By, Jacob Leveton, Director of Product for Veritone
Following the customer journey and seeing which interactions led to a buying decision has become easier for marketers over the years, thanks to artificial intelligence (AI). Take radio, for example. Before the digital age and the use of AI, humans were tasked with culling through thousands or hundreds of thousands of audio hours to find a link between an ad placement and whether or not it resulted in the desired outcome for the ad client, such as a visit to a specific physical store or the purchase of a certain product. It’s one of the reasons radio began to lose market share in advertising dollars, as digital advertising provided more detailed customer journey attribution tracking.
Today, however, it’s possible to perform attribution analysis from queries that pull a massive amount of data to make such correlations. With the use of machine learning (ML) and natural language processing (NLP), which rely on AI-based algorithms, these queries deliver results in near real-time and have become more accurate. That’s important when trying to distinguish the outcomes of broadcast ad placements not just on radio but across multiple formats. Whether tracking and analyzing pre-produced spots, organic mentions, live reads or website interactions, AI, ML and NLP make it easier to detect written and spoken nuances and sentiment to discover the value of sponsored content versus a mention in a conversation by an influencer or celebrity, for instance. Success can be measured granularly by criteria such as location, day of the week, frequency, time of day or length, which helps radio and TV advertisers further optimize their campaigns in a way that was previously unavailable and demonstrate ROI with advertisers, greatly improving their renewal and new advertiser contracts.
Marketing Technology News: DoubleYard and Laserfiche Partner to Provide AI OCR Technology for Both Digital Print and Handwritten Content
Now comes the metaverse, which introduces an entirely new way for brands and customers to interact with one another. While the digital (online) world can be seen as a closed circuit, with every new and existing site, interaction and action trackable and traceable in an ecosystem, the metaverse will likely mirror the real universe—which is ever-expanding. And it won’t be monolithic. Companies like ROBLOX, Sandbox and Decentraland have been attracting visitors to their metaverse-like platforms for years. And big brands are starting to take notice. At CES 2022, for example, Samsung launched Samsung 837X on Decentraland, where avatars are met by a virtual guide in the Samsung lobby who directs players to three unique interior spaces they can explore. While it may be easy to determine where an avatar goes and what it does based on structured data—think attending a concert or sporting event, or buying a nonfungible token (NFT) or digital product—it’s the unstructured data that will present the biggest challenges.
Attribution analysis becomes more difficult in the metaverse because avatars are in many ways limitless in what they can experience and create. As a result, their conversations and behaviors will be more unscripted and unpredictable. More importantly, there will be multiple languages and dialects to sift through. So, correlating some of this native behavior and things that are tied into a brand’s content will require hyper-intelligent, AI-based solutions to bring that tracking into the metaverse. Furthermore, just like Facebook, Google, Twitter and other platforms on which ad dollars are spent, every marketer and metaverse owner is going to be responsible for attribution analysis on some level. And the level of data portability and data availability is going to be a conscious decision that may help or hurt where marketers spend their money. Additionally, as metaverse creators think about how they want to build the ecosystem, privacy will be a significant factor. We see this now in the digital world, with privacy policies and data accessibility. While more privacy is good for users, it makes things far more complex for marketers. Finding the right balance in how much data will be available to the marketer, and how privacy policies are communicated to users, will add a layer of complexity.
Marketing Technology News: MarTech Interview with Amy Heidersbach, CMO at Persado
A couple of years ago, one of the big conundrums was that everyone wanted a multi-touch attribution suite for their digital marketing because they thought if they could bring all of the data into one place, there would be a clear consumer story that would tell them how to spend the money. And after marketers spent a lot of money and time building up a multi-touch attribution suite, they realized that they had too much data, and they couldn’t tell what was moving the needle. In the online world and with radio, Veritone has witnessed a great deal of success in generating consumer action with sponsored influencers and integrations that are more organic than prerecorded ads. Organic host-read or influencer-read spots are often better received by the audience but are much trickier to track without the help of AI and ML-based solutions. The metaverse should be no different, as long as markets and metaverse owners have a way to cut through all the noise and get to the right “signal.”
In the same way deep learning is showing great potential in wireless communications for improving signal processing, AI-powered attribution solutions have the same potential to help sort through millions and billions of data points to identify and validate which marketing touchpoints and/or influencers are driving specific business outcomes in the metaverse. It’s all about data processing at scale. So, the sooner brands recognize these challenges, the better they’ll be positioned when they are ready to enter the metaverse.