For years, we as marketers and business owners have been working with the data we know we’ve got: newsletter conversions, in-store traffic, online ad clickthroughs, and online user tracking data. This data is often delivered in a nice, neat dashboard where you can search for patterns in your customer’s behavior and, maybe as a result, adjust your next promotion according to your findings.
It’s certainly descriptive, allowing us to extrapolate future success based on past customer behavior. But it doesn’t actually predict the impact of a change in your marketing. That’s where predictive analytics with the help of AI comes in.
Let’s start from the beginning. What’s predictive analytics?
Let’s look at an example of predictive analytics in marketing. Say you run an auto dealership in Indianapolis. You already know from your marketing data analytics that a chunk of your customers are male race car enthusiasts and have responded well to in-store giveaways of race tickets to increase dealership foot traffic. You might use predictive analysis to determine whether lead gen will increase even more by spending 20% against advertising during the Indy 500. You can assume that marketing to this segment of male race fans will predict that based on past behavior from other promotions, your foot traffic will increase in-line with that result.
That’s predictive analytics “ it’s incredibly helpful to marketers but, like anything, it does have its limits. The more detailed and hyper-targeted your aims, the more niche your personalized campaigns become until your ideal target audience dwindles to a number not worth putting your resources behind. Marketers are required to make decisions that weigh significantly upon this type of data.
Got it. So, what’s AI?
Now AI, or artificial intelligence, is where it gets really creepy and really cool. There is an application of AI called machine learning “ which means that a machine can learn on its own without being programmed to do so by a human. Its system can automatically learn and then improve from experience. This changes our whole world as we know it!
How can AI help me?
AI machine learning can test and retest that auto dealership’s customer data to predict every potential customer-lead gen match at a speed and brilliance that no human being could possibly accomplish.
Now, your auto dealership can utilize machine learning to identify which of those race fans have families and would, therefore, want a less snazzy and more practical crossover SUV. It also determines that many female customers are equally likely to respond to advertising around race season and you may, therefore, want to widen your targeting. It will further determine that while you’re based in Indianapolis, the system will acknowledge other race-loving fans as potential customers in Daytona, which would prompt other offers as those fans descend upon the Indy500.
Another example of how AI can work for you: On the Growth Lab Show podcast, we interviewed Dennis Mortensen, CEO and Co-Founder of x.ai. His long-term vision is to “kill the inbox,” which led to the creation of Amy and Andrew, AI assistants who schedule meetings. When we start talking to our computers, we humanize them. And why wouldn’t we? They’re participating in dialogues with us. So, when Dennis and his team were building an automated machine, they built Amy and Andrew.
Let’s say Joe is a contractor. He’s always on the go and doesn’t really spend much time at his desk, so when he’s at a job and it takes him a few hours or even a full day to respond to a consultation request, that leaves a chunk of empty, nebulous time for a potential customer to ask someone else. It’s normal, but it’s not optimal. So, by using an AI agent like Amy or Andrew to book his consults, Joe is now removed from the conversation. It’s not Joe’s job anymore “ it’s the job of a machine.
I see the importance. What are the baby steps to help me get started?
If your business requires top-notch customer service, the low hanging fruit is what we can showily call conversational AI but essentially is a good old chatbot or digital assistant.
Chatbots can interact with your customers in a conversational way. You’re familiar with these on Amtrak when you need to change your train time, Whole Foods to search for recipes while you’re shopping at the store, or Lyft to request a ride. It almost feels”¦dare I say”¦human? Plus, because they’re available all day and all night, you can focus your customer service on more complicated matters.
As a marketer, accurate audience segmentation is your bread and butter, and you’re simply not going to grow your business or nurture your current customers without putting any effort into this space. Traditionally, you’re able to segment your customer base with demographic and geographic characteristics. (If you’re still with me on that car dealership example, you may find this helpful.)
Now with AI, your marketing data analysis can get far more granular. Now you can organize your customers based on their interests (race car driving, for instance, or golf, cooking, travel, etc.), past behavior (like a tendency to drop out of a course or be late to an appointment), or a purchase pattern (the likelihood that a customer puts items in their cart but never checks out or purchases an item but usually returns). This kind of segmentation can even help you identify customers who are likely going to discontinue your services, prompting you to initiate a marketing campaign to keep them in the fold.
Paid Search & Display Ads
Optimizing paid search ads is one of the biggest positives of jumping on the AI bandwagon. If you’re spending the money anyway, you might as well make it as effective as possible. AI can discover which type of ad is performing best, the optimum time of day to run it, and the most effective call to action. Then, it will automate the bidding process so that your programmatic ad is making the best possible use of your budget.
At LOCALiQ and the USA TODAY NETWORK, we created a tool to optimize ads using AI machine learning. Our platform is capable of determining which images, colors, and other design aspects will be most impactful for audiences “ leveraging intelligence derived throughout the full scope of its local news sites. By predicting how the creative of each ad will perform “ ranked on a scale between “very poor” to “very well” “ brands can roll out authentic yet informed campaigns with the USA TODAY NETWORK at large. The best practices are applied to all new ad creatives at the time they’re being designed.
Recently, advertisers using the tool have boosted click-rates between 10 and 33%, and 80% of A/B tests netted higher performance when following the tools guidance over original designs.
Bring AI Home
We’ve discussed a lot of high-level concepts here, but if you boil it down, you can begin work with AI like you do anything else “ a little bit at a time. Try something simple like a chatbot or programmatic ads, experiment, review your marketing analytics, and if you like it, scale it up!
And, if you’re looking for help incorporating AI into your marketing strategy, we’re here to help. Give us a shout today.