Amit BenDor

@amitbd

Voice Is Booming! Does Your App Deserve A Voice Version?

December 16th 2018

5 tiny tests to decide if you need a voice app

Voice seems to be one of the most notable opportunities in the market. In the U.S. alone 24% of households own at least one smart speaker (Google/Amazon), which means 80 million potential users.

This number keeps growing sharply (you can read more here and here)

You’re probably looking at these stats and thinking, ”hmm... I should build a voice app”. Well, the truth is that voice can be an amazing fit for your application, but also a terrible one, depending on the use case or scenario.

But How?

So how can you evaluate it? I came up with a fun, easy way to do that whether you’re an Indie developer or representing an enterprise — A game!

Or even more specific — a personality quiz for your app.

Game Rules:

Let’s take your use case and put it into several tests, in some tests it can earn points but it can also lose points for others.

  • If passed the test, add (+) the number to your total sum
  • If failed, subtract (-)
  • You’ll find how much to add/subtract in the headline of each test
  • In some tests, you can only earn points — those who don’t have a “-” sign in the headline

So let’s grab a piece of paper (or open the notepad) and get started!

The Game:

#1 The friction test (+/- _ points)

This is the fundamental test for your use case, and it’s just about counting actions — it is that simple.

count the number of actions the user will have to do in a traditional UI. (Web/App)
Subtract (-) it with the number of actions it will have to do in a voice UI

Voice-based interfaces have one big possible advantage — the user can trigger a complex action with just *one* sentence.

For example, using a bank voice assistant:

“What was my balance from December 15' not including any foreign currency?”

— “Your balance on December 15' was 1,000,001$ ”

In a traditional UI (web) you would probably have to:

  1. Google your bank name and enter their website
  2. Log in to your account
  3. Choose the balance menu
  4. Page to December 2015
  5. Filter foreign currencies

Tiring. isn’t it?

Yeah. I know how you feel

So in a voice UI we counted 1 action, in the traditional UI total of 5 Actions!

In that case, it will be 5–1 = 4 extra actions (!)

  • As I said, this is a crucial test so every extra action counts as 2 points!
  • As it is the most fundamental test in the process, we’d count each action twice
  • In our case, we’ll add 4*2 = 8 points.

To be more precise, measure the “happy path” of your app as well as the “error path” and compare both of them to a traditional UI.

Also, you give a different weight to longer actions such as searching, browsing and sorting, compared to shorter action.

#2 The “busy hands” test (+3 points)

If in a typical scenario your user’s hands are occupied or resting, this could be a great advantage for your voice app.

A few examples for that are:

  • Driving, Walking, Running
  • A mom carrying her child
  • Preparing food in the kitchen
  • Lying in bed

One of the most common voice commands I use is controlling Netflix on my Chromecast.

To be honest, I’m just too lazy to pick up the phone, open the app and do whatever I want to do. I prefer asking my Google assistant to “Play Narcos”. (don’t judge me 😅)

Typical lazy Netflix user

#3 The “narrow scenario” test (+/- 4 points)

This is another core principle in every conversational interface — voice or chat.

Are users` expectations aligned with your voice app abilities?

In a voice app, you do not have a menu or organized categories in front of your eyes, and so, as the user approaches the voice chat, she needs to know what it can generally do.

Otherwise, she might ask for things you did not expect and do not want to support. This might lead to frustration and miss-use of your voice app.

So you need to ask yourself “Is the scenario narrow enough? (or open-ended?)”. Keep in mind that it is not about your expectations, but the user’s expectations when she first approaches it.

Examples:

  • Narrow — A voice bot for For Filing A Car Insurance Claim
  • Open-ended — A voice bot for gaming
  • Seems to be narrow but open-ended — “A concierge voice bot”

In the last example, you can see how it is very important to see how the user perceives it in the exact scenario.

In my recently closed startup Gooster, our product was “A concierge chatbot” (for recommendations, tickets etc.), but for the user, the bot represented all of the hotel’s services including booking or any other strange question they did have. (“can you clean my socks?”)

#4 The too much information test (+/- 3 points)

Can you answer the user’s questions with short, concise answers?

“Ok Google, Why is my wifi so slow?”

— “Interference can also be caused by simply having another electronic device too close to the wireless router itself. If the wireless router is sitting on or very close to a TV or sub-woofer speaker, or other electronic devices, it can cause interference… ” 😕

(This is a real-life example for a long answer by my Google assistant)

Imagine that you’re speaking with a friend who just can’t stop talking. Something distracts you for a moment, and suddenly you have no idea what he was talking about.

Using a voice app, just like a normal conversation, requires almost our full attention as a human. When distracted, one loses bits of information so he might ask you to repeat it or worse, get lost in the conversation flow.

Furthermore, if the information given by your app is not what the user expected, she will probably have to wait until it finishes up talking.

So please, don’t be the friend who just can’t stop talking. Keep your answers short and precise!

#5 The Emotion test (+2 points)

Conversing with a voice agent can sometimes feel like talking to a human.
A well built voice bot can deliver more than just data but also ignite an emotion. 💓

How nice would it be if your insurance voice app will have the personality of your favorite movie star? or to hear Pablo Escobar on a drug selling voice app (just kidding 😆)

Emotion is a powerful tool in a conversation, and the most common way to achieve that is by giving your voice app a personality.

However, an iconic personality is not always a good fit for every industry or business, for instance, if an important feature of your app is to project a formal and trustworthy front.

Calculating The Results

Add up the numbers and see in which range your use case is in:

Above 15

Right on! You’re in a great spot, you should go ahead and create a voice version of your app tomorrow.

15–8

Not bad, there is potential for a voice version for your app, it can benefit at least some of your users, but it might not be life changing for your business.

Below 8

Hmmm… a voice version might not be the right choice for your app. work on the use case and retake the test. sometimes minor fixes can make a great change!

Conclusion

Creating a voice version for your app is not easy, I hope you’ve learned from this article about important principals of good voice apps and your decision is now easier.

➤ If you enjoyed reading this, give it a few claps 👏👏👏

➤ Did I miss something? write a comment!

➤ Special thanks to Mor Bendor, Gilad Dayagi, Noam Izhaki & Chen Feldman who helped me in the process of writing this article

Amit Bendor is a Voice & Chatbot Expert, helping companies to create outstanding voice experiences that scale. Co-hosting the successful podcast “Making Software” & former CTO at Gooster.

Visit amitbend.com to learn more

More Related Stories