Somewhere in a sleepy North London suburb, a shopkeeper ritualistically opened his daily newspaper.
Eyeing the weather report, he moved a bin of black umbrellas to the front of the store, just inside the door where they could easily be seen by customers needing a quick respite from the approaching rain. The year is 1861 and a weather forecast, first published in London’s daily newspaper The Times, had likely influenced the purchase of an umbrella or two.
Over a century later, American Jule Gregory Charney – who is considered the father of modern meteorology, teamed with his Norwegian and American counterparts in mathematics, meteorology and computer programming to develop the first computerized program derived for the prediction of weather. Their computerized approach was perhaps the first example of artificial intelligence (also known as machine learning) influencing consumer behavior through weather reporting.
With this predictive analytics, shopkeepers and advertisers could effectively move merchandise associated with changes in weather, from simple umbrellas to pharmaceuticals, clothing, vacations, and air conditioners.
Today, IBM’s The Weather Company provides actionable weather forecasts and analytics to advertisers with relevance to thousands of businesses, globally. Through the speed and agility of digital advertising, ad campaigns can flight and pause with the precision of changes in the weather… and as we know, the weather always changes. Ads for cold weather products can appear when local temperatures drop below 68 degrees, while ads for Caribbean vacations can target New York days before an approaching snowstorm.
In the last 20 years, artificial intelligence has flooded the advertising market by helping to scale operations through programmatic and content creation, emulating human conversation via chatbots and virtual personal assistants, and refining advertising platforms to understand consumer intent.
Just as our ability to forecast weather allows us to target advertising dollars, artificial intelligence is influencing more and more advertising decisions on our behalf. To this point, below is a brief history of advertising’s use of artificial intelligence and perhaps a glimpse of the future.
1998 – AI thinks you’ll like this book
The concept of clustering consumer behaviors to predict future behaviors began at Columbia University in a report on “digital bookshelves” by Jussi Karlgren, a Swedish computational linguist. And it was in 1998 that Amazon began using “collaborative filtering” enabling recommendations for millions of customers.
Today, Spotify recommends music you may like, Netflix suggests films and television programs you may like, and Facebook suggests friends you may know. This all comes from AI-based clustering and interpreting of consumer data paired with profile information and demographics. These AI-based systems continually adapt to your likes and dislikes and react with new recommendations tailored in real-time.
2013 – AI targets the labor of content creation
With the increasing popularity of content marketing, more content means more advertising opportunities. But the cost and pace of good journalism are considered too slow given volume of ads and eyeballs to be had. The solution: Yahoo’s Automated Insights Wordsmith Platform (now Verizon’s) uses artificial intelligence to scan billions of daily sports-related data points (scores, statistics) and structure the information in computer-generated articles summarizing games, informing fantasy sports fans, and reporting stats.
Articles are produced with speed and scale never possible by human journalists. The AI produces natural language content and adjusts for tone and personality, giving each piece a specific journalistic attitude. Automated Insights published 300 million pieces of content in 2013 and has far exceeded 1.5 billion annually since.
2014 – AI optimizes decision-making and reduces labor in advertising
Artificial Intelligence is making advertising easier, smarter, and more efficient. When programmatic ad buying was popularized in 2014, it introduced us to artificial intelligence-based ad buying, effectively removing the broken, laborious manual tasks of researching target markets, budgets, insertion orders, and layers of additional analytics tracking – not to mention high prices.
Through programmatic – a marketplace approach to buying and selling digital ads – the whole process is managed through intelligent tools that make decisions and recommendations based on the desired outcomes of the campaign. What was once used for ad remnants quickly and affordably became the new normal for digital publishers and some offline opportunities as well, with forecasts estimating over $33 billion spent via programmatic in U.S ad dollars.
2015 – A search result that understands user intent
Since the early 2000’s, artificial intelligence has been a compliment to search engines and their ability to provide a more logical search result. In 2015, Google introduced its latest artificial intelligence algorithm, RankBrain, which makes significant advances in interpreting search queries in new ways. Through RankBrain, Google has been successful in interpreting the intent behind a user’s search terms, making for a more relevant result.
If Google receives a search query for a term it is unfamiliar with or lacks proper context for, it can now leverage a mathematical database derived from written language that can pair the terms with related words that give it context. Through this artificial intelligence, Google can provide a more accurate result, pleasing both consumers and advertisers.
2016 – AI is listening, learning and responding
As Amazon Echo, Google Home and Apple charge forward with speech recognition on virtual assistant devices in our homes, a whole new opportunity beckons for advertisers. Just as ads ranked in Google’s AdWords, will advertisers bid to influence Alexa’s product recommendations? Amazon is currently in 15 million homes and is developing advertising opportunities for Clorox, Proctor & Gamble, and others to promote their products on Alexa.
Facebook Messenger, WhatsApp, and Slack began using AI to reduce the human labor involved in answering simple customer support questions – a cost center for any company of size. AI-powered chatbots respond to customer questions by chatting online under the auspices of customer support technicians and helpdesk prophets. These chatbots interpret the keywords in the users typed questions and form likely answers to questions.
What’s next for advertisers?
While it’s clear that AI is influencing the methods, targeting, and labor behind advertising, the future of AI may be in the ad itself. Case in point: in 2015, M&C Saatchi developed what is widely considered to be the world’s first AI-powered advertisement.
Built for a fictitious coffee company, Bahio, and debuting in central London, this virtual “poster” changes based on consumer reaction, with various text and image options that elicit different reactions. Given the consumer data collected by Google, Facebook and others, perhaps custom AI-oriented advertisements are the future of advertising with language, cadence, images and colors custom designed to appeal to the viewer. Could these custom ads be triggered by the user data in our cell phones?