Episode 3: Adding AI & ML to the Marketing Analytics Mix for Achieving Peak Performance

Shayla: Welcome back to another episode of our six-part marketing analytics podcast series Marketing Analytics Central – Conversations on Blazing Ahead with Data. My name is Shayla Wentz and I’m your host. In our last episode, Marketing in a Post-pandemic World, we talked about how embracing data-driven agile marketing helps organizations successfully navigate through this era of extreme disruption. How organizations can keep up with the ever-changing customer expectations, leveraging data, how the impact of disruption on marketing, and much more. Today, we are going to talk about how organizations can achieve peak performance in a post-pandemic world by adding AI and machine learning to the marketing analytics mix. We’re going to touch on the use of artificial intelligence and machine learning in marketing and how they can help optimize marketing budgets and investment, as well as provide and improve ad hoc analysis for better results. We will also talk about the importance of customer experience tracking and the role of AI to enhance the overall experience that brands can deliver.

Once again, I’m joined by industry expert David Edelman. David is the former CMO of Aetna, a global executive advisor for digital and marketing transformation, and a member of Grazitti’s advisory board. Thank you so much for joining us again today, David.

David: Uh, it’s great to be here Shayla.

Shayla: So first question. AI and machine learning for marketing have extensively changed the marketing landscape, creating a massive emphasis on personalization, behavioral targeting, micro-targeting, and several other parameters. The implementation of machine learning algorithms have also produced amazing results in the marketing space. So how do you think AI and ML are contributing in scaling businesses and is there a specific strategy to implementing them for maximum return?

David: I think we’re just at the beginning of how machine learning is going to influence many, many different stages of marketing. And there are different use cases where it’s going to make a difference. And I think just answering your last part first, is there a specific strategy? I think you have to really think through the use cases where AI can help and prioritize those. Now, how do you know which are the use cases where AI can help? It’s where there’s some kind of either decision that needs to be made, where there’s a lot of data and a lot of possibilities and you’re looking to make a decision, a decision on who to target, a decision on how to allocate money. That’s one side, the other side is informing you of new things going on in the market that can influence a decision. So looking for patterns that are disruptive. So for example, using AI to analyze all of your media spend against your plan to spot things that are maybe not going necessarily right. And many companies have incredibly complicated media strategies. For example, you know, thousands of keywords, tons of different targets they’re doing in social media, they have offline stuff.

Well, what if rates for certain keywords is starting to skyrocket? What if certain click-throughs are falling down in a very specific spot? AI can surface that and then AI can help you create scenarios to rebalance it. For example, there’s one company called Elsy – E L S Y. That is using AI to change media planning and management into something more like a wall street hedge fund, where every single day in real-time, you can see what’s going on with your media spend, spot things that are against plan and then model scenarios for how to reallocate your money.

That’s one area. Another area is using machine learning to optimize your testing. So we’ve talked about agile testing and how to think about the best way to motivate customers to take an action. Well, you may know which customers you want to take, what action, but you have no idea how to get them to do it. So there’s so many different variables that you can play with. You can play with creative, the offer, the media that you use the timing, and it’s not as simple AB test. It’s a multivariate test that has to constantly be refreshed. And this is the kind of capability that AI can bring. And there’s companies out in the market, for example, one is called Offer Fit that brings AI to bear in terms of helping customers construct and manage large Al multi-variate testing. So I think there’s a number of different ways for spotting things, reading the market, and making decisions where AI is going to make a difference. And the key is to think through in your business where those use cases are and prioritize them.

Shayla: Excellent! Now we know that AI uses larger rays of data to identify valuable patterns, essentially customer behavior tracking. With most organizations nowadays having rather large tech stacks, how can they integrate insights from AI across multiple tools in a way that costs them less time and money?

David: Yeah, one of the challenges with AI is pulling together data from many different sources of record. And those sources may not necessarily have common schema for how the data comes together and being able to look across those platforms is actually the value of AI, if you can get that data together. So there’s a number of different things going on in the data world. One of the biggest is the creation of CDPs customer data platforms, which are essentially places where you can aggregate customer data from a lot of different places. Um, you snowflake is one. Salesforce is building its own. There’s many different ones out there, um, where you bring together lots of different data based on creating some kind of identifier, which AI itself is using to figure out whether that identifier applies to different pieces of data, so it can stitch together your identity and pull that all into one schema. You’re seeing that also in the media world, I mentioned ELSY earlier where they’re using AI to pull together all the different data from many different media tracking tools into one media data lake, essentially, that is normalized, so that you can use, um, tools to analyze that at scale.

So the ability to use AI is hampered. If you aren’t able to bring that together, and most companies have so many legacy systems. And have had a lot of trouble bringing that data together. There are new tools now that are making that dramatically easier. They’re actually using AI itself to bring the data together and then applying AI on that data to make insights. And I think that’s something as companies start looking into this space, they can be much more open-minded about because there are possibilities. The art of the possible is opening up more than it had been in the past.

Shayla: Let’s move on to machine learning now. How does machine learning help marketing teams to create more meaningful customer interactions while keeping the analytics at the center of it all?

David: This is the key to all of this is using all of the data that you have about somebody to personalize an interaction, not just based on who they are, but on the context, based on what you think will motivate them. So for example, I mentioned the company earlier called Offer Fit that allows you to test many, many different variables, such as creative incentives and such, um, to figure out the best way of using it. And so, for example, um, one company that uses their, their tool is Brinks – the home security company and renewals are always a major challenge in their business, and they have used AI to figure out for all different kinds of populations, a whole different range of combinations, and they’ve come up with new ideas that they didn’t realize they could test before using the tool to essentially ride the demand curve and come up with just the right messages for each individual. Without, for example, giving 50% off to everybody!

So AI can help understand the patterns of what other people have responded to what’s appropriate for someone given their context and inform it. And essentially it’s moving from AB testing, which is very simple to AI testing, which is multi-variate and constantly, proactively trying different things, seeing what works and then using that to optimize how you contact people and AI tools now are enabling that to happen at scale. So that is really changing the nature of the concept of personalization and optimization. Frankly also reducing the need for having all kinds of data scientists in-house, but you first have to get your data cleaned up, get the processes in place and figure out what is going to be the way you operate those use cases.

Shayla: Excellent. Yeah, I think this is an area that we’re going to see a lot more growth and transformation over the next few years, for sure. Well, thank you again for joining us today, David. That’s the end of my questions for this podcast. Um, please be sure to join us next time for Enhancing Data Visualization for Better Health Care and Business Operations.