How Machine Learning Reinvented Advertising
You’re shopping online for car mats from your cubicle at work, and you find a good price. Is it the same price you’ll pay tomorrow, searching from home? More and more algorithms are determining what price you’ll pay based on a variety of factors: seasonality, competitors’ prices, where you’re searching from, time of day, your online shopping history, etc. In fact, algorithms have not only changed online retail, but they have changed sports, medicine and real estate – even what you see on Facebook and Instagram. And algorithms are getting smarter every day.
Of course, this is going to change the world of advertising.
Algorithms have advanced to such an extent that now they’re programmed to create new algorithms, which in turn solve for different and evolving criteria. This has given rise to well-known machine-learning computers like IBM’s Watson and Google’s RankBrain.
The information learned by RankBrain is already estimated to be the third most influential criterion in Google search results, impacting how webmasters produce positive search results for brands. We have entered the age of machine learning, where you can’t game the SEO system anymore because it’s working 24/7 to evaluate content and gradually integrate new ranking factors all on its own. Don’t believe it?
Google already has the ability to interpret the intent of complex and ambiguously worded queries across unstructured databases all over the web. It’s gotten good at context and syntax — the heart of semantics. So good that you don’t need to use the same word over and over again as a keyword to boost your search ranking, an old SEO tactic. Google understands the subtleties of similar words; therefore you should be thinking more about word groups and word clusters.
Moreover, the correlation between keywords and high search rankings has fallen over time.
There’s a completely new set of criteria for high search rankings in the age of machine learning—optimized for the human experience.
Similar to the way they disrupted stock market trading, smarter and smarter algorithms are buying display ads, faster and at better prices than humans ever could. They’re writing personalized content.They’re making real-time decisions about which “suggested articles” you might want to read after finishing a blog post. They’ve made significant inroads into predictive content creation and programmatic buying. They're better at creating audience segmentation than anything media buyers can create.
Given sufficient time, machine learning will be able to understand how to avoid the common mistakes made by humans, optimize revenues, reach more consumers and identify new opportunities. This technology has implications well beyond media buying and content creation. It’s already writing songs and scripts, and eventually it will be the guiding force behind marketing through the Internet of Things.
And you thought digital marketing was complicated now!