Default Effect and Usability
July 19, 2014
I had a conversation today about how the definition of college has changed in my 5 years at RIT. I will not expand on this topic, but topics that seem very sensical today may have been quite nonsensical half a decade ago. In the ’90s, I recall seeing engineering as being about the designer of the new big “invention,” and I always envisioned this as an individual invents the new big idea or gadget. Today, I see engineering as the new big idea, but instead of invention, I think as ideas owned by start-ups. The definition of engineering in my head has changed from individual design to team design (i.e., the app age). I do question if this is a result of a changing world view (did college change me or did the world change around me?)
A recent discussion has been on how Google took years to address a battery draining issue in its browser. In the Ars Technica article, they reference a post by a previous Google engineer, Mike Belshe, questioning the default “internal timer on wait-event functions” on a Windows machine being 15 ms, which in today’s standards is very long indeed considering the processors today versus the processors of yesteryear. This conversation really calls to point on whether there are other default settings out there in the science, engineering, or humanities worlds that are outdated or underanalyzed because they are accepted as the ‘norm.’
Applebee’s in Saugus, Massachusetts (Wikimedia)
As of SEC filings in Feb 2013, Applebee’s had 2,011 franchised restaurants in 49 states within the U.S., 1 U.S. territory, and 15 countries outside of the United States.
As of SEC filings in Feb 2019, Applebee’s had 1,768 franchised restaurants in all 50 U.S. states, 2 U.S. territories, and 13 countries outside of the United States.
As an engineer, I’ve been taught to question the way the world works and ask whether it is the best that it can be. It has been shown that setting the default has a significant effect on the decisions that users make (Default effect). A tough topic in the realm of engineering (like any field) is ethics and this is where the default can be important. For instance, Applebee’s and Chili’s new tabletop tablets have a default of 20% tip (despite the average being somewhere above 15%, but well below 20% – tipping in the U.S.). A similar tipping trend has been shown in adding tablets in NYC taxis (where the defaults are 20%, 25%, and 30%). These tabletop tablets have become a trend in the industry, as Red Robin and Uno Chicago Grill will also have the devices. Companies, not consumers, design technology. On the other hand, sometimes you wonder if the tip should really be quantified as a portion of the check, or based on the time you are at the restaurant.
Jeeves (Flickr) was the mascot for AskJeeves when I was a kid. He has since retired in 2006, but he still exists in the U.K., probably unaware of what they’ve changed his company into.
You may argue at this point in the name of waiters/waitresses and how the company is looking out for them. The default that I would argue the majority of the population is against is the default to install the Ask browser toolbar when installing or updating Java. One company found that 40% of people who had the Ask toolbar had it installed through the Java install. The article also examines that the Ask toolbar will install itself 10 minutes after the Java install to prevent the user from taking immediate action (PCWorld). (There’s a petition out to Oracle about this.)
When designing software, we need to keep usability in mind. But as the understanding of human-computer interaction grows, we need to engineer ethics into our creations. We need to keep defaults in mind too, as they direct human behavior. Do they really have to set the default value for the newsletter subscription of every website sign-up to “Yes”? Shouldn’t we provide the user the ability to understand why the default value is as it is, or place no default value in the first place?
A set of experiments in this paper at Aalborg University found that we are more prone to accept the defaults when the interface is familiar (consider any ‘typical’ software installation or website sign-up). An argument made here is that users in unfamiliar situations may employ prospect theory, a decision-making process where you choose based on the information of gains and losses given (e.g., this seems to be the typical tip for a taxi, so let me choose the higher one). The designers place bias into a question by suggesting a single default or multiple defaults (as the NYC taxi example shows). Imagine if a designer were to place a default donation of 20% of your next online order to a political candidate. Where do we draw the line on designers suggesting defaults? And when do we start educating our customers on what other consumers choose and why you would choose one option over the other?
“Ethics is knowing the difference between what you have a right to do and what is right to do.” – Potter Stewart
- “Recommendations Implicit in Policy Defaults” – McKenzie, Liersch, & Finkelstein (2006)
- “Watch out! Default values ahead…”
- “A lesson in usability anti-patterns from Virgin Blue”
UPDATE 1 (1/12/2015)
Recently in the news, there was some discussion on taxi fares. I Quant NY did some great analysis on NYC taxi fare data. The post comments on the fact that an official misinterpreted riders’ behavior for the effects of software that enabled the default effect. These taxi services have unclear and inconsistent methods for calculating tips.
Further update (7/9/2017)
There’s a small clip in a 2016 Data Stories podcast referencing this story.
UPDATE 2 (2/8/2015)
A Forbes article talks about default tipping, citing a concept called “What You See Is All There Is (WYSIATI)”. The author discusses the new reality of using payment apps, and how their spread could mean more and more examples of this.
An edit was made to this page on April 6, 2019 to comply with new Jekyll theme.