Five years on from the initial launch of the Amazon Echo in November 2014, many would claim that we’re in the midst of a ‘voice revolution’.
Certainty about this fact has only grown with every new announcement and product launch, from new voice assistants, to new integrations with smart products, to the much-talked about demonstration at Google I/O 2018 in which Google’s new smart assistant, Duplex, made a booking over the phone. Surely, these are all signs that we are now living in a much more voice and AI-dominated world?
Accordingly, companies have scrambled to respond by securing their presence on voice-first devices, optimising for voice search and developing voice-powered experiences. In late September it was announced that 100,000 Alexa skills had been launched worldwide, up from 50,000 in September 2018; as of January 2019, Google had more than 4,200 Assistant Actions. Major brands including Domino’s Pizza, HBO, Diageo, Sephora, 1-800 Flowers, BBC and Nike have developed voice experiences for the Echo and Home, and in several cases won awards for them.
Yet despite all of this, there is a surprising dearth of case studies available that talk about the results that these voice apps have yielded for brands. How much usage are they seeing? Have they yielded any revenue or brand lift, or brought in any new customers?
The benefits of being present on voice devices are assumed to be self-evident: developing a voice app, or optimising for voice search, is “future-proofing” for brands; it “gives them a presence” on what we are repeatedly told is a huge, up-and-coming channel. But if this is the case, then why aren’t more brands talking about the results they’ve seen from their voice experiences?
This conspicuous silence around the outcomes of creating branded voice experiences mirrors the lack of concrete figures provided by either Google or Amazon on how much Actions and Skills are actually used – as well as the lack of any evidence to suggest that consumers are using voice to search the web. So where did the notion of voice as a thriving, up-and-coming channel actually come from – and what is the reality?
The myth of voice adoption, or why smart speaker unit sales aren’t enough to prove a trend
The rapid rise of the smart speaker is often compared – favourably – to the rise of the smartphone. This is usually based on its widespread adoption: investment manager Ark Invest has predicted that by the middle of 2020, more than 50% of US households will have at least one smart speaker, putting its time between early adoption (10% of households) and mainstream adoption (50%) at just three years. The smartphone and the tablet, by contrast, both took four.
The rapid rate of smart speaker adoption and their impressive growth in unit sales are often held up as proof of a massive shift in consumer behaviour to favour voice-first devices. After all, these huge numbers have to be indicative of something, right?
There are two key problems with relying on sales figures as evidence of a trend in consumer behaviour, however. One, when we talk about the sales figures, we don’t tend to mention the extent to which Amazon and Google have stacked the deck by artificially pushing down the prices of their smart speakers (or even giving them away for free) – specifically to achieve this kind of market saturation. Ark Invest points this out in the article I linked further up:
“Why then are Alexa speakers being adopted at a record rate? Perhaps the biggest reason is that, as Amazon, Google, Apple, and others fight to dominate smart homes, smart speakers, particularly Alexa, seem to have become the best bargains in the speaker space.” In a footnote, the writer adds, “Amazon regularly discounts their smart speaker devices. This Prime Day the Echo Dot was 50% off.”
The second problem is that smart speaker sales figures don’t tell us anything about how consumers are using their devices – how often they use them, whether this usage drops off as the novelty wears off or remains consistent, and what kinds of functionality they make use of.
Some research has shed light on this: the most recent US Smart Speaker Consumer Adoption Report from Voicebot, released in March 2019, found that 26.5% of smart speaker owners used their devices “Never or rarely” – up from 12.9% the previous year. The percentage of monthly users of smart speakers also increased, from 23.5% in 2018 to 26.1% in 2019, while the percentage of smart speaker owners who used their devices daily was found to have dropped significantly between 2018 and 2019, from 63.6% to 47.4%.
Voicebot is still optimistic about this trend, writing that, “Three out of four smart speaker owners still report being [at least] monthly active users. As long as we see that type of consistent usage along with continued growth, smart speakers will continue to grow in importance as a voice assistant channel for consumer engagement.” However, you could be forgiven for thinking that a 16.2 percentage point drop in daily usage of smart speakers is a bad sign for long-term consumer engagement.
Research has also found that the vast majority of consumers are using their smart speakers as just that – speakers. Multiple surveys of consumer behaviour have indicated that playing music is the most popular function for smart speakers, with listening to the radio another popular use case. While this is good news for audio streaming providers and radio broadcasters, it doesn’t suggest that consumers are making much use of the wider voice functionality of their smart speakers.
A November 2018 report by Reuters Institute, ‘The Future of Voice and the Implications for News’, for example, found a vast gap between how frequently UK consumers use certain smart speaker features and how much they value them. It found that playing music was the most-valued smart speaker feature by a long way, used regularly by 84% of consumers surveyed and valued by 61%. By contrast, although 66% of consumers used their smart speakers to answer general questions (i.e. conduct voice searches), only 6% believed this to be one of the most valuable features.
When it came to the types of features that might involve voice apps, the prognosis was even poorer: only 11% of UK consumers used their smart speakers to play games, and no consumers surveyed indicated that this was a top valued feature; additionally, just 9% of consumers used their smart speakers to order products online, with only 1% indicating that this was an important feature.
Some voice advocates have quite reasonably made the argument that many early smartphone owners similarly weren’t aware of most of the capabilities of their devices, and tended to barely scratch the surface of the available features. However, this is where the comparison to smartphones falls apart, because at the time of writing, it has been five years since the US launch of the Amazon Echo, and three since the US release of the Google Home.
By anyone’s measurement, we should have moved out of those “early” stages where device owners stick to the conventional features that they’re used to and into a broader usage of voice device capabilities. Five years on from the release of the first iPhone (mid-2012), it would be very unusual to find a smartphone owner who used their device as just a phone. And yet this is the situation that we’re currently in with smart speakers – and have been for some time.
Wanted: a breakout hit in the voice app space
As I mentioned earlier, there’s no shortage of voice applications for consumers to choose from, nor of major brands creating applications for voice. Despite this, we’ve still yet to see any breakout hits in the voice app space. As one developer of a well-reviewed Alexa game told Bloomberg News in March, “This platform is almost four years old, and you can’t point me to one single killer app.”
Voicebot’s Smart Speaker Consumer Adoption Report 2019 surveyed US smart speaker owners on how they discover voice apps, and found that almost half of consumers – 49.7% – simply don’t. What’s more, this percentage is on the increase: when Voicebot asked consumers the same question for their 2018 report, 48.2% of consumers said that they don’t discover voice apps.
In more bad news for brands, Voicebot’s Smart Speaker Consumer Adoption Report 2019 found that the majority of users – 51.3% – have never used a third-party voice app at all. Taken together, these statistics suggest that far from trying out third-party voice apps and having a poor experience, most users simply aren’t interested in them.
In 2017, research by industry startup Voicelabs also found that when consumers did try third-party voice apps, retention was poor: on average, only 3% of consumers who enabled a voice app would still be an active user of that app by the second week. Now granted, we’ve also heard that retention for smartphone apps can be poor, but not this poor: according to a 2016 retention report by Braze magazine, 13% of Android app users and 11% of iOS app users are still active a week after downloading an app.
Given this, perhaps it’s not so much of a mystery why brands aren’t sharing any figures from their forays into voice. But what does this mean for the future of brands on voice platforms? Should brands forego investing in voice altogether?
In my opinion, brands should be prepared to take a very long view on voice as a marketing channel. Thus far, the hype around the “voice revolution” has largely been driven by tech companies vying with each other to produce new devices and assistants, but this situation can only sustain itself for so long if consumer behaviour isn’t there.
Smart speakers and other voice-controlled devices are an attractive novelty to consumers, but so far, they can’t achieve anything that isn’t already possible with a smartphone or computer. We’re starting to see the very beginnings of a unique niche for voice emerging with the voice-enabled car, but it will take a radical shift in behaviour for consumers to get used to interacting with their computers via voice and sound, rather than text and visuals. For that to happen, voice will need to have much, much more to offer consumers than a slightly more useful way of compiling a shopping list.
Brands who are interested in investing in voice should take a critical eye to the statistics, and question whether they justify the conclusions being drawn from them. They should be wary of any hyperbolic predictions that don’t seem borne out by reality. And above all, they must not lose sight of what consumers are actually doing – which holds far more sway over the success or failure of a voice application than a hundred marketing industry commentators.
There may well be merit in experimenting with and getting to grips with voice for when the long-awaited consumer shift towards voice computing does arrive – but don’t be surprised if that doesn’t happen for a very long time.