Information Cascades and Bad Website Decisions

I miss blogging. It’s been too long. I’m going to bang out a post that’s been stuck in my head and just hit Publish. Hope you enjoy!

A swirling vortex of letter-like symbols converges in a lighter point on a dark blue plane
“an information cascade” as generated by Stable Diffusion

The Component Gallery is a really amazing resource showing how different large design systems approach common interface elements. There are 81 examples of checkboxes! (I am legitimately very excited about this.) But anyway, that’s not really the point. The point was that on the About page, there is the following caveat that I need to quote in full:

Don’t assume that an opinion or technique is correct, just because it’s popular. Be wary of information cascades and, if you can, always base your decisions on research. Definitely don’t just copy what everyone else is doing. There’s lots of research out there already — look for design systems that give reasons for decisions based on real research.

This is great advice and gives a name to something similar I’ve mulled over in the past that I called a “worst practice cycle” (though it’s not quite the same). This definitely happens in web and interface design all the time. I often get requests that feel as if they’re primarily informed by what they see other websites doing rather than what helps website visitors effectively use websites (which is the goal of any website for any organization).

But why is this? Nobody wants to follow the herd without a good reason. Nobody wants to make a website that’s hard to use or inaccessible. Why are there so many websites that are frustrating and bad in the same ways?

The concept of “information cascades” was new to me and is a great way to think about these questions and phenomena:

Information cascades occur when external information obtained from previous participants in an event overrides one’s own private signal, irrespective of the correctness of the former over the latter.

Information cascades explain how what other people or organizations are doing can keep you from considering—or make you actively ignore—other sources of information that would lead you to a different conclusion.

People don’t want to make bad decisions, but we are tired, have a million other things to do, and don’t always have the expertise to evaluate a bunch of technical or nuanced design decisions. I am certainly not immune from this!

I’m very confident that this is why websites continue to use home page slideshows, overwhelming scroll-triggered animations, annoying popups, and similar interfaces with bad usability. There are so many other websites with them, that it’s hard to imagine they could be that bad.

I also sometimes wonder about features like “Escape” buttons on domestic violence websites or font-size changers on websites as prominent as the White House. These features aren’t “bad”, but I do wonder if they actually accomplish the goals for visitors that the site owners have. Is there evidence that people actually use these features (I’ve never seen it).

Even when you do put in the time to do the research, it can become very hard to buck a major trend when “everyone is doing it”. I’m certainly not immune to looking around to see dozens of websites doing something I don’t agree with but still struggling to ignore that nagging doubt of “what are all these other people seeing that I’m missing?”

Of course, for every social dynamic, people will try to use it to sell things. The Wikipedia article offers multiple examples of this, including this very enlightening bit on marketing:

Marketers also use the idea of cascades to attempt to get a buying cascade started for a new product. If they can induce an initial set of people to adopt the new product, then those who make purchasing decisions later on may also adopt the product even if it is no better than, or perhaps even worse than, competing products. This is most effective if these later consumers are able to observe the adoption decisions, but not how satisfied the early customers actually were with the choice. This is consistent with the idea that cascades arise naturally when people can see what others do but not what they know.

That last bit really drives home why information cascades are such a problem for organizations building websites. It is very easy to observe what all your peers are doing. It takes much more effort to do detailed research on usability, and even more time to do the critical work of seeing how people actually use your website.

I saw this all happen, in fact. I recently watched a sales pitch for a product that I strongly disagree with and would refuse to install. This is mostly a story for a future blog post, but the key point here is that the sales person primarily relied on the following points:

  1. Look over here at these other organizations that use our tool.
  2. These organizations are very happy with using our tool.
  3. Many website visitors click on our tool.

When pushed to offer positive feedback from actual website visitors or to explain how they had done usability testing on their product, there was essentially no answer. The sales pitch was: Look at all these other people doing a thing! Just trust us that it actually helps people. “This is most effective if these later consumers are able to observe the adoption decisions, but not how satisfied the early customers actually were with the choice.”

(Aside: I recently was able to ask two organizations why they were using tools like this. They both immediately pointed to other organizations as a key part of their reasoning, and said “we did lots of research”. But research about what? I never got good answers, but I have to wonder if it was the equivalent of researching the best air conditioning system for your house at the North Pole.)

Looking at other, bigger, fancier, well-known websites feels like the perfect shortcut to figuring out how to make your own website good, but it’s rare for small organizations to have any information about how effective websites are for other organizations or how they impact the people using them. It’s like having a famous house that’s horrible to live in.

But at least there’s some good news!

4. Cascades are fragile – because agents receive no extra information after the difference between a and b increases beyond 2, and because such differences can occur at small numbers of agents, agents considering opinions from those agents who are making decisions based on actual information can be dissuaded from a choice rather easily. This suggests that cascades are susceptible to the release of public information.

Put in plain language, the only data point in an information cascade is that “other people are doing it”. That might be 2 people, 1000 people, or a million people. But if compelling information comes out that what they are doing is definitively wrong, it becomes fairly easy to buck the trend. This also makes me wonder about trends in fields like education where it takes time for it to become clear whether a new practice or tool is actually effective.

And so now I’ll tell you the thing I saw being sold recently with all the hallmarks of an information cascade. It was an accessibility toolbar and there are tons of reasons why you shouldn’t use them. Please read that article (and this one too) and think about information cascades the next time to have to make a website decision. Never let other people’s work distract you from your impact on other people. You can help break the chain!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.