not respond to backlink requests for commercial websites, and to avoid including links to commercial websites in our pages. We sometimes make exceptions for  local, non-profit organizations. But since those organizations are part of our community and usually know someone at Lane, they tend not to reach out using a form email.

Remember, linking to a website from your page can provide an implicit endorsement of that page.  Make sure you trust what you’re linking to.

* ever ever ever ever ever ever.

Backlink Requests

Back in the resource pages post, I briefly mentioned backlink requests. Since if you don’t get these regularly they can be difficult to identify and understand, I thought we could dig into them. Backlink requests aren’t evil, and getting backlinks can certainly help your Search Engine Optimization (SEO). But the people making the requests don’t necessarily have the interests of our students in mind, so you should know how to spot requests and why they’re sent.

A quick refresher. Search engines mostly work by looking at how pages link to each other. Pages that get linked to a lot (that have a lot of inbound links) are considered more authoritative. Links from those pages are worth more than links from sites with fewer inbound links. If you have a website that you’re trying to get to rank higher in search, getting other sites to link back to yours will help.

One way to make that happen is to ask. You can do that many different ways. For instance, you could find a website with a broken or outdated link, and suggest to the website owner that they link to your site instead. Or there’s the testimonial strategy, where you write a testimonial and offer it for free to another website, with an expectation they’ll attribute your testimonial to your site. Offering a guest blog post is another. The most straightforward is to ask directly. And the easiest place to ask for your link  to be included? A resource page.

College websites are particularly attractive targets. While as mentioned above, Google certainly provides a lot of search engine weight to links,  there are all sorts of other factors they consider. One of those is domain name. Not every domain is easy to get. Anyone can purchase a .com, a .org, or dozens of others. But some, like .mil, .gov, and .edu are very hard to get. If you want an .edu domain, you need to be an accredited post-secondary educational institution. And you can only have one. 

Because an external agency has guaranteed the validity of the domain, Google is thought to give .edu sites a little boost over sites with more open domains (part of why you probably don’t want a .com for your department!).

When someone sends you an email, requesting you put their link on your site, ask yourself if they stand to gain financially from an increase in traffic on their website. The link they’re sending you might not sell anything directly. But that page might link to something that does or to a page with advertising – linking to a page on a site provides a boost to that page’s rank, which in turn provides a smaller boost to all the pages that page links to.

We should also remember that linking to a site provides something of an endorsement of that site. Particularly on resource pages we’re effectively promising students that we’ve validated these resources as good and trustworthy. Make sure they are.

Next post we’ll look at the structure of a backlink request and how to respond.

Top 25%

WebAIM’s analysis of the accessibility of the top million homepages came out the other day, and we made the top 25% for the first time. That puts us as #7 in the state for Community Colleges, #13 for Oregon public colleges. Well done SOU for being the highest ranked college in the state!

That’s two successive improvements to our rank. Hopefully we’ll move up quite a bit post redesign! See our entry on WebAIM’s site.

Are resource pages problematic?

As mentioned in the broken links post, there are a lot of pages on the Lane website that are what I’d call “Resource pages”. These pages don’t necessarily provide a lot of content on their own, but have a lot of links on them to help visitors find useful resources on other pages. I’ve heard some conflicting opinions on resource pages, so I wanted to dig into if these pages are problematic, or generally helpful.

Our resource pages are often full of broken or moved links

While we’ve made some progress on how many broken or moved external links there are still thousands of broken links on the site. Because resource pages tend to be full of links to external sites, they naturally collect bad links. I’ve spent a lot of time during the last week fixing pages, and I keep coming back to resource pages, just because they have so many links.

It only takes a few bad links on a page before visitors start to wonder if the other links are any good. If I’m on a page with a couple bad links, for me it’s an immediate indicator that this page hasn’t been reviewed in a while, and I probably shouldn’t trust it.

They’re only helpful sometimes

Resource pages are usually put together out of a genuine desire to be helpful. But they’re not as well used as you might think. Here’s some data for the last year on 5 different pages I’d consider to be resource pages:

Page External links on page % of external links clicked
1 17 47%
2 133 19%
3 27 19%
4 12 42%
5 25 52%

At best, only about half the links on a resource page are even clicked. And sometimes, that number is a lot worse.

There’s also a whole set of people who are never even going to see your resource page. In user experience discussions, we sometimes talk about searchers (who use the search box) and browsers (who explore the site using navigation). While obviously you can be both at different times, finding a resource page as a searcher feels like a long shot to me. More likely, you’d just perform a search on Google itself and not look for anything on our site.

To me, it feels like resource pages are for browsers. People who came to the site for something else, and then stumbled on these pages and have time and interest in what was is them.

Of course, that’s just my conjecture. So I thought I’d check the data.

Page  Views Entrance Exit % Bounce
1 774 30% 27.39% 38.2%
2 222 45% 38.74% 60.61%
3 93 5% 5.43% 0%
4 78 5% 17.95% 0%
5 410 14% 11.19% 64.29%

This surprised me. While pages 3 and 4 certainly match my thinking, pages 1, 4, and 5 don’t. 1 and 2 actually have a higher than average entrance and bounce rates – an indicator that these pages are being found via search and attracting traffic. But the higher than average bounce rate is an indicator that we’re not retaining that lands on those pages. The people who find those links aren’t staying around.

I did a quick check on Google Webmaster Tools to see what people were searching for. Unfortunately, there weren’t enough queries listed for Pages 2 or 4 to learn much, but Page 1 stood out. There were a lot of queries referencing a specific service that actually has a page of its own, but was linked from this page. That’s a potential issue: it means that searchers aren’t finding the page they’re looking for, but instead have to land on the resource page first, then figure out what link to click.

That said, at least page 1 provided context for the links. Some resource pages are literally just collections of links, without even headers, making it very difficult for people landing on the page to find what they’re looking for.

Can hurt us in SEO

To understand this one, you’ve got to know a little bit about how Google works. This will, of course, be a gross oversimplification. If you’re interested, the wikipedia page has considerable detail.

Here’s a simple directed graph:

A basic directed graph. A links to C, B links to C, and C links to D
True story: I once vehemently told a professor in college that I would never need to know anything about directed graphs like this.

Pretend each one of those circles is a web page, and each arrow represents a link from one page to another. Google uses those links to infer importance, and gives each page a score. If we start with each page having a score of 1, and we allow each page to transfer a third of its score to a linked page, we get something like this:

A and B both have scores of 1. They each give a third to C, giving that a score of 1.66. C then gives a third, or 0.55, to D, giving that a score o 1.55.
Turns out that professor was totally right, and I was totally wrong. Common experience for me.

If each page has a base score of 1, and gives a third of that rank across every link, you end up with A and B keeping their scores (no one links to them), C having a score of 1.66, and D a score of 1.55. So if there’s a search, and these 4 pages are in the result set, they’ll end up with C first, and D second. Links from C are more valuable than other links, but can still be beat out by lots of low value links.

You can likely predict how people try to game this: link farming. If you get a bunch of websites to link to your page, you’ll get ranked higher. For a while, you could even pay for links to your page. Google has since adjusted their algorithm to account for that and now, if they see your page as a link farm, they’ll bump it down in the rankings. So how do they know if you’re a link farm?

We’re not entirely sure. But we think one way has to do with the ratio of outbound to inbound links – that is, links to other sites vs links to your own. If we have a page with a few dozen outbound links, like on some of our resource pages, then suddenly that ratio isn’t so good. There’s no strict rule about what that ratio should be, and since Google doesn’t publish either its actual algorithm for ranking pages, or what your score is, we can’t really know if our resource pages make Google see us as a link farm. But it’s something to keep an eye on.

You should also be alert to backlink requests. These often show up as an email, where someone will ask you to include a link to their page on your page. It’s possible that their site is actually helpful, but more likely they’re reaching out trying to capitalize on your page’s score to improve their own.

Internal resources are just as bad

Of course, that doesn’t mean you should just start packing your page full of internal links either. There are a number of pages on the site that are entirely internal link resources pages, often including an abbreviated steps to enroll or a subset of our one master resources for students page. As previously mentioned, the intent is good, but if enough people add pages like that, we’ve just added more pages for our visitors to sort through, and set up a circumstance where pages can be out of sync (that is, content doesn’t match, and we tell students potentially conflicting things). What starts out as an attempt to fix a broken navigation on our site can end up exacerbating the problem.

There’s a place for resource pages

For instance, here’s a budget page. If I do an analysis like with the pages above, the numbers will look terrible. And this page definitely doesn’t provide any links in context. And compared to some of our pages, I can’t imagine it gets a lot of traffic. But that isn’t the point. This page gets used a lot in the budget planning cycle, as emails and presentations refer people to this site as a place to find specific resources. While it may be one of the most plain pages on the site, the headers make it quite functional. Sometimes a resource page is the right choice.

So what’s the conclusion?

As with anything else complicated, it depends. There’s a place for resource pages, but if you’re thinking about putting one together, you should think long and hard about if that’s the best solution. For instance, if you’re putting that page together because you genuinely think it’s important students learn about each of those links, you might be better off communicating a few resources every so often with a targeted population, or regularly posting resources to social media, or you might be better off working resources into each of your other pages, surrounded by related content.

If you absolutely must have a resources page, then make certain to put some context with the links, so that page visitors know why you’re bothering to put this link on a page. Be sure to revisit your links once in a while, and fix any broken or moved ones. And watch out for backlink requests!

If you have a page and you’d like to know if it’s working, reach out! We can dig into how people are finding it, what they’re clicking on, and if they go on to other pages.

 

 

Web Redesign and Broken Links

We’ve getting to the stage of the web redesign where it’s time to think about content. And our website has a lot of content. Even though our website is  smaller than it’s ever been, it still has more than 3000 pages and some 10,805 external links.

While I’m pretty good at clicking links, I can’t possibly click all of those. Instead, the website has a service which checks some links every fifteen minutes. Right now, every external links on the site gets checked every 3 days. Here’s a terrifying statistic: about 1 in 4 external links on our site needs to be updated somehow.

The vast majority of those are 301 errors. The page isn’t really broken, it’s just moved. Much of that is due to Chrome’s move to mark HTTP pages as not secure. But even though those links still work when we click on them, it’s important that we update them. Eventually, server redirections stop working, and if we don’t update the link before then, the link will eventually turn into a 404.

404 links are pages that can’t be found. Those are the next biggest category of broken links on our site. A 404 is the worst. We’ve all encountered them. You think you’ve found a helpful link, but when you click it, you’re left guessing about what was there. And they make us look bad, like we dont’t even have the decency to keep our pages updated.

Here’s another terrifying statistic: out of the 3439 nodes on our site, only 1 in 3 has no problematic links on it.

We need to do better. Lori and I are going to spending a lot of quality time over the next few weeks trying to fix some of these. I’m afraid this is another place where you can expect some nagging emails from us: if a page 404s, we may not have any idea what we should link to instead, and we may be asking you for help.

There’s one particular type of page which tends to have a lot of links with errors: resource pages. There’s a few dozen pages on the site which are little more than collections of links to other sites. While these page are often well meaning – they’re trying to help students find resources we’ve verified as being helpful – they’re sometimes misguided. We’ll dig into those pages deeper in a later post, but for now, when Lori or I find broken links on these pages, we’re just going to remove the link and move on.

Our new goal is to cut the number of links with errors in half before Fall term. Wish us luck!