How Artificial Intelligence is Changing Marketing
In every sci-fi movie, the future looks like a megapolis filled with big, bold, holographic advertisements. You must have noticed giant 3D ads in “Blade Runner 2049” or “Ghost in the Shell” and the mind-bewildering personalized recommendations in ”Minority Report.” The movie industry does make it look fascinating, but what will the actual future of advertising look like? It seems that artificial intelligence (AI) just might be the answer.
The future of marketing we are talking about is actually not that far away. You might have heard about the bizarre incident involving Target back in 2012 when an infuriated father demanded to speak to the manager because they had sent his teenage daughter coupons for baby items – are they encouraging her to get pregnant? Turns out, the girl was pregnant, just didn’t tell her father. Target’s algorithm figured it out by simply analyzing the girl’s shopping behavior, demonstrating to everyone the budding use of AI in marketing. It’s no replicants and androids, but it’s definitely a start.
What is AI and What’s in It for Marketing
Artificial intelligence is the capability of a machine to imitate intelligent human behavior, like the ability to reason, discover meaning or learn from past experience. The AI that we have now is called artificial narrow intelligence (ANI), and it’s very far from the sentient robots we’re used to seeing in movies. ANI can perform a single task at a time, it isn’t self-aware and can’t think like a human.
Over time, more and more AI algorithms are escaping technological labs and computer rooms to spread across industries like healthcare, insurance, entertainment, banking, fintech and, of course, marketing. A great part of marketing is actually collecting and analyzing consumer data. So, for the industry, the adoption of AI means more personalized customer experiences, smarter search and recommendations, more targeted ads and, consequently, more conversions.
Both marketers and consumers will benefit from implementing AI algorithms in advertising. For marketers, AI may be the key to building strong personal B2C relations with each of their hundreds of customers individually, while consumers will be pleased with personalized advertisements tailored specifically for their shopping habits (and, according to a research by Adlucent, 71% of responders like that). It seems that using AI in marketing is a logical next step for the industry that’s so dependant on data, but what does it look like in practice?
AI in Marketing Today
Let’s look at the various ways artificial intelligence is being used to improve digital marketing.
If you’re on a tight budget, personalized advertising is a great AI marketing tactic for you. It allows you to show the right type of commercial to the right people, resulting in better conversions. Why waste ad views on people who aren’t interested in your products? Instead of hiring teenagers to hand out your leaflets to everyone in the street, you can now use smart algorithms to help you deliver personalized ads to potential customers. The data of the user visiting your website can be quickly gathered and analyzed in real time to select the banner that will most likely end in conversion.
As far as re-targeting is concerned, don’t you hate it when an ad of a product you accidentally viewed a month ago is following you everywhere for weeks? Well, it seems that AI will finally fix this issue by using very specific and more complete data sets.
User segmentation is something that has been used in marketing for a while now. The idea is to have an algorithm break your user base into different groups based on common characteristics for further analysis. You can select any characteristics you want, like location, browser, age, interests, operating system, etc. and even mix-and-match them. Say, you have different groups for young people who like hiking, senior people who like hiking and young and senior people who don’t hike. This makes marketing to them more precise and potentially more successful.
There is also cohort analysis that can be used in a similar way. Users are broken down into cohorts – related groups with common characteristics within a set time-span. Cohort analysis helps marketers see certain patterns of time depending on when the user was active. For example, it will show if users that purchased something during the Christmas sale ever returned to buy something else and if so, what was their future shopping behavior like.
Recommendations make wonders. If your friend recommends you a movie, you’ll go see it. If an online store recommends you to also buy a case when you’re buying a phone, you’re more likely to purchase them together. You might not have thought about buying a case yet, but hey, if the online store says it’s a good deal, why not? There are two kinds of recommendations: user-based that offer buying products that users like you have previously bought and item-based that offer buying products like the ones you’ve selected.
The reason recommendation systems are so popular is that they are very profitable. Take Amazon, with a third of their income coming from recommendations, or Netflix that shows three-fourths of movies based on recommendations. All of this thanks to machine learning.
Personalized search is the difference between “fork” in “Cutlery” and “Bicycles.” If you know that the user’s search and order history included spare parts for bikes, the algorithm may show products from the “Bicycle” category first. But if the user has previously bought a table and has searched through ten pages of tablecloths, the fork they’re looking for must be in “Cutlery.” Customized search makes the shopping process faster and easier, offering a feeling that the shop knows exactly what the customer needs, which means they’re more likely to return.
What do a psychic, a meteorologist and a data scientist have in common? Forecasting abilities. Amazon has been successfully collaborating with at least one of them to create their so-called “anticipatory shipping.” This predictive logistics tactic allows delivering products before users even ordered them. Sounds surreal, but it’s actually pure time series analysis that identifies if your orders have a seasonal character. Maybe some of them peak on holidays, celebrations, seasons, months, days of the week? Sales forecasting can analyze tons of raw data and consider many factors simultaneously to show you when and where the demand for certain products may rise so that you could amaze your customers with the delivery speed.
Let’s make something clear right away: no, the AI we have now can’t create content and won’t leave copywriters and designers out of work for at least a couple of decades. Artificial intelligence is constantly evolving, but the neuron networks we have now simply aren’t smart enough to generate anything creative. So, do not fear, fellow humans, ANI is as far from taking over the planet as Apple’s Siri is from Tony Stark’s Jarvis.
However, all the examples that we’ve pointed out in this article are already being actively implemented by business giants as well as little firms. Direct marketing, personalized advertising, chatbots and virtual assistants, marketing automation, user segmentation, sales forecasting – all of these tactics have been introduced or enhanced thanks to the implementation of AI in marketing. And it seems that, unlike in the movies, the future will bring less advertising, but the ads that will be shown are going to convey just the right messages to the right people.