What You Need to Know About Artificial Intelligence in Marketing, Right Now!
Artificial Intelligence (AI) is a hot topic today and deservedly so as it’s going to be a massive revolution in both how work is done and the way we live. In fact, it’s often referred to as the Fourth Industrial Revolution, comparing it to revolutionary inventions like the steam engine and computers themselves. But in this article, I’m going to review just what we need to know as marketers – what are the AI applications that we’re already seeing in the marketing world and where is this likely to go. As you’ll see, artificial intelligence is already here and very soon will impact every aspect of marketing.
If you haven’t read Jeremy’s earlier blog post, “Machine Learning for the Modern Marketer“, you should as he provides a good explanation of what Machine Learning is, how it works, and why it’s so powerful. Machine Learning is the main type of AI being used in Marketing, especially Deep Learning (a specific type of Machine Learning which has been modeled on how our brains work) as it’s been the catalyst for most of the recent advances in AI in general, not just in marketing.
Current Marketing Applications
AI has been used in Digital Marketing for years now (at Vovia, we’ve been using bidding algorithms to improve the performance of our campaigns for at least seven years now), but more so at the enterprise level. But more recently, Google and Facebook have made big investments in AI (they’re both widely considered two of the leading AI companies) and have been rolling out a steady stream of Deep Learning powered features available to advertisers of all sizes. AI products and features are also being developed by just about everyone else in the business as cloud AI services have now been made available to the masses thanks to Amazon, Google, Microsoft, IBM, and others, so developers all over the world are leveraging those AI clouds to create a myriad of AI-powered features for their existing products as well as create a whole host of new products, so AI features are showing up in just about every bidding, targeting, and analytics solution there is.
Here’s a list of just some of the AI marketing applications that you can already use today:
- Smart Bidding – Bidding tools powered by Machine Learning algorithms have been around for a long time (we’ve been using tools like Marin’s “Big Algorithm” and MediaMath’s “Brain” for years now). Google has also been rolling out smart bidding features directly in Adwords, such as their Target CPA and Smart Display campaigns, and likewise, Facebook and other channels have introduced automated bidding features as well. These tools can take over bidding for segments of your campaigns and will often produce better results than humans do as they can take into consideration hundreds of factors, while the average person can only take into account 6 or 7 at most. So does that mean that we don’t need to do any bidding manually anymore? No, not yet. AI powered bidding tools need lots of data (lots of clicks and conversions per ad group) in order to make good decisions, so currently this only really works for larger campaigns. Think 50+ conversions/month at least per ad group. In the future, Smart Bidding tools may develop some sort of extrapolated intuition (the ability to make educated guesses based on smaller amounts of data), but at the moment they learn mostly based on trial and error, so they need lots of data in order to be effective.
- Smart Targeting – Many of the channels (Google, Facebook, and various DSP solutions) have been providing Affinity Audiences and In-Market Audiences for at least the past couple of years and 3rd party data providers (such as Bluekai and Audience Science) have been providing them for even longer. These AI powered targeting options look at people’s browsing habits to help us focus in on people who are interested in our product/service/topic of interest and can also distinguish between a prospective buyer and a student doing research for a school project. Google also recently launched Life Events targeting, allowing us to layer into our targeting key life events (such as graduating college, marriage, moving soon, or recently moved). AI targeting features are very effective, often lifting campaign performance by 10% or more, so they should be part of every PPC campaign.
- Creative Testing – For over a year now, Google has been recommending that advertisers should load up alternative creatives and let Google’s AI choose which one to serve to each potential customer. Their AI will decide based on not only the person who the ad is being served to but also the moment it’s being shown, in order to choose the ad which is likely to be reacted to the most positively. You can even provide a bunch of creative elements (headlines, descriptions, images, and logos) using Google’s Smart Display campaigns, and let their AI combine them and find the best combinations for you.
- Intelligent Analytics – There are a number of analytics and data platforms which will mine your data for insights and bring forward trends, anomalies, and insights (see Datorama, QuanticMind, etc). In some cases, they will also bring forward suggestions for improvement. Google has also added a feature called Analytics Intelligence to Google Analytics which does just this (it finds insights using AI) and also provides an easy to use interface for querying your data, providing answers to regular language questions, such as “where is my traffic coming from?” or “which channel converted the best?” It’s early days for AI driven Analytics solutions (they’re not yet as good at surfacing insights as we’d like), but when they work they’re a huge time saver.
- Personalization – Personalization is one of the main areas for AI marketing application with a wide range of solutions for website personalization, shopping suggestions (recommendation engines), email personalization, and ad personalization. One-to-one marketing is the holy grail for marketers but has been too difficult to implement at scale until recently. AI is now making personalization available to marketers on both the large and small scale in the future.
- CRO – Is standard A/B testing too slow for you? Then AI can help test a multitude of different ad optimizations (between 8 and 50 at one time) and come to a much quicker result. Most times, the optimal result will be obtained in a fraction of the time compared to a standard series of A/B tests, leaving you more time to take advantage of the increased performance of your ads and/or website.
- Social Listening & Management – AI has been a cornerstone of Social Listening products like Crimson Hexagon, Radian6, and Sysomos for years now, powering their sentiment analysis engines (which categorize posts as positive or negative, depending on the language used). AI is also now being used to identify influencers, find new audiences, better understand your audience, intelligently schedule your posts, and even assist with content creation.
- Content Development – There are AI tools which check your grammar (if you’re not using Grammerly yet, you should give it a try – their free version is fantastic!); enforce quality and content guidelines; research and suggest content topics, and optimize your content for keywords. There are even some which attempt to write content based on data you provide (e.g. it can write a description of a quarterly financial report) or write content based on topics your audience is discussing.
- Chat/Social/Email Bots – Bots aren’t just for chats anymore. There are bots for all text-based communication, including email and social. In fact, there are even outgoing email bots which will identify and approach potential customers for you (yes, cold calling bots!).
Artificial Intelligence Trends in Marketing
Innovation in the area of Artificial Intelligence is not slowing down at all. In fact, it’s picking up speed and is expected to continue to do so, due primarily to the decreasing cost of computing power (AI requires a lot of computing power, so as Moore’s Law keeps increasing the speed of computing power, it’s making AI more effective and affordable). We are also seeing large improvements in the understanding and synthesis of human speech, enabling AI to interact with us in a more natural way and opening up the possibility of new applications (check out the video of Google making a phone call appointment).
So where is this all going? Over the near-term, AI powered systems and features will continue to make astute marketers (those who make effective use of them) more successful, driving superior results for their campaigns. For example, over the years we’ve experienced an average lift of 20% in performance from smart bidding solutions. That’s a huge advantage! AI solutions will also continue to make us more efficient in what we do, automating routine tasks and doing so in a fraction of the time it’d take us to do them manually. By augmenting our intelligence, AI will make us smarter and better at what we do. Just think of how much smarter the Internet has made us by putting all of the world’s knowledge at our fingertips. AI will do this as well, but will go out and find answers to questions for us and sometimes to questions we haven’t even asked yet! By looking for trends, patterns, and anomalies, AI will bring things to our attention that would have otherwise taken us a lot of effort to find or which we wouldn’t have noticed at all. In the end, it’ll make us more effective and efficient marketers.
However, that’s not to say that all AI solutions work all of the time – we also need to understand both the strengths and weaknesses of each solution and use them appropriately. Of the many solutions we’ve used, none works in all situations nor even in most situations, so knowing when to use them is crucial. We also need to constantly monitor their performance and correct them when things go awry (AI marketing solutions can sometimes produce strange results when left to themselves and can be thrown off by changing variables such as goal changes or increased levels of competition). So don’t expect an AI robot to replace marketers anytime soon. Instead, we believe that Augmented Intelligence (AI making us smarter and better at what we do) is the more likely outcome.
No Fate But What We Make…
These are just some of the ways in which AI is already being used in marketing. If you’re using Artificial Intelligence, Machine Learning, Deep Learning, or any of the AI marketing applications listed above to boost the performance of your marketing, let us know. We’d love to hear your thoughts! Or reach out if you’d like to find out how AI can help improve the performance of your existing campaigns.