Artificial Intelligence (AI) is an enormous field -- both in terms of size and complexity. And though the technology to put AI to use is available, the implementation of these technologies has lagged somewhat.
Indeed, learning about AI, its many varied implementations, and its broader implications and use cases is enough to quickly overwhelm any marketer who doesn’t have a full understanding of the basics yet. Still, digital marketers need to be paying attention to AI now if they want to remain on the cutting edge.
Let’s start with a basic definition:
AI is a broad term that encompasses many different types of technology, but roughly defined it is the ability of machines to simulate human intelligence and decision-making.
Machine learning is what’s most often being referred to when digital marketers discuss AI. It describes the ability of algorithms and software to act based on learnings from previous data, as opposed to requiring the direct input of humans to influence behavior. Based on previous data and examples, these algorithms can make predictions around which to base future behaviors.
Currently, some key applications of AI for marketing and advertising use include business intelligence, programmatic advertising, and campaign optimization.
A word on why AI matters:
In a nutshell, AI has the power to improve both efficiency in humans, to return better performance on campaigns than humans are able to on their own, and to analyze data and predict outcomes in ways humans will never be able to. That doesn’t mean that AI is going to replace humans, only that it can help us do our jobs more efficiently and with better results.
AI allows marketers to provide better, more personal experiences to their users and, ultimately, to gain better visibility and insight into the trends hidden in large data sets that most have us collect, but are ill-prepared to properly analyze.
Ways to implement it:
There are a seemingly limitless number of ways that AI can be implemented to benefit almost any business’ marketing strategy. All of these implementations require time and investment upfront, though some are considerably more difficult to get off the ground.
Below are a few AI uses that are on the easier end of the implementation spectrum.
Chatbots are programs that use artificial intelligence to interact conversationally with humans via text or speech. They can be employed to serve many different purposes:
- augmenting or partially replacing human-based customer service
- helping users source products or content
- providing use case instructions and guidelines
- ordering a service or product (a ride from Uber, delivery from Pizza Hut)
- completing an ecommerce transaction (using Alexa to order products from Amazon)
Chatbots don’t have a lot of visibility among mainstream web users yet -- the technology is still new enough that many people have yet to interact with one, or do not realize exactly what they’re interacting with when they do encounter one.
An eMarketer survey estimates that as of early 2016, roughly 70% of users did not know what a chatbot was:
However, of those users (in the millennial set) who have interacted with chatbots, most indicate having had positive experiences:
2. Predictive Analytics
Most companies currently use analytics to measure the effectiveness of previous efforts and campaigns. Today’s analytics offer a powerful way to determine the ROI of campaigns that have already been run. They allow marketers to identify patterns over time, and to optimize future campaigns based on those patterns. Predictive analytics, on the other hand, employs machine learning to create algorithms that can analyze huge datasets and make predictions about future performance.
For example, predictive analytics can help companies find new customers by identifying people who are more likely to convert based on a set of relevant criteria that identify them as valuable prospects (eg; they have the same profile as customers who have converted before or a specific set of attributes associated with high-value customers) The business can use that information to proactively send marketing messages to those customers.
Predictive analytics can be used to score the leads a business already has, by identifying leads that have the highest potential to convert and helping to weed out the less qualified ones. The result is better visibility of high quality leads.
Another use case is in understanding the lifetime value of a company’s customers. Companies can use predictive analytics to identify customers who exhibit behaviors indicative of someone who is thinking about jumping ship and reach out to those customers to reduce churn. Companies can also identify customers who are well-positioned to be amenable to an upsell, and use that information to further increase the lifetime value of the customer.
3. Programmatic Advertising
Programmatic advertising automates the ad-buying process through the use of AI technology that can optimize bids and placements based on a number of relevant criteria -- audience behaviors, demographics, etc. Programmatic reduces the amount of manual labor required to manage campaigns and, through the use of machine learning, is able to optimize those campaigns in real-time.
Programmatic advertising is used to increase efficiency in the ad-buying process, to reduce campaign management on the advertiser side, and to optimize spend and performance across campaigns.
Real-time bidding (RTB):
ads are bought via auction in real-time. Advertisers specify categories where ads should appear, but are not able to specify the exact publisher site. There is lots of inventory available and advertisers have the opportunity to reach large audiences, but control over where exactly an ad will appear is somewhat limited.
some publishers reserve inventory that isn’t sold through an auction. Programmatic direct is a way of automating direct ad buys with those publishers. It offers more control (buyers know where their ads will appear) while still automating much of the process of ad-buying.
programmatic advertising offers advertisers the ability to determine what the most effective creative will be for a specific audience and serve a version of the creative that’s more customized for them. At its most basic, this could include images of recently viewed products. At a more sophisticated level it could include subbing in images and headlines that are more applicable to the user based on their known attributes.
Programmatic ad spending is experiencing a meteoric rise. From just under $12 billion in 2014 to nearly $38 billion predicted in 2018 which nets out to 82% of all digital display spending.
There’s a lot to learn when it comes to AI. The field has been expanding rapidly over the past few years and there are many changes and advancements to stay on top of. Marketers need to know the tremendous power AI has to augment and improve their initiatives. Staying on top of these changes now means success for your company or your clients in the future.