How to bring AI into your digital marketing strategy

bglen-2han – The sheer amount of information you need to personalize, optimize, and eventually automate digital marketing communications is becoming too massive. And yet, in that massiveness lies the granularity to understanding the variability that enables personalization for each consumer. Hiring more data scientists that can wrestle with the snowballing avalanche of data is an unsustainable arms race.

The key to unlocking it is AI.

You cannot wade through this unending pool of data unless you have the support of AI/ML tools. We interviewed David Edelman, Former Fortune 50 CMO, Harvard Business School Professor, Executive Advisor, and an AI marketing expert. Edelman shares insights on the most pressing questions for marketers who aim to bring AI into their digital marketing strategy, including:

  • How marketers can navigate the plethora of AI solutions available
  • How AI can help brands deliver an exceptional customer experience
  • Why AI can bridge the online-offline data gap to enable attribution

Q: How can brands navigate the plethora of AI solutions available?

David: “AI is now being used to bring fragmented channels together by creating a unified picture of the customer journey. Advanced algorithms can create an identity map that can link information about a customer across systems, even if they do not all have the same identifying ID. In doing so, AI provides marketers with support in four key areas.

  1. Analytics: Simply put, what is going on? AI analytics get a sense of where customer journeys are happening, and what these journeys involve, and then derive attribution. For example, it can help marketers understand what drives costs.
  2. Automation: There are now AI tools, such as those from a company called that can create 100 different variations of a particular email that you want to send, test these variations, and tag them. Large-scale marketers are starting to use these tools rather than wasting creative energy on doing this manually. That’s one example of automation – there are, of course, many more.
  3. Personalization: What is the best outreach and message for each customer? It’s a question of constantly testing, experimenting, and figuring out the right combination of creativity, timing, and channel. Based on these smaller tests, you can then take the learning and use it to personalize what you send to a broader group of people. There are so many variables that you can play with, but traditional A/B testing just does not get you there. A company called OfferFit, for example, has tools that allow you to completely set up all the test cells, figure out what personalization works for whom, and then send the right message to the right person at the right time. Ride the demand curve with the right incentive for each person, delivered in the best way for them to engage.
  4. Optimization: Over time, all these systems learn from what is happening on an ongoing basis in terms of customer behavior. Based on these learnings, they constantly update their models.

You also need to consider integration. None of these tools will work perfectly and do everything fresh out of the box. All these AI tools must integrate into your environment because they pull the data down, do something with that data, and then often feed data back into execution or a reporting tool.“

Q: How should brands use AI to manage the customer journey?

David: “Marketers can use AI to piece together vast amounts of data into customer journeys, and then make predictive judgments. An example of a provider who can do this is Pointillist. It provides customer journey analytics, using AI in two ways:

  1. Matchmaking: It filters through the channel databases that a company has and stitches together the journey for each customer. It timestamps the data and provides insights into the full longitude of a customer’s journey.
  2. Pattern and anomaly detection: It shows the most common trajectory of customer journeys and where variations from the journey occur. For example, a company experienced a spike in call-center calls. The customer journey analytics showed people calling had just used the Android version of the app within 15 minutes before calling and were trying to pay a bill. This showed a bug in real time and allowed them to make a change.

Systems like this, which are being used upstream and in customer journeys, can pull together all that information. Whether it shows you a problem you need to fix, or a customer to reach out to, these AI-based tools are allowing marketers to manage customer journeys at a micro and macro level.”

Q: How can it help brands to personalize post-cookie?

David: “Thanks to AI you can now build a better identity graph. By linking the different touches that you have with the customer, you can use that to not only identify somebody but have a richer sense of the journey they have been on.

If, for example, I’m sending a simple email piece, I should look at the background on the picture, the font, the color of the font, and the spacing of the words, to name a few variables. I should look at when the customer responds, what time of day they respond, whether they just click once or if they come back to it, how long, and whether they kept it up on the screen.

Personalization is all about variation. The more data you can create or collate, the more variation you’ll have and the more you can tag. You can then model your future interaction with the customer. The previous environment of basic cookies didn’t offer this capability and data. Now you can create more granular data and a much more powerful arsenal with which to personalize.”

Q: Why is it an important part of the marketing attribution formula?

David: “Marketers have been constantly challenged to find data links to close the loop and to connect (online and offline) channels. Loyalty cards and programs are often proposed as a solution, but that’s only relevant for a certain range of industries. And unless you give money away in the loyalty card to get people to use it, penetration is not necessarily going to be high.

The other thing to consider is the timeline for decision-making. Sometimes, especially for higher-end products, there are many steps involved. B2B products, buying a car, buying appliances because you’re redeveloping your kitchen. They all take time. So, for measurement and attribution, how long should that lag time be? What are other influences that come along the way? Over a longer timeline, it’s tricky to weigh different influences and get a longitudinal view of performance.

So, for all the digital development over the past twenty years, marketing still has smoke and mirrors to it. It is not always perfectly attributable. Some techniques can get you close, but you’re not going to be able to completely close the loop. AI can help you get dramatically closer.”

Q: What impact may AI have on marketing in the future?

David: “As I look ahead, the thing that excites me the most is innovation. Marketing is still about emotional connection. You’ve got to find a way to make that connection, and humans are going to need to be a part of the picture to figure out how we can best elicit emotion and come up with new ideas.

The companies that I have worked with that are using these tools are changing the way they think about marketing. Instead of just having standard campaigns where they incrementally run a simple A/B test and push one improvement, they can now throw out 20 new ideas every two weeks.

An amazing personal example occurred when I was buying solar panels for my house. I had communications from various providers, but one stood out from the rest. I received a physical piece of mail personalized to my house address.

It had a personalized URL that took me to a site that showed my house from Google Earth with Solar Panels super-imposed to show me what it would look like. They used AI to figure out how many panels could go on my roof and by the angle of my house and its tree cover, how much power those panels can generate. Using Zillow, they could get data on the square footage of my house and estimate how much energy I used. Using these real-time calculations, they could tell me how much I could save. It was utterly seamless and remarkably powerful.

30-second summary:

  • Marketers are faced with massive amounts of information. AI/ML tools are the key to unlocking insights from a snowballing avalanche of data.
  • We interview David Edelman, Harvard Business School Professor and AI marketing expert to understand how marketers can harness the power of AI.
  • Edelman advises on topics including navigating the growing landscape of AI tools, attribution, and post-cookie personalization.

Leave a Reply