While undertaking new concept development within the Anthemis Foundry, I couldn’t resist tapping into the possibilities of voice interfaces across industries.

I have a passion for the possibilities that voice technology unlocks and see an opportunity for scalable businesses in this space. Albeit over the long-term, the bottom-line is that I am excited for what may come.What better way to understand a space, than to immerse yourself in it. A colleague and I attended the Mobile UX London (MUXL) designing voice interactions with Alexa

It was a great experience to put research into practice, tapping into the expertise and knowledge-base of judges and mentors to explore the full potential of the Alexa device as well as what is projected to come.

Gratifyingly, our team won the event.

VOICE IS THE FUTURE

By 2025, the digital voice assistant (DVA) software market is expected to reach 1B+ users to yield US$7.7B globally.¹ In the future, nearly twice as many millennials (29.6M) will interact with voice assistants on a monthly basis in-comparison to their Generation X counterparts (15.3M).²

This gap is projected to widen over the next three years.

Driven by a a mobile-based B2C market, a Mindshare report estimates that currently 600M global users engage with DVAs at least once weekly.³ However, second-week retention sits at an alarmingly low 3%, which proves cause for concern. There are outliers, yielding 20% retention in some instances, but this is rare.⁴

Problems do still exist with the rudimentary technology — It lacks emotional quotient. Whether this contributes to the above, needs clarification. However the first-date effect of failing to understand accents, dialects, context and complex sentences does not aid the cause.

Irrespective, considering the market is in its infancy, this should be expected.

Nevertheless, market activity is rife. Tech behemoths Amazon (Alexa) and Google (Google Home) lead the way while Samsung (Bixby), Salesforce (Einstein Voice), Microsoft (Cortana), and Apple’s (Siri) trail (despite Apple’s first-mover advantage).

In fact, the recent Alexa-Cortana integration with Microsoft proves Amazon’s market clout, and further boosts it, with ~145MN active Cortana users.⁵ Considering Microsoft’s strong product suite, there is potential to leverage and build PowerApps in the future. But let’s wait and see about that.

MNO-led initiatives from Orange (Djingo) and Telefonica (Aura) enhance DVA offerings overall. However, with 70%+ of the DVA market, it is Amazon that aims to create a third major computing platform. Yes, you heard correctly, step-aside PC and smartphones because Amazon is coming fast…REAL FAST.

THE AMAZON ECHOSYSTEM

The key cornerstone to enriching application environments in the space really comes down to creating sustainable ecosystems. Robust app ecosystems need:

• Distribution
• Retention
• Monetisation

This, combined with an emphasis on Machine Learning activities will reinforce contextual understanding capacity for intelligent assistants. Hence, Facebook, Baidu, Microsoft and Amazon have (and will continue to) invest in ML and AI engines. Recent evidence of this, is Amazon’s announcement of Alexa Hunches, in which it proactively recommends and suggests actions on the user’s behalf.

While Amazon has the first point covered, its ability to retain customers and app developers both through engaging offerings and monetisation efforts, remains questionable. So let’s explore.

(1) DISTRIBUTION

“We want customers to be able to use Alexa wherever they are.”

Land-and-expand is working. With 20 000+ third party Alexa devices, Amazon has experienced fivefold device growth in 2018 alone.⁶ This is a direct result of its Alexa Voice Service (AVS) software development kit (SDK). Amazon aims to extend interoperability into every home device, teaming up with chipmaker NXP to incorporate Alexa into all kinds of devices. This also allows for lower-end mobile integration as well as additional new form factors such as home appliances, PCs and automobiles.

Anyone should be able to buy the kit, build a product, download the Alexa software, and get everything running without any prior knowledge or any help from Amazon. They might not even know the product exists until it hits shelves.

In fact, it is suggested that companies are reaping the rewards of building Alexa-enabled devices, which, together with the Works with Alexa certification is encouraging business growth of between 40 and 50 percent in the nine-months post release.⁷

The conglomerate is also pushing its mobile platform play harder, partnering with Huawei and Motorola to introduce customised hardware solutions. It has also integrated voice functionality into its Android and iOS Alexa apps, positioning Alexa as an AI platform via Android Open Source (AOSP) to circumvent Google licensing terms and associated bloatware. This is all with the aim to become the ‘Google Play’ of smart home apps and future voice-services.

(2) RETENTION

But, many users are not engaging with their devices beyond simple voice commands. This is where the term Zombie Skill originates from i.e. these skills are accessible but not heavily used or appreciated. Hence, ~53% of consumers only engage with one to three skills overall and 14% haven’t enabled a third-party skill before. Around 30% of Amazon Skills have more than one consumer review.⁸ The same applies to general smartphone apps, with the average mobile user downloading 33+ apps on their device, but spending 80% of their time interacting with just three of them.⁹

Despite the above, Alexa skill development is high. Total skill availability is up 400% to ~50 000 since Q1 2017 and continues to grow.¹⁰ However this presents a two-fold problem:

(i) Discoverability (ii) Usability/Practicality

Borrowing the idea from Apple’s app store, Amazon does issue email newsletters to promote discoverability but is this enough? Naturally, it is hard to keep users aware on an audio-only device. Together with the fact that most skills are information-providing B2B2C, this is problematic. Hence, the introduction of the screen-integrated Amazon Echo Show models.

Are they working? It’s too early to tell, so let’s wait and see.

(3) MONETISATION

Amazon has gone back-and-forth on their monetisation policies which has rattled developers in the past.

We all make mistakes

While Amazon originally introduced an Alexa advertising network, it changed its policies and inevitably shut it down. The leading way for third-party developers to monetise from their Alexa skills was through VoiceLabs’ interactive Sponsored Messages. VoiceLabs would collectively group skills together and sell each group as an advertising network to brands. While developers previously used creative means to advertise, Amazon wholeheartedly banned “any advertising for third-party products or services”.

Perhaps the ban came because the market simply wasn’t ready? While this was back in May 2017, I wonder whether sentiment within Amazon has changed? Particularly as Google Home continues to gain traction across geographies. Considering how critical it is to encourage further skill engagement i.e. conversational brand engagement, perhaps we will see the reemergence of the ‘brand voice’ in a different form. Especially since the number of brands within the Skill ecosystem is up ~200% to 3 500.¹¹

To circumvent the above, developers can use account linking to implement a freemium or paywall model i.e. connecting the identity of the user with a user account in different systems. The downside risk is that this can disrupt the UX and frustrate users by misleading them into purchasing unwanted full versions of apps which threatens an app’s review rating.

A new hope

Amazon has implemented a new monetisation policy that has been interpreted favourably by the market. By introducing in-skill purchases, they are encouraging the sale of premium content and digital subscriptions within their skills. Developers will earn 70% of the listing price, applied before any discounts or Amazon Prime incentives.

Furthermore, Amazon has also authorised third party developers to accept Amazon Pay for in-skill purchases. This suits the growing number of game-based skills immediately, and should help to encourage further development of high-quality skills — improving the in-skill experience and engagement with other skill types.

Additional incentives to mitigate the Zombie Skill problem include Alexa Developer Rewards to encourage developers to focus on maximising customer engagement through cash incentives. Developers can also generate additional revenue by selling physical goods and services through Alexa, using Amazon Pay.

An exciting industry, yes, but it doesn’t come without its challenges. I should highlight that voice technology has existed for 60+ years and it is still in its infancy. Clearly Amazon continues to test the waters, as it should. With Google hot on its heels — healthy competition is what’s needed here.

While Amazon pushes for cross-product integration through a B2B2C/B2C model, Google’s white knight is its push to partner with retailers. The cognitive load on voice is 50% less than text, and Google wants to take advantage of this. You see, Google wants to be the go-to for answering questions (voice-based search) and monetising via adverts.

And it’s clawing back market share too. In fact, 2018 is the first time that Google has outperformed Amazon in the space…

Game on.

If you or your network is passionate about the space, and would like to learn more, please feel free to reach out.

E: [email protected]