The bullish AR, voice & chatbot predictions that didn’t come to pass by 2020: stats

Sometimes, these are just for fun; other times, companies are basing their strategies on how they think things will evolve.

The year 2020 has had perhaps an above-average share of predictions attached to it, thanks to its roundness, its symmetry, and the futuristic possibilities it represented. I’m not talking about predictions made in 2019 about the following year, but the more far-flung and ambitious types of predictions set down in the early to mid-2010s. I’m sure you can think of a few off the top of your head, cited in presentations or dropped into talks for colour: “By 2020, the average person will have more conversations with a chatbot than with their spouse.” “By 2020, 30% of browsing sessions will be carried out without a screen.”

I thought it would be interesting to check in on some of the more ambitious 2020 predictions and see how they measured up against reality. Some of them are patently wide of the mark, based on the belief that a particular trend would dominate the next several years, but which then failed to materialise; others are less far-fetched than they might sound.

Chatbots

“By 2020, chatbots will power 85% of customer service interactions” / “By 2020, the average person will have more conversations with chatbots than their spouse”

These two predictions – often cited one after the other – were popularly presented in the mid-2010s as evidence of the chatbot revolution and the way that chatbots looked set to transform our interactions not just with brands, but with each other. Both statistics come from Gartner, but it may surprise you to learn that they were made a full five years apart.

The “chatbots will power 85% of customer service interactions” stat actually comes from a brochure that was produced by Gartner for its Customer 360 Summit in 2011. It didn’t originally reference chatbots – unsurprisingly, since chatbots weren’t really a “thing” in 2011 – but predicted that “By 2020, customers will manage 85% of their relationship with the enterprise without interacting with a human.”

This is a bit different to “powering 85% of customer service interactions”, as it refers to 85% of each customer’s relationship with a business – not 85% of all customers’ relationships with businesses. The context suggests that Gartner was thinking of this in terms of interaction with CRM systems and other automation tools.

It’s still a far-reaching prediction, and I don’t think we’re quite there yet in 2020 – in part because companies have been slower to implement automation technology than might have been expected, but also because human interaction in customer service continues to be valued in the right context. A 2017 report by MIT Technology Review and Genesys found that 91% of businesses with high levels of customer experience and world-leading brand recognition use AI solutions to some degree to increase customer satisfaction, while 60% believe they have the right mix of “live” and automated channels. Leading firms were also found to be 20% less likely to resort to a technology-based solution for customer support than the average company.

Come the chatbot craze of the mid-2010s, Gartner’s statistic suddenly took on a life of its own and was repurposed to support arguments for the impending dominance of chatbots – paired with another prediction that Gartner made in 2016: “By 2020, the average person will have more conversations with bots than with their spouse.”

No creative reinterpretation was needed to make this stat fit the narrative, but it’s clear that it wildly over-estimated how prevalent bots would be – or wildly under-estimated how much people communicate with their spouses. (I’m not sure whether those of us who aren’t married are meant to be included in the statistic – perhaps for “spouse” substitute “close family members”, “housemate” or “cat”…)

Gartner seems to have recognised this as well, as in 2018 it issued an updated prediction that by 2020, 25% of customer service operations will use virtual customer assistants. This one seems more likely to hold true, since more than half of organisations have reportedly invested in virtual customer assistant (VCA) technology already and found significant reductions in support query volume and improvements in customer satisfaction. Chatbots might not have taken over, but they can still be very useful under specific circumstances.

Voice

“By 2020, more than 50% of searches will be carried out via voice”

I had to knock this one on the head early. Those of you who are regular Econsultancy readers will know well what’s coming, as I have a bit of a vendetta against this stat. I’ve debunked it quite extensively elsewhere, so I won’t rehash it all in too much detail here, but this is probably the most persistent prediction about the year 2020 despite also being the most demonstrably inaccurate.

Suffice to say that you shouldn’t be citing this stat in articles, talks or presentations (least of all in 2020) and anyone who does cite it uncritically deserves more than a little bit of scepticism directed their way. It seems (to me) quite obvious that we aren’t using voice technology enough for 50% of searches to be coming from voice in 2020, but this stat seems to have a near-mythical hold on our collective imagination, to the point where it is still cited by some surprisingly senior marketers.

For a detailed explanation of how this statistic originated (hint: it didn’t come from comScore) and why it has become so distorted in the retelling, you can read my piece on why we need to stop repeating the “50% by 2020” voice search prediction. I also ran the numbers on what it would take to get to the volume of searches this prediction needs to come true the future of voice search. Finally, for a look at voice in the year 2020 and why I think talking about “voice search” at all is misleading, have a read of my piece from January, ‘Voice in 2020 is (still) not about search’.

“By 2020, 30% of browsing sessions will be carried out without a screen”

This prediction from Gartner in 2016 is cousin to the “50% by 2020” voice search stat and often pops up alongside it, although it hasn’t suffered from the same level of distortion. It originates from a similar vision of the world in 2020 as a much more voice-driven one.

Despite the adoption of smart speakers into millions of homes, however, they are rarely used for any activities that could be classed as web browsing. Just as searching the web has turned out to be a much less popular use for voice devices than many marketers believed, browsing the web has seen a similar lack of traction. Voice shopping has yet to take off in any meaningful way, and voice applications have also been met with disinterest from the majority of consumers.

Both articles I’ve linked above delve into exactly why voice devices aren’t being used as experts predicted they would be, but the short version is that there is little that they can do which screen-based devices (particularly smartphones) can’t do better – and in a way that people are more accustomed to. Gartner’s prediction may have overestimated the extent to which consumers would be willing to shift their internet browsing, buying and searching habits for the sake of the occasional hands-free convenience.

And far from doing without screens thanks to the prevalence of voice devices, we’re instead enjoying a new breed of voice devices that incorporate a screen into their design – meaning that the trend which was meant to realise this stat (screen-based devices giving way to voice-based ones) has ultimately reversed itself.

AR & VR

“AR and VR will generate $150bn in revenue and disrupt mobile by 2020” / “By 2020, there will be 1 billion augmented reality users”

The mid-2010s was an optimistic time in marketing: along with the hype around voice search and chatbots, enthusiasm for AR and VR was also at its peak. In 2015, writing for Tech Crunch, managing director of Digi-Capital Tim Merel predicted that by the year 2020, AR and VR would collectively generate $150 billion in revenue, and have disrupted mobile completely.

Merel forecast a smaller share of that revenue ($30 billion) coming from VR, whose addressable market he believed was confined to gaming, 3D films and niche enterprise uses. He believed AR had a much larger potential market with “hundreds of millions of users”; applications in commerce, voice, enterprise and data; and the ability to drive high revenues from hardware (like Magic Leap and HoloLens).

Around the same time, it seems, another prediction sprang up that augmented reality would have (or gain) one billion users by the year 2020. It’s difficult to pinpoint an origin for this stat, and like many dubious statistics, it’s typically vague: what constitutes a “user” of AR? How often do people need to use AR for it to count? Is this a global prediction or does it refer to a specific part of the world? It’s easy to dismiss it as a characteristically unrealistic marketing myth, but as it still appears in roundups of AR statistics, I decided to take a look at it.

Both of these predictions deal with the perceived dominance of AR and VR in 2020, which is why I’ve grouped them together, but they actually relate to quite different things. One is about how profitable AR and VR are, and the other is simply about how many people use them; it doesn’t need to be a use that generates much – or any – revenue.

We know that in 2020, neither AR nor VR has turned out to be the commercially successful phenomenon that was expected just a few years ago. In AR’s case, the technology just isn’t that good yet, outside of a few specific applications like virtual makeovers. Most of the promised use cases for it (like commerce, or interactive marketing) haven’t materialised. As for VR, Merel was right in that its applications are mostly confined to gaming and other niche use cases, primarily entertainment, but even there it hasn’t attained enough mass adoption to be particularly profitable.

So, how much revenue are AR and VR generating in 2020? According to AR Insider’s Global AR Revenue Forecast for 2018-2023, revenues from AR in 2020 will total less than $10 billion and are only projected to reach $27.4 billion by 2023. Meanwhile, AR Insider puts the global revenue from VR at $8 billion for 2020, growing to only $10.6 billion by 2023. All told, it’s a fraction of the amount that Merkel predicted in 2015 – and neither technology seems likely to disrupt mobile any time soon.

But when we look at the number of users of augmented reality, here’s where things get interesting. eMarketer figures from 2019 forecast that AR will have 77.7 million users in 2020, and 85 million by 2021, in the US alone. According to eMarketer, they revised their user estimates for AR upwards between 2018 and 2019 “to reflect changing market dynamics” like the number of AR users on social networks like Snapchat, Pinterest, Instagram and Facebook. It’s not Merel’s “hundreds of millions”, but if we assume that regions like Europe, Australia and Canada (which are harder to find figures for) also have tens of millions of AR users, we could easily get to 100 or 150 million.

A marketer’s guide to Augmented Reality (AR)

What about the “one billion AR users” prediction? Well, you might have noticed that my list of regions above is missing some key players, particularly one country with a vast population that accesses the internet through mobile first and foremost. I’m talking, of course, about China. And while I think most of the predictions about the dominance of AR in 2020 were concerned with the ‘west’, it doesn’t do to leave out China, which is far more of a hotbed for AR technology than the US and other western regions.

With its mobile-first approach to the internet, AR has found a ready-made consumer base in China where it is integrated with everything from short videos to maps, search and ecommerce. And while China is still finding ways to drive revenue with AR, there’s no question about its popular uptake: search giant Baidu’s AR applications alone were reported to have more than one billion monthly users in 2017.

While I think that this statistic being on the mark is more of a coincidence than a well-made prediction, it goes to show that some of our visions of the future in 2020 are playing out – just not where we might expect them to be.