Why Some Apps Are Intentionally Slow

February 27, 2020

In tech, speed is usually a major priority. But did you know it’s actually possible to create a program that is too fast? So some websites and apps are slowing down. From TurboTax to Facebook’s Privacy checkup, this phenomenon occurs more than you might think.

FULL TRANSCRIPT

Andrew Davis: [MUSIC]

Andrew Davis: If you ever use TurboTax before,

Andrew Davis: you probably remember a part of

Andrew Davis: the process that look something like this.

Andrew Davis: Where the program would have you wait for a few seconds,

Andrew Davis: while claimed to double check

Andrew Davis: your tax returns before you filed.

Andrew Davis: What you probably don't know though,

Andrew Davis: is back in 2017 The Atlantic

Andrew Davis: took a closer look at TurboTax source code,

Andrew Davis: and what they found was actually pretty interesting.

Andrew Davis: That animation was fixed.

Andrew Davis: It did not appear to be communicating with

Andrew Davis: TurboTax as servers once it began playing,

Andrew Davis: and it always took the same amount of time to

Andrew Davis: complete regardless of the user.

Andrew Davis: Basically, it appeared that

Andrew Davis: TurboTax's developers had inserted

Andrew Davis: this delay for the user experience.

Andrew Davis: It's called an artificial wait

Andrew Davis: and it's not just TurboTax.

Andrew Davis: Remember when Facebook launched

Andrew Davis: its privacy checkup a few years back,

Andrew Davis: that friendly little cartoon robot

Andrew Davis: with a stethoscope would have you wait for

Andrew Davis: a few seconds while it scanned through your profile and

Andrew Davis: offered you personalized feedback

Andrew Davis: on your privacy settings.

Andrew Davis: Yeah, that was a fake way too.

Andrew Davis: Fast Company revealed that

Andrew Davis: Facebook's massive server infrastructure

Andrew Davis: can perform your privacy checkup almost instantly.

Andrew Davis: So what could prompt tech companies to do this?

Andrew Davis: Well, it turns out that you're probably

Andrew Davis: thinking about waiting or wrong.

Andrew Davis: [MUSIC]

Andrew Davis: Most of us assume waiting is bad,

Andrew Davis: but the reality is much more complicated.

Andrew Davis: In many cases, people expect to wait

Andrew Davis: and if that expectation is not met,

Andrew Davis: speed can actually undermine

Andrew Davis: their trust in the final product.

Braden Kowitz: The goal for almost every company is

Braden Kowitz: deliver experiences that are great for people,

Braden Kowitz: and most of the time that

Braden Kowitz: means something that's really fast.

Braden Kowitz: But your experience also has

Braden Kowitz: to meet people's expectations.

Braden Kowitz: So if I go into a nice restaurant and order

Braden Kowitz: a steak and it comes in 30 seconds,

Braden Kowitz: that breaks my expectation,

Braden Kowitz: and I might start to wonder

Braden Kowitz: what's happening in the kitchen.

Andrew Davis: That same principle can apply in tech.

Braden Kowitz: Computers can run incredibly fast,

Braden Kowitz: and in some cases you actually have to

Braden Kowitz: slow the service down.

Andrew Davis: Steven Hoober has worked on

Andrew Davis: the user design experience for media,

Andrew Davis: technology and telecommunications companies.

Andrew Davis: He says, he stumbled across artificial wait times by

Andrew Davis: accident when working for a major cell phone provider.

Steven Hoober: We were told to build a plan optimizer.

Steven Hoober: The user would tell us a few bits of information and we

Steven Hoober: will then find the best service plan

Steven Hoober: for them based on that stuff.

Steven Hoober: Mathematically at least, it's not that complicated,

Steven Hoober: it ran instantaneously and

Steven Hoober: gave them very accurate results.

Andrew Davis: But when Steven's team started testing

Andrew Davis: this plan optimizer they were so proud of with users,

Andrew Davis: they got a big surprise.

Steven Hoober: Everybody hated it and

Steven Hoober: we couldn't figure out why for a while.

Steven Hoober: We finally realized, if you are gonna say,

Steven Hoober: let me check that for you.

Steven Hoober: Then you press the button and it immediately acts,

Steven Hoober: turns out people don't trust it.

Andrew Davis: Steven's team sat there scratching

Andrew Davis: their heads for a few days until he said.

Steven Hoober: So wait, what if we add a delay? That's all we did.

Steven Hoober: We added a random delay and then we started

Steven Hoober: seeing people actually interested.

Steven Hoober: They became engaged in literally most of

Steven Hoober: the users leaned into the screen like,

Steven Hoober: uh what's it doing?

Steven Hoober: And then we're terribly excited to see the results.

Andrew Davis: That excitement in Steven's testing alludes to

Andrew Davis: another reason developers might

Andrew Davis: use these artificial wait times.

Andrew Davis: A little wait can not only help prevent

Andrew Davis: a user from disliking or distrusting an app,

Andrew Davis: but can also make the user value that app more.

Andrew Davis: Waiting can create suspense,

Andrew Davis: hyping up the anticipation of results.

Andrew Davis: It can also draw focus to all the work

Andrew Davis: the app is doing on the user's behalf.

Andrew Davis: Increasing the perceived value of those results.

Andrew Davis: This phenomenon is called the Labor Illusion.

Andrew Davis: Ryan Buell and Michael Norton from

Andrew Davis: Harvard Business School coined the jam.

Andrew Davis: It all started with a visit to

Andrew Davis: the early version of kayak.com.

Andrew Davis: Buell and Norton noticed that

Andrew Davis: the site had a fun little animation that showed

Andrew Davis: users in real-time all the airlines

Andrew Davis: the algorithm was searching.

Andrew Davis: They wondered how this impacted the user experience.

Andrew Davis: So they ran some experiments with

Andrew Davis: the homemade generic version of

Andrew Davis: the Kayak site that tested different search scenarios.

Andrew Davis: What they found is users can actually prefer waiting

Andrew Davis: for results when that

Andrew Davis: waiting is accompanied by the animation,

Andrew Davis: telling them what the algorithm is doing,

Andrew Davis: even over getting the results instantly.

Andrew Davis: But they also soon realized that this information didn't

Andrew Davis: even have to be real to make that enhanced value stick.

Andrew Davis: In short, just creating

Andrew Davis: the illusion of the site labeling by listing off

Andrew Davis: airline names during a wait can

Andrew Davis: increase the perceived value of the results.

Andrew Davis: So now we can see why developers might have

Andrew Davis: an incentive to insert fake wait times into apps.

Andrew Davis: But the presence of little illusions like

Andrew Davis: these got some experts thinking,

Andrew Davis: is all this ethical?

Andrew Davis: [MUSIC]

Andrew Davis: Professor Eytan Adar from

Andrew Davis: the University of Michigan and a few of

Andrew Davis: his colleagues from Microsoft pioneered

Andrew Davis: research on this question back in 2013.

Eytan Adar: The way we think about it from

Eytan Adar: the ethics perspective is who benefits?

Eytan Adar: So am I actually

Eytan Adar: doing something in the interests of the user,

Eytan Adar: are they better off, um

Eytan Adar: given the way I've implemented these things.

Andrew Davis: For Eytan, it all really boils down to whether it's

Andrew Davis: a harmless benevolent deception

Andrew Davis: or a harmful malevolent one.

Andrew Davis: So what about our TurboTax example?

Andrew Davis: Was that benefiting the end user?

Andrew Davis: That little artificial wait with

Andrew Davis: its claims of double and triple checking

Andrew Davis: the results could obviously increase

Andrew Davis: the perceived value of the experience,

Andrew Davis: which in turn could help sell more TurboTax.

Andrew Davis: But building confidence in a product

Andrew Davis: here might help users as well.

Eytan Adar: With things like money, right?

Eytan Adar: If it came back within a second and say, "No,

Eytan Adar: I'm sorry I can't find anything," you will feel worried,

Eytan Adar: you will start looking through

Eytan Adar: the documents much more closely, uh,

Eytan Adar: you will potentially go out and like make tweaks on your

Eytan Adar: own which might result in you filing the wrong taxes.

Eytan Adar: So there might be some cost in you

Eytan Adar: feeling like the software is

Eytan Adar: not doing what it's supposed to do.

Andrew Davis: So these waits can be helpful and a little bit worrisome.

Andrew Davis: Just knowing about them though,

Andrew Davis: can change the way that we think about technology.

Andrew Davis: I think a lot of us had this idea in

Andrew Davis: our heads that computers are sort of infallible,

Andrew Davis: that they're sort of in voluntarily truthful.

Andrew Davis: But even this video's quick look at artificial waits can

Andrew Davis: show us that the reality is much more complicated.

Andrew Davis: Instead, developers can bake

Andrew Davis: their own human qualities into these systems.

Andrew Davis: Whether that's an inspired design experience

Andrew Davis: or a fib here or there.

Andrew Davis: Knowing this can nip a lot of

Andrew Davis: misconceptions about technology in the bud.

Andrew Davis: Ensuring that those little benevolent deceptions

Andrew Davis: remain just that benevolent,

Andrew Davis: rather than taking on a life of their own.

Andrew Davis: [MUSIC]