As consumers flit between smartphones, web pages, call centres and stores, retailers struggle to keep track of them. Eight out 10 retailers say a single view of the customer is an issue for their current or future agenda, and only 8% have successfully achieved it, a study from SAP and PwC has found.
Just as they realise the extent of what we might call the customer intelligence omni-challenge, a new source of data is about to arrive. According to Gartner, worldwide spending on virtual personal assistants, or smart speakers, will reach $3.5bn by 2021, up from $720m in 2016. By 2022, says Juniper Research, 55% of US households will have a device such as the Amazon Echo, Google Home or Sonos One.
This growth should interest retailers. Amazon is already selling products via its voice platform, while in the US, retailers Walmart and Target have formed alliances with Google. In the UK, Tesco sells through the Google Home device via an IFTTT channel.
The question for retailers and other consumer-facing organisations is what this new voice channel means for their understanding of customers and how voice data will integrate with existing investments in customer intelligence.
Kees Jacobs, Capgemini’s global head of insight and data for consumer products and retail, says: “It should be the next wave of CRM [consumer relationship management]. Companies that have good omni-channel CRM can build on that. The goal is to truly understand all aspects of how the consumer thinks and acts to form meaningful, empathetic relationships with them. No company is there yet, but mature CRM may put you in a good position.”
One of the initial difficulties organisations face is getting hold of the voice data, says Jacobs. “At the moment, it is early stages – a bit of a wild west,” he says. “Conversations are orchestrated by Amazon and Google. If that is going to remain the case, the primary data is collected by these firms as they deliver the device and intelligence.”
Kees Jacobs, Capgemini
However, retailers and consumer goods manufacturers that have developed applications – Amazon calls them Alexa skills – of voice platforms can collect some data from the interaction.
“It is still a world where only retailers or manufacturers that have skills will collect data,” says Jacobs. “That is the situation right now.”
Hive, a provider of home control services and devices which is owned by UK utility British Gas, provides skills for Amazon Alexa that allow customers to control heating and lighting with voice commands. Jo Cox, Hive’s commercial director for UK and Ireland, says it keeps data on interactions with its skills and uses it, along with other customer data, to develop new offers and services.
For example, after analysing data with the help of subscription application provider Zuora, it now offers a Mimic Mode, a security feature that allows users to replicate lighting patterns so it looks like they are at home when they are out.
“At the moment, the services we provide are quite generic, but we will start to personalise them through our own data,” says Cox. “Amazon doesn’t share data with us. We will be using our own data platform to personalise Hive: 80% of that will naturally be available in Alexa.”
Customer intelligence opportunities
Accenture’s managing director for advanced customer strategy, Rachel Barton, says that although smart speaker and voice platform providers such as Amazon and Google currently hold the data, in the future it is likely to be available to third parties, creating opportunities for customer intelligence. Those that have access to the data will find it valuable in shaping offers to consumers, as well as future products and services, she says.
“What organisations have never really known is the behavioural pattern of consumers’ lives,” says Barton. “Voice command will provide a very different relationship. Consumer companies could know when and how consumers make decisions about what they are having for dinner. That is a huge opportunity to tailor the proposition and marketing message to a relevant template. It is much more need-driven than ever before.”
But few organisations are thinking about voice data in this way, says Barton. “There are brands that are sophisticated in the way they use and store data, but in general, there are a really small number of brands that are thinking beyond traditional datasets.”
Those that are already making progress in an omni-channel understanding of their customers will be well placed to benefit when they get greater access to voice data, says Barton.
“It is going to be a case of understanding your ecosystem and being able to respond in the most optimised way, rather than thinking of a series of channels,” she says. “It will still be about the consistency of experience across channels.”
The wake-up call for organisations will come when they realise that voice assistants could have the power over choice of brand, says Barton. Brands or retailers that are unable to serve voice content based on customer intelligence could be literally left out of the conversation, she adds.
“If you think of ordering washing powder, many consumers are less bothered by the brand,” she says. “Shopping on the internet, you can pick whether to buy brand A or B based on familiarity. If you are using voice commands, you do not get anything to prompt you to buy a brand.
“There is going to be period of transition. There will be swathes of customers buying on these platforms, so how are brands going to influence them? Through traditional means, such as advertising? Or are we going to see digital assistants helping us select the brand we want with the consumer setting the options, such as ‘I want a non-bio and eco-friendly washing powder’. If that is the case, then digital assistants have huge power to kill a brand. This subtlety has yet to be fully thought through.”
Tone of voice captured
Mike Lowndes, Gartner research director digital commerce, says one major advantage of voice data will be knowing the identity of the shopper, which is not always the case with other interactive customer technology, such as conversation bots running on a website. But organisations could also understand the mood a customer is in when he or she is making a purchase, he says.
“In terms of voice intonation and emotion, that is still in the research phase,” says Lowndes. “It is being worked on and there are papers published in Nature and Science. It is academic now, but within five years, we could see companies using it.”
Consumers using voice assistants to buy goods and services shifts shopping towards a concept that commentators are calling “conversational commerce”. But enterprise data will not only have to cope with new forms of customer data, says Lowndes – consumer-facing organisations will also have to understand what their products “mean” in a conversational context.
“Many retailers have metadata for their own products, but that is not good enough,” he says. “They need a service that will enrich metadata around products, that will then get a more accurate response to search queries. It is the shift from key-word to intent-based searching. It is not about the customer asking, ‘can you find a radiator with capacity of 80 British thermal units?’. It is more, ‘I want a radiator to work in this room’. That is something companies need to understand about their products, and we are not there yet.”
SAP, as a CRM software supplier, is taking an interest in analysis of customer voice data using artificial intelligence (AI), but is not currently offering products in this area. Its chief value adviser for retail industries, Shane Finlay, says: “The way you say something to voice ordering is something that can also be collected. With voice, you can capture emotion.
“SAP is active in the discussion on AI [to analyse voice data]. We are reaching to meet customer demand with partnerships such as the ones we have with Google, Facebook and Apple. You need a large ecosystem to work on the use-case and go-to-market strategy.”
Access to customer and consumer voice data is set to be hotly debated for some time.
Read more about modern approaches to CRM
- How CRM systems are helping couples say ‘I do’.
- Artificial intelligence was the main topic at the PegaWorld 2017 conference in Las Vegas, with keynotes, breakout sessions and the 4,000 attendees discussing the CRM industry’s latest trend.
- Larry Augustin, CEO, SugarCRM: six 2017 predictions for a shifting CRM market.
Amazon says audio information is not stored in CRM systems or used for targeted advertising or customer recommendations. Similarly, the company says it does not share voice and audio data with developers or third parties – it is only used to improve question answering and AI models.
An Amazon spokesman says: “There are a number of ways Alexa helps customers shop. Alexa may look to see if the item you are requesting is one you have purchased before, offer a similar item or use the Amazon’s Choice algorithm to find a well-rated, well-priced item that ships with Prime.
“Amazon takes customer privacy seriously and we have taken measures to make Echo secure. These include hardware control via the ‘mic off’ button, rigorous security reviews, and encryption of utterances between Echo, the Echo app and Amazon servers.”
The Echo device allows users to delete all voice recording history via a specific button.
However, Amazon would not comment on whether it plans to use Echo voice data and consumer queries for CRM, promotion and product development in the future.
For Capgemini’s Jacobs, the dominance of a handful of players controlling consumer voice assistants and access to consumer data, whether Amazon, Google or others, should be a concern. He feels it could hinder broader industry development of the voice platform and consumer acceptance of it.
“Ultimately, there needs to be clarity and more openness in sharing voice data and insight,” he says. “It relates to the decoupling of devices, the intelligence behind them and the consumer engagement and business intelligence behind that. Right now, that is in a few hands. For consumers and the industry, we need to go to a level playing field.”
Whether Google, Amazon and the handful of firms popularising voice assistants agree is one question Alexa is unlikely to answer.