As Alexa rapidly joins the ranks of Siri, Cortana and Google Now in the growing ranks of consumer voice user interfaces, or digital personal assistant to most of us, many marketers are starting to ask themselves just how to overcome the digital gatekeepers. It’s not a new thought – search engine marketing has been attempting to overcome or work within the boundaries of algorithms for some years now – but this new element of machine learning fuelled by developments in AI technologies adds an additional layer of complexity to marketing within a digital landscape.
It appears that these digital personal assistants are becoming more integrated into our lives as people become more used to issuing voice commands, and the assistants themselves become more functional in responding to how people speak, rather than sticking to set vocal terms. What is perhaps less well known in the context of marketing is that machine learning, and responding to it, has been around for some time already, especially in retail and social.
Most people, for example, understand that their Facebook feeds are chosen selections of the most engaged-with posts, based on their previous interactions and preferences. Other content publishers are now tapping into this filtering technology, enhanced by insights from Facebook based on those users’ likes and preferences. In retail, we’re seeing brands moving on from rudimentary content recommendations based on previous purchase data, or ‘other people shopping for this item bought x’ suggestions, and instead presenting product selections likely to be interesting based on peoples’ individual profiles. One iPad purchaser, for example, may be more likely to see suggestions for a pink case to accompany it, while another could receive recommendations for a charger, stand, or a blue case.
The risk ahead was best described by Upworthy’s Eli Pariser called ‘filter bubbles’ in his standout TED talk. Content is increasingly being sifted on people’s behalf, and with these assistants becoming widespread across tablet and mobile devices, their roles in this filtering process as they learn and make recommendations about their owner’s preferences will grow. This is already present in Google searches between one user and another, even if they are not logged into their own individual Google accounts at the time. So the challenge here for brands is to understand and overcome these new barriers to reaching their target audiences online, as there are no ‘standard’ sets of search results anymore. Factor in the fact that Amazon is growing its own brand grocery offering as we speak, and it becomes more of a pressing issue to brands that when a shopper asks its phone to remind him to buy more batteries, it is the right brand for that person which is delivered.
The issue of AI and growth in machine learning isn’t all concerning, however. They can be turned to a marketer’s advantage, and plug some of the knowledge gaps we have struggled with for years. For example, smart machine learning and game theory are now started to be deployed to explain how different marketing techniques can work together collectively to influence a sale. Advertising has been unable for some time to truly attribute which part of an integrated campaign ultimately triggers a purchase, mainly because it is often a cumulative process and one which will differ from individual to individual. In using game theory to understand how channels collaborate, and machines to spot trends in the data which humans simply can’t identify, the industry is getting a lot closer to true attribution and weighting of cause and effect.
Artificial intelligence is beginning to impact the creative industries we all work in, and it will continue to develop and adapt as the number of channels at a marketer’s disposal proliferate. The growth of personal assistants could easily become a boon if they truly filter preferences to make sure information is tightly targeted to people’s needs and preferences without bias – or at least, without bias beyond that individual user’s own likes and dislikes. If, however, they become the digital PAs who stop brands from easily accessing their customers through their devices, are we likely to see two types of marketing develop - one targeting the machines, while others focus on the human? Only time will tell.
By Jonathan Moore, SEO group head at equimedia
PrivSec Conferences will bring together leading speakers and experts from privacy and security to deliver compelling content via solo presentations, panel discussions, debates, roundtables and workshops.
For more information on upcoming events, visit the website.
comments powered by Disqus