Search is no longer the traditional text-input field on the search-engine page. No – something completely old-fashioned and non-digital is celebrating its return: the spoken word. Google Home, Alexa, Cortana, Siri – almost all the Big Tech players are betting on Smart Assistants that can interact with the user, with or without the aid of a screen. Some of them see the next big thing coming over the horizon, while the others see a hype that will soon peak and fade.
So will Voice become the new Touch? How will voice control develop further – and what are the implications for customers and companies? We asked two thoroughbred search experts, Melanie Schlegel and Marcel Kollmar, the Heads of SEO (Search Engine Optimisation) and SEA (Search Engine Advertising) respectively at OTTO Online Marketing.
Marcel: In qualitative terms voice recognition made a giant leap in 2017. The question now is how quickly people can get used to talking to their devices and what Use Cases the Digital Assistants can fulfil – that is, how good the results are that they can deliver to the user. For questions such as “How much does a mouse weigh?” Google is already great – give it a clear question and you’ll get a clear answer. It works. Long-term, though, with Voice I don’t just want to search for information, I want to interact, navigate further, or carry out transactions. So dialogues need to be developed in which the machine can also ask checking questions.
Melanie: To begin with, Voice is naturally an alternative to keying in queries on a tiny touchscreen. That said, voice-controlled Digital Assistants are no replacement for smartphones or large screens, and therefore on their own will not be the next big disruption. However, voice-controlled Assistants are becoming part of the networked digital ecosystem, the much-discussed Internet of Things. They interact with TV screens, manage lighting, temperature and other devices. So the topic is suddenly becoming a big one.
Marcel: For 20 years we have been accustomed to inputting the closest-match keywords and operators so that search engines can produce accurate results. However, this is not a natural way of using language, much more a means to an end. Voice is leading us back to normality, and this also means a return to ‘W’ questions – that is, whole sentences. Here, the latest machine-learning algorithms help derive meaning from the interplay of numerous words and to assign them to their intended context.
20 years ago, everyone laughed at the people who typed “Dear Google…” into the search field and formulated their query like a letter. Well, in a way those people were ahead of their time!
Melanie: There’s also a funny side to this. 20 years ago, everyone laughed at the people who typed “Dear Google…” into the search field and formulated their query like a letter. Well, in a way those people were ahead of their time! Today of course, this is precisely the natural approach to language that the Big Tech players, the GAFAs, are preparing for. So I therefore see the greatest use of Voice amongst young people, who have rediscovered natural speech through WhatsApp voice messaging for instance. But Voice is a boon to older people too, who find natural speech far easier to use than a small touchscreen.
Melanie: With a short voice command we can now order a pack of paper hankies or a litre of milk – that is, ‘Fast Moving Consumer Goods’. There’s no great risk here, as the prices are so low and the purchasing decision carries no emotional weight. However, no-one would say, “OK Google, put whatever dress in my shopping basket.” Especially for emotional, complex purchasing decisions like these, firstly as a customer I need a screen on which I can view what the system is looking for, and secondly the online shops need outstanding filters and product data. The Assistant gets better the more data it has available. It needs to be able to select from this data-set in order to provide expert advice: “Would you like a red dress or a green one? Strapless, long, short, what kind of fabric?” At OTTO we have real goldmines for product information, especially with customer evaluations. The software could make a selection based on these evaluations and then give tips such as: “Be aware that according to other customers this dress is bigger – perhaps you’d like to order one size smaller?” The most important thing, though, is to start actually using the new software and to experiment with it, as we have already done for the German launch – for instance with OTTO Action for Google Home, which is currently under further development.