Using AI to learn what ads actually work

While outdoor advertising was one of the earliest, and stays one of the most obvious, methods of getting the word out about a product or service, it’s historically been extremely difficult to actually gauge how much people are paying attention. To understand how one company is using AI to learn which billboards are being read and which are being ignored, we spoke with Stephen Buko, founder and CEO of Kerb.

ABERMAN: Why is what you’re doing at Kerb such a big deal?

BUKO: What we do is so innovative, and so exciting, because outdoor advertising is kind of the one space that is really hard to get metrics on. And so you know, you advertise on Facebook because you have the number of clicks, you advertise on ESPN or CNN because you can tell the number of people that go to your website. In the physical space, there’s nothing like that. You can’t track the number of people that see your advertisements, or engage with your marketing campaign, or go up to your trade show, and so, that’s what we aim to do.

We aim to have artificial intelligence-enabled cameras, and actually count the number of impressions that outdoor advertisements actually see. It’s something that nobody else is doing, especially here in the Washington D.C. area. And so, we’ve been around for about a year or so, and we’re really excited to get rolling.

ABERMAN: Why do advertisers care so much about this issue? I think it’s worth mentioning because I’m not sure everybody listening really knows why it matters that much.

BUKO: Yes, I completely understand. If you have a 10 million dollar marketing budget, you’re going to spend 5 million on Facebook, and maybe 5 million in Times Square. But the number of metrics that you have for Facebook are very, very detailed. That’s why Facebook is so valuable. But now, the latter part of that equation, there’s no metrics. You can’t justify your five million dollar spend in Times Square. And so, that’s what we aim to do. We we aim to help chief marketing executives really justify their strategy, and then tweak their strategy.

Subscribe to the What’s Working in Washington podcast on iTunes.

ABERMAN: So if I’m an advertiser, ultimately I want to know a couple things. I want to know whether or not somebody’s seen my ad at all, and I also want to measure engagement. And I’m really interested, you mentioned artificial intelligence. It pops up in a lot of different contexts these days. How is your technology revolutionary? You are you using AI to actually let me know not just if somebody is passing my ad, which I assume you could just measure with the camera, but whether or not they’re actually paying attention to what they’re passing?

BUKO: It’s pretty exciting to be able to say that we have now have our artificial intelligence cameras on mobile billboards circling Manhattan today, actually. And what that means is that, as you’re driving down the street, you have this gigantic mobile billboard, pick your favorite company. We have cameras on both side of that billboard that capture the sidewalks, and we use artificial intelligence to look at the people on the sidewalk and check to see if they turn and face the advertising.

Basically, do they look at the mobile billboard? And what we do is, we partner with a mobile billboard company, and that company gives those metrics back to the advertisements that are that are on that billboard. And so, it’s a revolutionary way to apply computer vision and artificial intelligence to a real world environment, and then back into marketing

ABERMAN: It sounds to me like this is an example of something that I certainly am very interested in, which is how A.I. is applied not to substitute for human labor, but actually to make it more useful. Is that what’s going on with your company?

BUKO: Yes, yes, exactly. In a previous life, I was a management consultant at Deloitte, I worked at Lockheed Martin for years, and so, I was big into the future of work, and I understand it very, very, very well. And you know, artificial intelligence, that’s kind of its bread and butter. You know, replacing and automation of certain tasking. Whereas we’re expanding a category, we’re actually not contracting it.

ABERMAN: That’s my view too. I think that there’s a lot of confusion between what I would call specific intelligence, the ability to do a task really well, which A.I. is really good at that, and general intelligence, which is the ability to just do whatever is necessary without having all the facts. It seems to be, for a long time, that’s going to be where humans are going to continue to excel.

BUKO: Absolutely. Put on natural language processing in LP, and you mix that with some of the data we collect, you eventually will get to general intelligence. We’re about 10 to 12 years away from even the earliest version of general intelligence. But, it’s extremely exciting, especially with the advent of 5G coming here to the Washington D.C. area. We’re excited here at Kerb because we are part of the Verizon Wireless 5G startup program, and so we are test driving their technologies early, and we’re implementing that into our AI cameras.

ABERMAN: That’ll be just so wonderful for the Greater Washington Partnership, as example to hear, since they’ve been such a protagonist for 5G. Going back to something else, it strikes me, as our conversation continues, cameras, out in the field: I’m sure that you hear from people often, jeez, what about privacy? How do you react to that?

BUKO: I’m going to point you to the San Francisco local city council, who is now pushing the facial recognition bill through their local government, and it basically describes that you can’t do facial recognition and facial analysis without my consent. And so, we fully subscribe to that model. We fully subscribe to the European regulations of GDPR passed in 2016, and we’re building our business around that. So basically, we do facial detection, not facial recognition. There’s a big difference. I can understand that there’s a consumer, there’s facial detection, and some some high level attributes about that consumer. But facial recognition is an invasion of privacy, it takes my information without actually asking me for it. And so, we draw a very strong ethical line with that. And so, we’re fully compliant.

ABERMAN: So, the conclusion to be drawn is that technology in itself is neither intrusive or not intrusive, it’s ultimately whether or not companies like yours are willing to protect people’s privacy in how they apply it.

BUKO: Absolutely. Collecting anonymous data is what we strive to do. And unfortunately, there aren’t too many regulations out there that require that. We could do facial recognition, but we don’t. And that’s on purpose. And so, we can gather a lot of useful information for marketing executives and for companies, but we believe that certain regulations will come to the D.C. area, and come to the United States, similar to Germany, which have the strictest privacy regulations in the world. So, the ability for us to comply with that is something that we’re striving for, and we’re excited to get rolling.

ABERMAN: Well, I loved hearing about your company, Stephen. I think Kerb AI is another great example of why D.C. tech is unsung, but outperforming a lot of other places. Thanks for joining us.

BUKO: Glad to be here, thanks.

Copyright © 2019 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.