Our behaviors are changing: we ask Alexa to take our shopping orders; we rely on Maps to recompute our driving route due to traffic; we rely on Netflix to suggest the next show or movie; and we even let our cars detect the probability of collision and steer or brake as needed.
There's an evolving community around us, one that relies heavily on intelligent machines. Whether we're at an AI tipping point or not, it's clear that artificially-instilled intellect is growing. As AI continues to change consumer behavior, brands must recognize how it can impact marketing initiatives.
1. Moving from keywords to context. Is the long-tail really a tail anymore? As search becomes conversational, particularly through the explosion of voice (Alexa, Google, Siri, Cortana), extracting context is more relevant, as possible permutations of keywords will rise exponentially (e.g. 4 core keywords = 24 queries; 8 keywords = 40,320 queries). While there are tactical ways to manage this, the imperative is that context becomes more important and AI can be used to extract it.
2. Using structured and unstructured data together. Marketing channels are intricately interconnected, and not all data points are structured. There's been a surge in brands trying to understand this connection through various attribution solutions. While a step in the right direction, attribution approaches are still mostly linear and largely use structured data. As AI-based solutions evolve, structured and unstructured data (signals, images, emoticons, sound, video, etc.) will become nodes in a neural network, dynamically determining exposure to each touchpoint. This will cause a shift in how channel performance is managed, measured and reported.
3. Creating shorter search funnels, with more content optimizations. Cognitive AI solutions aim to connect users with the information they're seeking, faster and with a higher degree of relevance. As these systems get more predictive, it's plausible that drill-down searches will be reduced, and the responsibility of delivering the right information, in the context of the search, will fall on the systems that render the content. This implies that content optimization will not only need to be connected tightly with the ad delivery system, but will also rely more on AI and decisioning algorithms.
4. Training the AI system. The most challenging task in any good AI application is training the AI algorithm. This requires large amounts of good data (and content) for training, as well as validation, which can be an intensive and costly effort. As brands look to use AI systems, they'll need to account for these costs.
5. Rethinking segmentation. Segmentation is about creating "segments" that are more likely to convert. As AI systems evolve, a user's interaction with the "next" stage is always maximized to the highest probability. As such, preformed segments lose their meaning. Accurate understanding of context will drive conversion. Thus, we'll need to rethink when and where we use classic audience segments vs. AI.
6. Assessing your data science capabilities. The boundaries around what machines can do are being redefined. Therefore, brands must also redefine their data science capabilities. Cognitive AI systems, neural networks, predictive analytics using AI, self-learning systems -- all these things require a heavy dependence on data science. To adequately consider these factors, a robust analytics team is a requirement.
Brands should take this plunge, as leading this evolution of intellect is quickly become a key lever in driving performance.