A few weeks ago, I wrote a post outlining a few characteristics that have contributed to Amazon Echo/Alexa’s popularity in consumers’ homes. In that article, I wrote that:
…the success of Amazon Echo / Alexa in our home can be attributed to three things (in order of increasing importance):
- a first-mover advantage,
- a presence in the “hub” of our home (i.e. the kitchen), and
- the ability to effortlessly complete tasks we engaged in multiple times per day.
In a footnote in that post, I mentioned that Alexa also incorporates variable rewards as a way to engage users. For today’s post, I thought I’d explore how Alexa incorporates variable rewards and why they are effective. First, though, it’s important to understand why rewards and variability are important to a product experience and why, when combined, they become a powerful driver of habits.
The Role of Rewards
Much like teaching a dog a trick or encouraging a child do their chores around the house, rewards are an important component in a product experience because they reinforce specific user behaviors. When a user completes a specific action in a product, the product responds in a way that delights the user and, in turn, the user is eager to complete the action again.
Rewards can come in different forms. One of my favorite examples of an in-product reward was from the photo editing site, Picnik. Picnik had a “freemium” model where users could access some basic functionality for free and could choose to upgrade to a premium version of the service that enabled access to an expanded set of photo editing tools. When a user chose to upgrade, upon completion of the checkout process, the site would explode into a screen full of colorful balloons in celebration of the new customer.
This kind of pleasant surprise is a psychological reward for users who complete the desired action. The reward activates a specific part of the brain that corresponds with pleasure. As you might expect, rewarding a user for completing a desired action tends to drive an increase in that behavior. *
The Effect of Variability
So, rewards are effective, but combining a reward with an element of variability in how it is delivered hypercharges the effectiveness of the reward in motivating repeat behavior. The variability invokes a sense of intrigue and excitement by stimulating the specific area of the brain that is responsible for our feelings of desire. When a reward is delivered in a non-predictable way, our brains activate a feeling of excitement and anticipation — basically stressors that want to be released.
The classic example of variable rewards — including how powerful they can be at driving repeat behavior — is a slot machine. With each pull of the arm on a slot machine, the player experiences a sense of excitement about the unknown result to come. While the slot machine wheels are spinning, the reward (or lack thereof) has not been revealed yet but the pent-up anticipation of a possible future reward amplifies the experience for the player regardless of the outcome of the individual spin. Subconsciously, the player is thinking (feeling, really):
“What will happen? Will I win something? If so, will it be a small win or a large one? Ooh, I can’t wait. I can’t wait, I can’t wait!!!! Shoot, I lost. That was fun, though. Let’s do that again and see what happens…and then maybe again, and again, and again…”
The upshot is that when rewards and variability are combined in a product experience, they become an incredibly powerful driver of habitual engagement. We see the presence of variable rewards in games and in mobile apps, but they are also present in a unique way in digital assistants like Alexa, Siri, Cortana and the Google Assistant.
Alexa’s version of variable rewards
Broadly speaking, Alexa and other digital assistants deliver rewards when they do what you expect them to do, or more importantly, when they do more than you expect them to do.
Here’s why. Even though Siri has been available since 2010, the ability to speak to a computer and having it respond is a new experience to most people.
A digital assistant with a touch of personality is an even more foreign concept. Also, our collective experience with new technology is that, most of the time, it is imperfect.
Given these factors, most people who first encounter Alexa or another digital assistant, go into the experience not actually expecting it to work 100% of the time. So, when it does work, and Alexa actually does what you ask it to do, it’s a bit like receiving a small reward. When it goes above and beyond what you expect it to do, it’s a bit like receiving a big reward.
One way that Alexa delivers a big reward is by demonstrating a personality in its interactions. Alexa does this by subtly hiding a handful of amusingly clever responses to specific queries. For example, in a deferential nod to Star Wars fans, if you say to Alexa, “Alexa, I am your father”, Alexa responds by saying, “No. That’s not true. That’s impossible.”
Siri did a good job of this when it was first introduced. The developers of Siri incorporated a number of clever responses that it would give to specific questions. Entire blogs were dedicated to finding and documenting the experiences. Nir Eyal, in his book titled Hooked: How to Build Habit-Forming Products, describes these experiences as “rewards of the hunt” because searching them out gives us a feeling of accomplishment which is a type of reward.
Because digital assistants aren’t perfect and often do not do what you expect them to do there exists an element of variability in whether the “reward” will be delivered.
There are literally dozens of variables at play with any given Alexa request (e.g. a misunderstood voice command, unexpected phrasing of a request, background noise, spotty WiFi connection, etc.), meaning that there are many more ways Alexa can screw up a request than there are ways it can successfully execute it.
The cumulative effect of the failures and the successes is analogous to playing a slot machine. Just like with every pull of the arm on a slot machine, there’s a chance that the machine might pay off. With Alexa, every voice command you issue comes with the chance of Alexa being unable to fulfill your request or Alexa surprising you with a better than expected response.
Of course, Alexa’s developers continuously improve the service, and users learn how to interact with Alexa so that the interactions are more likely to be fruitful than not — which is important for a product that ultimately needs to serve as a utility in peoples’ lives. But the success rate probably won’t get close to 100% for some time and that’s ultimately a good thing for Alexa in these early days of trying to grow consumer adoption.
The reason that imperfection, in this case, is good is that, despite being occasionally frustrated with some of Alexa’s failures, it’s the uncertainty that subconsciously motivates us to come back and give Alexa another chance. It’s one more reason why I believe Alexa and other digital assistants will continue to enjoy growing success and eventually grow to become a very common part of our everyday lives.
* I recall a talk by Jonathan Sposato, the founder / CEO of Picnik, where he explained that the balloons were so well received by the site’s users, that the Picnik team decided to add a link at the bottom of the Picnik web page that simply read “Balloons!” that allowed visitors to invoke the balloon explosion on demand. A link like that belongs in the footer of every site IMO.