Paul Roetzer is the Cleveland-based founder of the inbound marketing agency PR 20/20 and the creator of the Marketing Artificial Intelligence Institute and the Marketing Score blog. Roetzer also authored The Marketing Agency Blueprint and The Marketing Performance Blueprint.
Recently, Roetzer joined Act-On’s CMO Michelle Huff on the Rethink Marketing Podcast to talk about artificial intelligence and its role in the future of marketing.
This transcript has been edited for length. To get the full measure, listen to the podcast.
Michelle Huff: Thanks, Paul, for joining us. Maybe you can tell our listening audience a little bit more about yourself, about PR2020, and about the Marketing AI Institute.
Paul Roetzer: Thanks for having me. I started PR2020 in 2005. Blogging and social and things like that weren’t really big when I started it. And the majority of what we’ve done through the years is content marketing, really lead gen nurturing and conversion for B2B companies. A lot of tech, lot of manufacturing, insurance, professional services.
But then around 2011 I developed a fascination with artificial intelligence. That to me became an, oh my gosh, what if that technology was applied to marketing? And I didn’t know at the time what AI was, really, or if this was even possible. But it started this journey for me of discovery. And it led to late last year, November of ‘16, we created the Marketing Artificial Intelligence Institute with the mission to identify the current and future potential of AI and provide that education to marketers so they could look for ways to transform their own marketing, but also their careers.
What do you think is the first thing marketers should know about AI?
Michelle: That’s awesome. … It’s just a huge topic. What do you think is the first thing marketers should know about AI?
Paul: I think the first step for a lot of people is just to understand the basic terminology and what it actually is. Because you hear AI, machine learning, deep learning, natural language processing, natural language generation, image recognition. There are all these terms. And for me even 12 months ago it was just this mashup of words.
So I always start by explaining that AI is the umbrella term. Machines on their own, they don’t know anything. They don’t know a table from a chair. They don’t know how to learn and get better at a task. They’re trained to do this using data and different types of processes to do the training. And so AI is that. It’s this big picture idea of enabling machines to get smart.
And then underneath that are categories like machine learning, which is the most common one you hear. So the key is to not be overwhelmed by the terminology or even the idea of it. Despite the dystopian views that are out there and the things you see in Hollywood, at the end of the day AI is really there now and for the foreseeable future to enhance what you do as a marketer. And the sooner you embrace that and seek ways to have it help what you’re doing and make things more efficient and more personalized, you’re actually going to get ahead of everyone else.
What are things that people might not even think of as Artificial Intelligence that really is?
Michelle: In many regards there’s a lot of artificial intelligence that’s actually being used today and we might not even know about it. What are things that people might not even think of as AI that really is?
Paul: Yeah, it’s an important point. Because my general guidance to people is: Your life is already machine assisted, and your marketing will be, too. And you just won’t know it. And so, as you were saying, you probably as just a general consumer or a person living on this earth will interact with AI dozens if not hundreds of times every day.
If you watch Netflix, Netflix has massive AI investments. Google is an AI-first company. On your Gmail app on your phone when you to go reply to something, if you look at the bottom you’ll see recommended responses, usually like two to five words. Those are called smart replies. There’s AI all over that. If you’re lucky enough to drive a Tesla, Tesla autopilot is enabled by AI using deep learning. So yeah, it’s literally everywhere.
And I think for marketers, they’re going to start seeing that, like the platforms like in Act-On, where you’re using them anyway, they’re just going to start getting smarter and they’re going to start introducing little features into them that make your life easier. And you may never actually go looking for an AI tool to do send-time optimization. It’s just all of a sudden going do it. And you’re going think it’s like magic. In reality it’s AI.
Can you explain what are the five P’s of Artificial Intelligence?
Michelle: Exactly. … So, a lot us in marketing, we all know about the five P’s. Can you explain what the five P’s of AI are and maybe share some examples for each of them?
Paul: We really struggled to understand how to categorize the different technologies that were out there. Over time we started to see patterns developing where we could start to more logically categorize these so they could make sense to everybody. And so we ended up settling on planning, production, personalization, promotion, and performance.
Now each of those categories, some of them are very immature, so the technologies aren’t very far along yet. But I’ll walk through some examples of each so they make a little more sense.
At the planning level, if you look at something like search engine optimization, keyword selection, topic clustering ‒ that tends to be a very human-driven process. That’s something that a machine in the near future should be doing for most marketers.
Production we look at as the curation and creation of content. So specifically in 2015 we started looking at can we use AI’s natural language generation, being the kind of AI we were looking at, to write blog posts, because we do a lot of blog post writing for clients. So over time we realized that you have to create the templates and train them the different branching logic. But once you do that, you can tell a data-driven story at scale hundreds or thousands of times instantaneously.
Personalization is where we’ve seen most of the money going. Things that right now a human has to set rules for, the machine can absolutely do that better than a human if it has enough data to do it. So you’re going to see a lot of personalization over the next 12-to-24 months. That’s where most of the use cases for marketers will emerge.
Then you get into promotion. That one is also ripe to be disrupted. Not a ton of great tools in that space yet, but more developing. An example of that would be Albert, which does digital media buying. You just give it the budget and the creative, and it runs all the infinite variations, and makes all the changes itself based on performance data.
And then the last one would be performance. And that we mainly look at as taking analytics data, and finding insights out of it, and then figuring out what to do next. That space is also extremely immature.
Those five Ps then enable us to look at all these different AI-powered tools. And we’re tracking over 500 of them.
How to get get started with AI and Machine Learning?
Michelle: Is there an area that you would recommend that someone should get started with AI? Is it personalization? Or how do you typically recommend people think about it?
Paul: There are two general recommendations I have for getting started. The first is to pick a single-use case. And so by that I mean take a look at your existing marketing structure, your average monthly spend, and where your time goes, and look and see if any of those are really data-driven and really time-intensive, that once you understand what AI’s capable of you could say, well, that would be logical that an AI tool might exist to do that, and go do a search for a tool for that.
The other is to go talk to your core martech stack. So if you have a marketing automation platform, email marketing platform, whatever it may be, go talk to them and say: What are you guys working on? Are there any more intelligently automated features that you’re either beta testing or that are coming up that we could experiment with to start better comprehending what’s possible? And I would actually maybe even reverse those and start there.
Do you think there’s a world for both marketers and AI?
Michelle: Do you think there’s a world for both marketers and AI?
Paul: I think in the near term, which I would look at that three-to-five-year range, more than anything AI is going to enhance the knowledge and capabilities of marketers. And the ones who take the initiative to understand it, embrace it, and apply it, they will have a competitive advantage over their peers. It basically gives you superpowers in certain areas. That’s the reality of what most people will experience.
Will it replace jobs? Yes. It’ll transform the industry within the next decade. Which ones? I don’t know. Any great technological advance in the history of society has done that. It takes jobs, but it also gives jobs that you can’t guess would exist. And I think that’s what marketing will see. I think the industry will continue to grow, lots of opportunities will continue to exist for marketers to evolve. But they’re going have to embrace the opportunity evolve. If marketers just sit back and pretend like AI isn’t going to have this impact, then those are the people that would be in trouble.
Michelle: How can we learn more about you, PR2020, and the Marketing AI Institute?
Paul: Our website is just pr2020.com. That’s the agency. And then we do have the separate site which is marketingaiinstitute.com. As of today it’s just a content hub. We try and publish two to three times a week. We do a lot of interviews. And we’re really just trying to connect marketers to the resources right now and see where that site goes.
What are your big marketing predictions for 2018?
Michelle: We have one bonus question. Do you have your big marketing predictions for 2018 lined up?
Paul: Do I? Yeah. Oh man. It’s so funny. I oftentimes try to avoid these prognosticating things. I think there’s just going be a mass of acquisitions, honestly. I believe a lot of these marketing platforms, the major players out there, they have no choice but to infuse artificial intelligence. And most of them aren’t staffed to do it.
And I think most of the bigger platforms are very aggressively looking for opportunities to infuse intelligence into what they’re doing, and whether they build that or buy it. It makes it really hard for marketers to go find 10, 15 tools to do each of these unique things. Because there’s a pretty good chance if a company does one of these narrow tasks really well using AI, they’re going to get bought.
Michelle: Well, thanks so much for coming on the show. It was really interesting talking to you.
Paul: Well, thanks for having me.