How AI is revolutionising social media marketing
Updated: Feb 3
UCL FinTech Society Emma Prevot explains what social media marketing is, how it works, and investigates the role of Artificial Intelligence in digital advertising.
Social media has changed a lot during the past decade, from a “virtual place” for meeting new friends and connecting with old ones, to a vital and effective platform for advertising and promoting. What was presented as a free platform is now a marketplace. You may ask, what’s the product… Well, it’s us, the users.
Don’t worry, the scope of this piece is not to depict a dramatic picture of social media and convince you to delete them, there’s already a Netflix documentary “The Social Dilemma” which did it perfectly. This article is just to give you a brief idea of how advertising works on these platforms and the power of Artificial Intelligence behind it.
First, let’s start with a quick introduction on how/why advertising is used on social media.
How Advertising Is Used On Social Medias
Before social media, advertising was done via bulletins, roadside billboards, store windows, sides of buildings… Or television/radio commercials, website windows, newsletters… Nowadays, Social Media marketing is the second biggest market within Digital Advertising. In 2019, the worldwide revenue was US$89.5 billion, and it is expected to grow to US$138.4 billion in 2025.
What are the advantages of advertising on Social Media? Well, first advertisers can take the initiative, they search for “users”, rather than the other way round. They can take advantage of user demographic and choose the specific audience to target their ads appropriately, based on several parameters like age or gender and more complex ones like social behaviour or search preferences.
Nevertheless, we are all on 24/7 news overload, social media platforms are continuously trying to persuade us to stay on their platforms or to buy something, dozens of notifications overwhelm us as we wake up, emails, messages… With all this noise, in this extremely fast-paced world, it’s becoming more difficult to hold people’s attention. How can advertisers get their offerings to prospective customers? The answer is easy: by letting AI do the work.
Marketers are using AI-powered tools to persuade customers and gain their attention.
The Ai-powered Tools Behind Social Media Marketing
The big player behind the scenes is Artificial Intelligence. It dictates how the advertiser’s budget is being used, who will see the ads, how the ads will look like and finally the effectiveness of the campaign. The ability of AI to learn and improve without human involvement is crucial in new content creation: there are some tools able to write hyper-personalised ads for you, based on the feedback they previously received, to find the best times to communicate and optimise the targeted audience through clustering and pattern matching. As you may have noticed, we stopped seeing posts in chronological order on almost every social media platform; we now see what an AI algorithm thinks we’d like to see.
The Internet and social media have a large amount of data that can be used for digital advertising; studies show that users make around 95 million new and unique posts on photo-sharing sites daily. While humans are not able to process all those data, AI can take full advantage of this information. That is why the market of AI in social media is exponentially increasing: from US$0.6 billion in 2018, this market is predicted to hit US$2.2 billion by 2023.
Let’s see how some major companies are using AI-powered tools to tailor advertisements efficiently.
Google is collecting our information; we all know that. However, it’s not just what we do on our device, but we are willingly giving data from all our connected devices. Google’s AI algorithms use machine learning to gather insights on what and when we search, what we do after searching, what we buy or how much we spend…
On the other hand, the tech giant is providing an advertising tool for advertisers, streamlined by the Google marketing platform, called responsive search. This program asks companies to provide creative layouts for their products, which are then analysed by AI algorithms. Each layout is then fitted to a particular ad and a particular group of people. With all the amassing information gathered from users, AI is able to predict our online behaviour, what we like and what’s our budget. This means that multiple people, searching for the same thing, will get different results.
Video streaming sites like Netflix use those data to suggest what we should watch next and to keep us engaged. The site has also started to use a responsive search engine like Google. The algorithm decides which thumbnail photo to display based on our search/watch history to get users to click on that show. For example, if a user often watches shows with a female main character, Netflix will show a thumbnail with a female character.
Spotify and YouTube instead show ads during the content, “forcing” people to watch them, and use predictive analysis to make sure those ads are relevant.
Facebook and Instagram are using the AI-powered tool Phrasee do to what the company calls “AI-Powered Copywriting”: writing effective ads from scratch, based on the sentences that previously worked. In terms of data collection, Image and Face Recognition algorithms are used to scan specific features from pictures posted such as brand logos, colour palettes, if the user wears glasses or not and even the user’s mood and humor. Doing so, Facebook gains valuable insight into what users like, what is their aesthetic, what do they use daily and also about their emotional state, in order to offer very specific product recommendations.
Moreover, Facebook (including Messenger, Instagram and WhatsApp), is using a deep learning engine called DeepText, which can understand with near-human accuracy text in more than 20 languages, to filter text-based data and map any given text to a particular topic.
With 1 billion monthly users, TikTok has become the perfect place to advertise products for 13-40 years old people. The general aim of any social media platform is to keep its users engaged and hold their attention; TikTok is the best one in doing so. The app can engage its average new users up to 10 minutes (three times what Instagram is able to do); and the average time spent daily on the app by users is 52 mins (26 mins, 29 mins and 37 mins for Snapchat, Instagram, and Facebook, respectively). Trivially, advertisers could witness their brands blow up thanks to this platform.
The entire application is based on an AI algorithm creating the “For You” page; the user does not decide what to watch, it is the algorithm which dictates and creates every user’s page, making them all unique.
The renowned Chinese investor Connie Chan said that “TikTok is the first mainstream consumer app where artificial intelligence is the product”. The app heavily relies on the user's personal information (age, gender, location, search history, previous like etc.) to hyper-personalise the feed and insert some specific ads in between videos. Moreover, not only brands can advertise paying for TikTok ads, they can also work with Content Creators (influencers). The TikTok Creator Marketplace is a platform run by TikTok where brands can search for creators whose audience best fit the target market. The contact between companies and influencers is eased by the platform.
Twitter uses AI to determine tweet recommendations and tweet appearance. In fact, this platform is using Machine Learning to crop photos in order to display them in a much more appealing way. In 2017, the billionaire Mark Cuban said he purchased Twitter stock because “they finally got their act together with artificial intelligence.”
Are we truly aware of the amount of data collected from social media platforms? Maybe not, but despite undeniable privacy controversies, social media marketing is rapidly evolving, and is here to stay.