It wasn’t that long ago that Big Media — magazines and broadcast media — was the only way marketers could buy massive reach. Then came the internet, with all its disruptive power, shifting ad dollars from a handful of media companies to an infinite number of websites where users could be microtargeted at the very moment they were searching, buying, and reading.
The interview with
Tom Goodwin at Zenith Media
So-called programmatic buying made it possible to almost sidestep media altogether, reaching audiences in millisecond auctions at massive scale. Then came mobile, with the opportunity to capture users with the right offer, right content, right help, wherever they happened to be at just the right time. Although you’d think marketers would be happy with this profound power, in truth, clutter has only grown, and as it has, the continuing search for speed, scale and insight to capture the hearts, minds, and wallets of consumers through new technologies — from chatbots to voice activated devices, VR, and AR — continues.
That’s where Tom Goodwin plays. As EVP of Innovation for Zenith Media, the world’s largest media buying company, Goodwin knows what’s coming next in mobile innovation. Here Goodwin speaks with Marco Lafrentz, tyntec’s top CPaaS leader, to tease out the implications of next generation programmatic buying across digital mobile technologies.
Q: Let’s start off with what we mean by programmatic buying. How do you define it and what does it actually do?
Tom: It literally means we are using smart algorithms to buy and place media better. Unfortunately, for most people that’s become synonymous with remnant inventory and the deflationary pressure that pushes towards automatic placement of ads against things a lot of people find abhorrent. What makes it interesting is that we are starting to buy audiences rather than content. Through most of the history of media, you would buy against content, and content was a proxy for audience. If you wanted a rich a 27 year old guy that liked cars, you would find a TV show about cars and place the ad there because you know that’s what he’d be watching.
Programmatic allows us to buy against the audience wherever they are. The same person that watched the car show on TV now goes online and checks their email, and you can serve ads against them in different places. You can now reach people more cheaply, more effectively, and more efficiently.
Q: How does this affect publishers?
Tom: Terrible news for premium publishers. If you were making a yacht magazine in 1995, you could get 10,000 people a year to read your yacht magazine, and the net worth of those people and scarcity of opportunity to reach people in that context was incredible. That was a great business model, but now we’re in an audience buying landscape where we can find out when people have done a Google search for yachts and then next time they get on YouTube, just buy an ad against those people. It’s completely decimated the value of premium publishing, and to some extent niche publishing as well, because when you can buy audiences based on their behavior, niche content becomes far less valuable.
Q: But this is good news for telecom isn’t it? If you’re buying audience, it’s clearly better to buy media where people are in their mobile context, than specific media.
Marco: Totally right. Telco providers used to aggregate communication access and then sold single channels to enterprises to help them reach end consumers. Now we’re shifting to a space where the telco has already aggregated those channels and is providing the enterprise with capabilities to get the most value from them. Operators are opening up APIs and access platforms to create heat maps that analyze and pinpoint audiences. For example, there’s an operator in Turkey who still holds the subscriber relationship, so they know the identity of each and every subscriber. They’re not allowed to expose that, of course, but they can combine it with additional information. So, say you want to open a shop a barber shop in Istanbul. You can go to the operator and integrate with their services to look at what time men in Istanbul pass certain streets to find the perfect location for my barbershop. Combining their channel data with impartial audience intelligence and augmenting it from a service provider’s offering is an increasingly successful strategy for operators.
Q: Is that creating a new kind of opportunity for marketers to buy? Location-based buying? Location-based intelligence?
Tom: Yes, indeed. About four years ago we moved to buying audience instead of content. The next leap is to buy context, marrying data to describe what someone’s doing. Maybe the context is shopping in a shopping mall. Maybe it’s buying someone wandering around trying to get to the next meeting. The key is, it’s got many layers of data: search data, location data, demographic data. So we’ll soon be buying against context rather than audiences. And there may even be a step beyond that, where we’ll buy against intent, based on understanding what people have been looking for recently, on calendar information about what they’re planning to do that day. Instead of buying a demographic, an audience, or even a context, you’ll be buying the moment people need a taxi, or are stressed and need a holiday, or are fed up with their bank.
Q: Won’t that create big privacy and protection issues?
Marco: Yes it’s obviously a sensitive issue, and one reason reason why many enterprises silo marketing away from CRM. But there are some very smart ways we can design an application layer on top of the CRM to perform a hidden authentication or build smart data encryption algorithms that augment all the silos so the marketers get insight and context without hitting those data protection issues.
This is also where the REST API has been so successful. Doing authentication in-flight on each of those channels and then getting encrypted access to a specific set of user data is a very successful trend, as well as masking identities using temporary phone numbers in certain interactions.
Q: Tom how do you make customers feel like they’re being protected?
Tom: No one in advertising ever wants to talk about privacy. It’s very complicated. If you ask someone, “Do you want us to take your Google search behavior and serve you personalized recommendations?” everyone freaks out and says no. If you say, “Can we use your data to help make your searches more relevant?” Most people say yes. If you say, “Can we ensure that you are served nothing irrelevant?” everyone says, “Yes, absolutely, please do it now.” It all comes down to transparency and what I call value exchange — and different in different countries and demographics. For example, my understanding is that in Germany people are very worried about their privacy while in the UK and the US, people are less concerned. For me and most of the people that I talk to about this, it’s all about what you get in return for your personal data. I don’t think I even have a concept of privacy anymore. I spend a lot of time on Twitter and Instagram, and upload intimate posts to Facebook — the idea that I should somehow be shocked that a company knows everything about me is very strange.
Q: Is there a technological fix? Some people say AI and chatbots are becoming the new filter through which the optimization is taking place. What’s your take?
Marco: From our perspective, chatbots help create meaningful value for both enterprise and consumers. It used to be that you could send an SMS or get a voice code to get access to channels, but now we are moving away from that to maximize ROI by making smart decisions about whether even to send an SMS or make a voice call. Which sort of translates to “OK, do you want to transmit persistent information or do you think your communication should be of a more an intimate relationship?”
Bots take this decision making to a higher level. The bot is trying to figure out the context of the communication — mimicking what you’d experience in a person-to-person communication and providing hidden authentication of users at the same time. Both sides, users and brands, benefit from this abstraction of human behavior.
Q: Sounds great — unless I work for an agency. This takes place entirely between the brand and the consumer. What happens to the agency in this model?
Tom: I’m not sure we have anything to worry about. For one thing, chat bots don’t have much to really do with AI at the moment. It’s basically a logic tree right now: “Press 2 for this, press 3 for that.” But the idea of brands communicating in instant messaging is extremely exciting. You can do far better customer service that through instant messaging, but I doubt it will ever be a particularly good place for advertising.
If I see a beautiful video this afternoon in Snapchat that tells me about the latest Diesel jeans, then when I am on Facebook and I start talking to a friend about Diesel jeans, maybe a chat environment will open up where it will be possible for me to buy these jeans because my credit card details are stored on my phone and because I can just press a button with touch ID to buy them. That’s not the death of advertising. That just means advertising will create new demands fulfilled in another place.
Q: The opposite perspective is that AI and chat bots are a breakthrough in customer communications. Marco do you agree?
Marco: Definitely. Chat bots will soon be enhanced with real AI, so they are less like a tree and a greater resource for innovation. If you combine multi channel services so you have interaction with voice, interaction with text messaging and interaction with media all in real time. Think of an interactive video tutorial that allows you for example to demo a product with an interactive web real time environment — we are already seeing this for a lot of audio books and video games, where people are thrilled by the interaction with a data source.
The problem with advertising is that once you have seen it, the next time it’s highly repetitive. AI actually adapts to and can change the outcome. A video game may have a set plot, but depending on the mood or the context of my audience, there could be 15 or 20 different outcomes depending on the situation of the end user. The same thing applies with chat bots. Maybe you can currently recognize an angry consumer, but how do you react to that? If you have AI that actually learns to recognize that sentiment, potentially even with video or voice analysis, you can create powerful media that can boost the power of the experience of an advertisement.
Q: Is anybody actually playing with these tools in this way?
Tom: “Please can we have a chat bot,” is by far the most common request we get. But again, I think we are far away from being a time when the technology is good enough to do this. I probably spend about a hundred thousand dollars with Delta airlines and if they are going to start replying to my instant messages with a bot, that would be the end of me doing business with Delta. Until we get to the point when this software and the kind of technology behind it is so good that it effectively passes a bot Turing Test where I can’t tell that it’s not a human being, I won’t tell our clients to use it although of course we will do many experiments with it.
Q: But then on the other hand we do have these voice activated appliances that are moving from our pockets to our wrists and now to the living room, and it’s going to jump into the cellular (not wifi) environment. Marco do you agree?
Marco: Yes, and in fact, you could take it to the extreme and say you can mimic the complete experience of a person. But that’s not necessary to have sufficient customer engagement in services. One of the best examples is when you call say, American Airlines to re-book a flight, you no longer don’t need to re-authenticate yourself. This is why telecommunications is currently investing heavily on making the user experience go smoothly when you switch channels.
Q: Marco, do you think that this requirement for seamless communications between OTT capabilities and mobile data is changing the telecom business?
Marco: Traditionally, the core of our business model was aggregating access to communication channels, but this market became very mature and the margins got tighter as the competition grew. So our original job was to aggregate access to channels and optimize the cost to clients, and we’ve done that so well it is now in the “back end” of our business. Our business model is now turning into CPaaS (Communications Platform as a Service), where we are saying, “we’re not just aggregating channels, we’re providing valuable services on top — like automated authentication.
Q: Is that something that’s happening in the media space too — where channel aggregation is being forced to deliver innovation by the “race to zero”? Are your customers putting pressure on you to provide more value?
Tom: That’s a very good question. Certainly, we’ve had to bring in new people who have different skill sets and are conversant in areas like data analytics and optimization. We’ve also had to change most of our planning tools with new technology in mind. People don’t go online to do mobile commerce anymore, they buy and watch stuff, so we’re changing our processes to better reflect that environment.
Whole new areas like search and performance marketing we now need to integrate social and shared media and to some extent, PR. We are more likely to have relationships with e-commerce people or PR people and less with media. It hasn’t been easy, and I don’t think we’re there yet.
Marco: In the cloud communications world, we see platform business models taking over. Instead of offering discounts on voice access or sophisticated authentication methods or web clients, you see companies popping offering you a platform where you bring your own communication solution, and then the platform offers a variety of channels, application, and functionalities that you can pick and choose from.
The beauty of that is that for enterprise, we are able to bring many technologies together so they seem to be coming from one provider. It’s very attractive for them because they don’t have to commit to one provider, one channel, one feature, or one functionality. They can just pick and choose, and pay a little bit extra for a service based on actual use. It’s basically one-stop shopping, a pay as you go model that lets you potentially commit and to redistribute your investments in different ways so it’s an easy use case for you. Increased access to services is a really nice trend we are seeing in a lot of industries.