Episode 158 / Ricardo Rey / AB InBev / Head of Digital Consumer MAZ

Podcast: Investing in AI and A/B Testing for Better End Results

In his role as Head of Digital Consumer MAZ at AB InBev, Ricardo Rey works on connecting better with consumers through digital investments and developing digital strategies that are tested within real scenarios. This is why his Shiny New Object is AI in A/B testing, optimising the results from any marketing campaign and investing upfront to receive a better ROI down the line.

Ricardo has had a very varied career, including working in digital agencies on brand planning before moving into the media landscape and ultimately starting his own company based on digital anthropology. This involved understanding users’ behaviour and selling the findings to agencies and clients to help them build their own digital strategies. After working with the Discovery Channel in Brazil on their own digital transformation, Ricardo moved to AB InBev in Mexico, continuing his digital career.

It’s no wonder, then, that Ricardo’s best marketing tips are all about succeeding in digital, a constantly evolving landscape. He believes in building strong teams and accepting that no one person can be the expert on all things digital, but instead complementing one’s own skills with those of others around them.

One of these skills he focused on in our conversation is the art of investing upfront in developing a brand’s knowledge of their audience, then conducting extensive A/B testing with the help of artificial intelligence. Ricardo shared how AI enables him to automate a lot of processes, creating ads at scale to allow him to test multiple scenarios with real-life consumers, gathering data and feedback and ultimately delivering better ad campaigns for the client.

In A/B testing, he advises brands to accept that they’ll need to invest time to understand their audience first, to be able to use the right triggers and test the correct hypotheses. Secondly, brands need to be willing to invest in the experiments/hypotheses being developed to understand what works the best. In the long term, the outcome will be an optimisation of budget, time and results / ROI.

Listen to Ricardo share the challenges of using AI in A/B testing, as well as his top marketing tips and advice for students looking to join the industry, in the full episode here.

Transcript

The following gives you a good idea of what was said, but it’s not 100% accurate.

Tom Ollerton 0:00

Today's episode is brought to you by Media for All which was set up to help encourage more black, Asian and other ethnic talent into media and to provide a support and mentoring network to ensure talent flourishes in the media industry that we all love. If you're looking for a mentor, or would like to mentor young ethnic talent, check them out at mediaforall.org.uk and it is all 100% free.

Hello, and welcome to the Shiny New Object podcast. My name is Tom Ollerton. And this is a weekly show where I interview the leaders of this industry and get them to share their vision and their passion for what's gonna come next. And this week, I'm on a call with Ricardo Rey, Head of digital consumer MAZ at AB InBev. Ricardo, I try my best to say your name how your mom says it, but apologies to anyone who thinks I did that awfully. Could you just give the audience an overview of who you are? And what you do?

Ricardo Rey 1:09

Yeah, thank you, Tom. Don't worry about it. I know it's complicated in a foreign language and thank you for the invitation. Yeah, basically, I've been working with 3d for 13 to 15 years. I started working in digital agencies here in Latin America, basically working from the planning point of view, brand planning. So I worked with different types of clients: airlines, massive consumption companies. From there I went to work to the media landscape. I worked with a very big soccer brand here in Latin America basically in Colombia mostly and how to transform the offline presence into an online presence in the video landscape in in the media. After which, I got back to work for agencies that work for video in Colombia, for Century video, leading the digital content structure within the agency and was to build the entire team and how to convince brands to invest in digital content. Afterwards, I went to have my own company based on on digital anthropology we call it we use different tools, digital tools to understand users' behavior, it's in the retail context. And we sell sold those that information to different agencies and different clients so that they can build digital strategies. And afterwards, I worked for Discovery Channel in Brazil, leading the digital transformation in Brazil being the user manager over there. It's basically how to compete with giants like Netflix and Amazon Prime with a discovery content. And nowadays, I'm working here in Mexico with AB InBev leading the digital consumer strategy within the region in the middle Americas and also investing in new technologies to connect better with our data consumers.

Tom Ollerton 3:10

So if someone was listening to this podcast, thinking, I'd love to follow into Ricardo's footsteps. But they're like a student who haven't got into the industry yet. Or they're just kind of starting out early days, what advice would you give to them?

Ricardo Rey 3:26

Well, digital is a constantly evolving landscape. So the best thing that I did was to start young. Even when I was in the in the university, I went to work for agencies, even if they didn't pay me, I went to learn and to understand the different worlds within the digital landscape. And that gave me a very complete overview of the of the digital market. And afterwards, I got to choose where I wanted to focus my efforts because they get it there are no digital experts. It's so huge. And so for retail, that you have to choose one path. And starting early, even if I didn't get paid would be the best advice that I could give anyone who wants to start in the digital landscape.

Tom Ollerton 4:11

And as well as that, what would you other bit of advice would you give? What would you say your top marketing tip was? What is the thing that you've learned either when you first started out or later in your career that has stood you in good stead.

Ricardo Rey 4:23

Well basically, in line with the tip that I just said, is that you have to build very strong digital teams, because there's no single person that understands the digital landscape, and it's an expert in that way. So you need to complement yourself with other digital knowledge savvy in the different territories of the digital landscape. So I will strongly advise I would strongly advise to encourage you to build strong teams with different expectations, and that way you can be triumphant in the digital landscape.

Tom Ollerton 5:06

And how would you advise that people do that? So like if I'm an SEO expert, for example, and I see that I need someone to do PPC, but it's yeah, it's very close, but very, very different. That and I want to build a strong team. What do you look for when hiring someone or building a team when they've got a different specialism to you? And you know what it is, but you don't know how it works? How can you be sure you're hiring the right person?

Ricardo Rey 5:33

Yeah, basically, well, you have to understand first, what's the objective of the product that you're working on. And it's going to deliver different expertise that you need, if you're working on, if you're an SEO expert, I would strongly advise you to, to hang with a very strong development expert. Because SEO, it's not only content, but it's also the way you build the products, so that the engines get to know your content, right. So if you know about the theory about SEO, but you don't have a person who builds that specific product for your SEO strategy, maybe it's not going to be as successful. And also, if you get to think that SEO is how people are searching for different products or services, it would be good to understand or to bring someone in a team that understands the consumer through social listening or trends. That way you can understand what people are thinking, what people are talking about, and get better at SEO, and specifically, basically understanding the the chain that needs to be addressed so that they get into a digital product, and cover the different opportunities with people who are experts in that part of the chain.

Tom Ollerton 6:52

If you love all things innovation, and want to understand how brands plan to emerge stronger from the current situation, don't forget to check out Madfest London on the 11th and 12th of November. My good friend Dan at Madfest knows how to put on a cracking event. And there's always plenty of amazing speakers, beer and cool people to meet. Check it out at www.madfestlondon.com.

So we're kind of at the halfway stage, now, we're gonna move on to your shiny new object, which is AI in A/B testing, which, if you're not in this industry will probably won't make any sense whatsoever. And this is a topic incredibly close to my heart and kind of very close to what we do as a business anyway, so I'm really intrigued to hear how you do it. So no one's gonna know exactly what that means. So can you explain what AI in A/B testing is? And then we can move on?

Ricardo Rey 7:50

Yeah, basically, the, the uniqueness about data is that you can measure almost everything right, or you can get a sense of everything of every action that it's been taking on digital. So when we're talking about AI in A/B testing, it's basically I'm going to start with A/B testing, A/B testing is how you formulate different hypotheses, to launch a campaign or to address a specific problem. So for example, if you're trying to communicate with people between 18 and 24 years old, that visit specific sites, you can reach them with different formats, when I talk about formats, different sizes of banners, for example, or different pieces of content, or different types of creative. And when you get to analyze this matrix of these, these three variables that I'm talking about, you get to have millions of ways to communicate with these people. For example, I have a red banner, a purple banner, a green banner, and you get to measure which banner connects more with the audience that you're trying to address. So what we do is we build these hypotheses of how we can connect with this audience that will lead to impact. And that's when we use AI. Basically, we use artificial intelligence, to automatize process to build different creatives or to build different contents in real time to launch these campaigns and connect with these audiences. So that way, I don't need to hire a designer to build me 10 different banners in one week, putting an example. But I get to use platforms that can build a lot of creatives in one day with different rules with different templates with different copies, and they get to launch them with our control budget to understand which are getting a better ROI return on investment to my campaign, and then after which what the intelligence the artificial intelligence tells me is that there are some performing better than others, right? So I eliminate the creatives that are not working that well. And I invest more money in the creatives that are performing well. And I get to tune up themes and learnings from the creatives that are performing well. So I create new, new creative to impact this, the science says that way I optimize my budget, I optimize my time and my results and my return on investment.

Tom Ollerton 10:30

And what are the challenges you've had in trying to get this process working well?

Ricardo Rey 10:35

Well, the biggest challenge is that you need to formulate the right hypotheses, and the right strategies before using the A/B test and AI technology, if you don't know your audiences, and what are the possible triggers that you should activate, then the campaign will not be successful. So basically, we do other things before launching these campaigns. And we also use social listening to understand what consumers are talking about the trends, we learn from past experiences, to formulate these hypotheses. All of these hypotheses may not be 100% accurate. But the more time you invest in these hypotheses, and gathering information to build them, the more effective your campaign will be.

Tom Ollerton 11:24

And so for marketers listening to this thinking, yeah, I really want to get involved, what would you say the first steps that you should take to start building a strategy around AI in A/B testing?

Ricardo Rey 11:38

Well, first of all, I would advise them to change their mindset in two ways. First, you need to invest some time to understand which are the best creative, disruptive, performing addressed hypotheses that are performing, and you're willing to, the second thing is that you need to be willing to invest some money in in those experiments, to understand what's working or not. So when I have experienced as a consultant, also, and I work with some small brands, and I know that budget it's like, not, so the company don't have a lot of budget to invest, to when you tell them to invest a little bit of money, have a small budget to understand what works, what doesn't work, they reject that idea. But when people do that, they understand what's really working. And at the end of the day, the investment that you're making with us with a budget is going to be more efficient that if you don't do A/B test at the beginning. So that would be my advice to them. Invest time in understanding your audience, and invest budget in understanding your audience before you invest all of the budget that you have planned.

Tom Ollerton 12:51

So where do you see this space moving in the future? How do you think it's gonna play out in the next year, two years?

Ricardo Rey 13:00

I believe that the that in the Internet of Things, in the metaverse, for example, right now, I'm giving examples of banners and something that's not up to date. But what we're trying to do is, for example, how we connect these kinds of technology, with screens that in that are in out of home, and in betting technologies that help us impact a person that's buying a certain product within the store within a physical store, understanding their behavior. And so I believe that's where we're going. And not only in stores, and in retail, but also in connected devices. How you can address in, in an intelligent refrigerator, in your car, how you can reach them with the right message at the right time, by crossing different variables that you get to analyze and cross by gathering the data of the consumers.

Tom Ollerton 13:53

And so how do you see the feedback loop? Working with out of home? Because there isn't necessarily anything to click, I guess you could use QR code or a URL, but how are you solving getting enough data back because with banner ads, you've got billions of impressions and a lot of interactions. So you can work out what's the best ad based on a click through rate or, or whatever. But in terms of you might get 1000s of people seeing an ad 10s of 1000s people seeing an ad in an outdoor situation. But then how would you get the feedback loop from the from the outdoor?

Ricardo Rey 14:30

Well, you have to be very conscious that you're asking for very valuable information for for a human being right, so their purchase history, how they buy products, how they behaving out of home. So then, so that then that what you're returning to them needs to be huge needs to be valuable also so they give you information. For me to identify a behavior or a specific patient I know to have to have that the ID, for example, or the cell phone number, or their credit card in the best of the cases right, but people are very conscious of giving these details nowadays. So if my promise of giving something huge back, if you give me your cell number, or your phone or your address number, it's not good enough for the people, you're never gonna give me that data. And therefore, I'm never, I won't be able to cross your data with my databases. And therefore I will not be able to provide A/B test in out of home. So you need to be conscious that you're asking for very sensible data for people. And that the reward that they're getting from you as a brand needs to be very specific of what they need me to create value for them, so that they give you their data and the chain completes.

Tom Ollerton 15:56

So unfortunately, we're coming to the end of the podcast now. So if someone wanted to reach out to you and discuss the things that we'll be talking about today, how would you like them to do that?

Ricardo Rey 16:07

Basically, we trying to understand, which are the growth drivers for audience right now, right? Basically, we're trying to have a change in our mindset that we're not talking more about marketing funnels, but we're talking about growth loops. And there's a lot of documentation and theory out there. And there's a good platform called weforge, where you can find a lot of information about this. And it's how to identify the different triggers that make a user come back to my brand to get value. aside the product, it's what my brand can give them and identify which are the triggers so that he can no he or she can always come back to my brand. So I would like to to be addressed. At starting a conversation with growth loops. How would you identify growth loops for your business? For this information that you have from other countries? Because that way we get to identify the truth is we get to do A/ B testing, we get to use AI and we get to grow your business.

Tom Ollerton 17:12

That is the best answer we've ever had to that question. So thanks for finishing the podcast on such a high and Ricardo, thank you so much for your time.

Ricardo Rey 17:21

Thank you so much. Have a nice day.

Subscribe to the ‘Shiny New Object’ Podcast on Apple PodcastsSpotifyYouTube and Soundcloud.

Subscribe to our Newsletter

Watch ‘Advertisers Watching Ads’

Check out our Blog

Get in touch with Automated Creative