Smart Talks with IBM: Online Privacy is More Important Than Ever

just look for the ones that's say smart Talks. And now let's jump right into our conversation about advertising and AI with Sherry Backstein. Sherry Backstein, Welcome to the podcast. Hi, great to be here. Let's start with your professional background.

How did you end up doing what you do now? Well, that's a really interesting story because I actually started my career in journalism, um and worked for one of the national networks doing the news and then got recruited by the Weather Channel TV way back in my career, and

that's really how I it all started from there. Um, you know, left the Weather Company after a while, went to the digital world as we saw changes in television coming, and then came back to the Weather Company about thirteen years ago on the digital side of the company and have been here ever since. So why are you passionate about the work you do? The work that we do at at the Weather Company. It makes a difference in

people's lives. It helps people make the right decisions for them personally, for their families, for their businesses as it relates to the weather. And what's interesting is and before I had the role I have today, and early on in my career, I was actually a storm chaser for the Weather Channel, and so I got to go out into the field and cover all these weather events that were happening, and it was fascinating work, um. And it was heart wrenching work as well, because you saw the

destruction and the power that weather has. And there wasn't a type of storm I didn't get to cover, between tornadoes and hurricanes, nor easter's and you know, it was really just it was fascinating work. And you know, it's it's interesting that when, um, you're in the middle of a storm and one day I was covering it was a F five tornado. The tornado was a mile wide, ripped through Norman, Oklahoma, and I went there the day after.

I couldn't believe the destruction. It was just it's something that unless you see it yourself, you just really can't believe it. And you know, I was talking to people that had impacted and these people would just come up and they would just hug you, and they would say, if it wasn't for you and the Weather Channel, we would not be here today. It's your alerts, your you know,

the information you provide is so critical. And so when you get to experience something like that and you get to touch another human being in that way and to help them, you become very passionate about your work. And I have to say everybody at the Weather Company feels this passion for what we do. That we are saving lives um and then and then doing something just as simple as helping people plan their day. So that's part

of the reason I'm so passionate about it. Well, to all of the weather reporters, storm Chaser's meteorologists out there, from from the bottom of my heart, I say a genuine thanks, well, we appreciate that, and you know, again we're happy to provide the service. So IBM bought the Weather Company in I'm curious about that. What's the relationship between the two companies. Well, it really came down to data.

So as the weather company, we have a tremendous amount of data from a forecasting perspective, from a weather perspective, and so it came down to really providing businesses with a weather strategy. So every business should of a weather strategy because it impacts really everything you do around your supply chain in most businesses. And so it really came down to being able to take our weather data and infuse it into all of our customers within IBM to

better help them make decisions around their business. And then of course as the Weather Channel, we have such a large consumer business as well, and so it is a really great touch point for IBM to reach consumers at such a mass scale. We have three and fifty million users every month that use our platforms. But it's just a great way for IBM then to to reach a consumer.

Um you know, from that perspective. So one of the reasons then you're saying is that, um, the weather base is data that can be used by IBM, but also does it go the other way or their overlaps in the world today between artificial intelligence that serve say the weather forecasting side. So we have before we actually became IBM, we used AI in our forecasting and we still today. But what we have seen is we've been able to accelerate the use of AI in other areas of our business.

For example, we are able to predict uh, you know, the flu in areas at risk associated with the flu um, you know, health risk associated with allergies by using Watson's AI. So that's a great example. And then most recently, we are leveraging AI on the advertising side of our business to create AI driven advertising solutions, not only for us as a publisher, but then these solutions that could be used by other publishers or marketers or others within the ecosystem.

So with the Weather Company being part of IBM, it's been very beneficial from not only an AI perspective, but we just put out a brand new Weather model earlier this year and it leverages the Power nine supercomputer of IBM, and so if we were not IBM, we wouldn't have been able to do that. So there has been you know, benefits on both sides of the companies and us joining together. So you brought up advertising, and obviously that's one of the main things that we wanted to talk about today,

So um to set the stage. We know that today's Internet and technology sphere is paid for or at least subsidized in huge part by advertising. And people argue about whether there could be a better model, but for better or worse, this is the one that's in play today. Give us a picture of what that landscape is like, how does the Internet and the technosphere make money through

advertising and how is that changing? So you're right, most content on the Internet and digital content is underwritten by advertising, and it's done that way so people can get it for free so they don't have to pay for it. And certainly we've seen an upcrease and subscription where people are paying for that content because they don't want to see advertising. So there's really those two choices. But for decades, advertising has really underwritten all that is content. And what's

important is advertising has changed through the years. Is kind of twofold one for the user. So you're on a platform, you're using an app, using a website, and the ads that are targeted to you. It's a better experience if they're relevant, if they're something that you might be interested in, and so targeting has improved over the years, and that's

become very very important in the industry. On the marketer side, that targeting is important because as a marketer, you only have so much budget to reach audience that are are going to be meaningful, audience are that are going to take the action that you want to take, and so being able to target just to that specific audience is really important from an efficiency perspective and you know the

best case for results for you as well. So that's really the advertising landscape UM as we know it today, UM, and it's changed through the years. You know, it started where it was really more direct sale is where you a sales person would go into a marketer or a brand and they would make this relationship and they would

sponsor certain segments on an app or a website. And then about eight to ten years ago, programmatic came on the scene which automated that process UM and so you didn't have to have as much face to face contact as a publisher, you know, to a marketer, but it can all be done through an automatic exchange. And so that's where we are today as far as the advertising industry is. But that is changing. And what's the role of what's known as third party data in all of this,

what is that and how does it work? So third party data, So most websites have first party data. So it's that relationship between the user and the brand. So in the case of the Weather channel, you know, you come to our platform, you type in a zip code, or you let us take your location, you give us permission, and so that's first party data. We know where you're interested in getting the weather, we know where weather is where you live, if you click on allergy or flu,

that's all first party data. Well, not every publisher has a lot of first party data or means to collect it, and so then you can leverage the actions that you do on one platform and that can travel with you to another platform. So then that first party data becomes third party data on someone else's platform. So when someone from maybe a news publisher comes to us, we know what they were interested in maybe on the news publishers site, what sections they went to, and then advertisers can target

them appropriately on our website. And so that's how the third party data has grown and it's become again valuable in order to provide that specific targeting to make your experience from advertising perspective relevant to who you are to where you are and what your interests might be. That was the case in the in the ad supported web we have today for it, content that really the gold standard is targeted advertising. I mean is untargeted advertising, uh

sort of going nowhere? Well, untargeted advertising isn't very desirable from anybody's perspective. I mean, I'm sure you have seen ads that just seem so obscure to maybe what your interests are, just you know, um, something that you would not be interested in all. And from a marketer perspective, you know, it's kind of like you're just throwing an ad out there and just hoping you reach someone that

might be interested. So it's a waste of money. And so really targeting advertising is beneficial for both the consumer and the marketer. You know. It's as a publisher, we talk to our users a lot, do a lot of surveys, and we've asked them what they prefer, if they prefer relevant ads or they just prefer they don't really care what ads they see, and they do tell us in most cases they prefer relevant ads, um things that they're interested in that they actually might want to purchase or

might want to find more information about. And so it does improve that user experience. But I guess, UM, one of the distinctions here is whether the data that's being used to target the user is somehow unique to them or something that people would feel is personal or private information. And one of those things I think might be mobile identifiers. Could you talk about the idea of mobile identifiers and

what role they play in UH targeted advertising today? Yeah, So the mobile identifiers essentially an I D that's associated with that device, and so then whatever you do on that device can be tracked through advertising, and there's permissions around that. So it's not that it's done in any kind of secret way because on your UM you know, in the case of I D f A, that's that's an iOS supported identifier, you have to limit, you know, put that AD tracking on, or you can say I

want to limit the AD tracking. So that user has always had the decision on the iOS device, UM, you know, to make that decision whether they want that tracking or

not to happen. But that identifiers just like your I D. So then when you visit maybe a different app or you go to a web browser on your phone, that I D would follow you UM, and so then your behaviors and things that you're interested in kind of build on that one I D. So then advertisers can target that I D. So this brings us to the relationship between targeted advertising and privacy. What is the conflict here

as people receive it? So it's an interesting question because as I mentioned, like with I d f A, users have always had the ability to turn it off. What's changing is that instead of that being opted in for a user, it's now going to be opted off. And so then the user has to explicitly say that they

want to share their information, which I personally agree with. UM, as you know, a pub lasher myself, our consumers privacy should be respected and they should be able to make that decision on what information they want to share and to be able to control that data. So I think it's critically important. So that's really what's changing on I d f A is the opted off versus what has

been opted in. And I think we all know when we get our cell phones, um, there's so much to go through when setting up a phone, UM, and there's so many, you know, different areas you can go in with an operating system, and most people just don't do it. So then what happens is you have articles that come out or media that comes out that starts scaring people that all of these things are happening and and and you're being tracked in in you know, in certain ways

that make you feel uncomfortable. And I'm not saying that there's not bad players out there, but for the most part, people are doing it to support all the free content UM that consumers are are using and the services in order to fund that, and so because of that, it's all about transparency. And so the transparency about what's being done UM is really improving, and I think that that's

really important. Yeah, I think we certainly take digital privacy concerns very seriously on the show, and personally, I tend to find that with targeted advertising, people don't seem to mind it when the mechanisms are clear to them, when it seems based on their consent. And I think that means uh, very importantly for people, their conscious consent meaning that they understand what they're agreeing to and not just signing, say,

a big agreement with lots of fine print. Uh. But what I think people often don't like is that the feeling that they have been observed without expecting to be observed or that they are being manipulated in a way that feels sneaky or hidden or tactically arcane. Would you basically agree with that? I would absolutely agree with that. I think private see really boils down to that value exchange between the consumer or the user and the content

or tech provider. And so you know, if you look at a privacy policy, they certainly can be daunting um and and rarely I think people look at them. So companies that are trying to make that more digestible UM and really helping to understand what you're doing with that data. That's really important and it's easy to do, and it's

it's the right thing to do, so people understand. Because the more fear we take out of it, the better off I think everyone will be, not only from you know, making a better user experience, but then that that content camera remain free because the fear is that you know, companies have to underwrite their content with advertising or else they'll probably turn to a subscription model UM and then

you know, consumers will have to pay for that. And companies that have both options and we certainly do, because there's some users that just don't want ads, they're they're intrusive to their experience. They have that option, which I think is great. How have companies like Apple and Google played a role in privacy versus targeting, well, both from

an iPhone and um android perspective. I mean, they own the operating system and so they are the ones in the case of like I d f A UM they're making the change that says, hey, we should have users opt into this and versus them opting out of it. And so that's how they're definitely evolving their privacy. They've always been very privacy conscious, UM at least at least

Apple from that perspective, has always been. But they're making it even more transparent now, and so I think that that is certainly important UM, and that people understand how their their data is being used, especially you know, with Google, because they're not only an operating system, but they are one of the major ad platforms, UM, you know, in

the world. This may be kind of a tangent, but I'm a little bit curious did this conflict between targeted advertising and people's perception of their you know, the limits of their privacy. Did this arise in as a surprise or has this been widely predicted back in the earlier

days of the Internet. That's a really good question. I think the changes that we see are really changes that big tech companies are making to be more transparent, and by that I mean Apple and Google because they are the ones that are now making those changes to impose these new rules that then publishers and marketers have to adhere to. And so that's really I think where most of this privacy changes are stemming from is a result

of that. But I do feel consumers are becoming more privacy conscious as time goes on, as more information as being shared, and you know, as they see, um, certain companies coming under attack, you know, from from different points of view, from a legislative part of view as well, and so there's I think more awareness now of what

actually is happening, and people are getting more educated on it. Um. You know, information and knowledge is always you know, as time goes on, is really important and the more people understand it, the more than they're taking more control over

their own personal privacy and making those decisions. So the way things stand today under the current model, publishers need ad targeting in order to pay the bills and target effectively while a lot of individual people want more privacy and control over their data, and I believe you're suggesting that AI could help accommodate both of these needs at

the same time. Can you explain that, Yes, So the changes that are are happening in our industry with the identifiers and with cookies doesn't mean we can't have targeted advertising. It just means that those identifiers are not going to be used in the future, so we have to find

new ways to target advertising. And at IBM what's in advertising, we believe AI is going to be the new backbone of the advertising industry, and that's because it doesn't rely on identifiers, and so it can be anonymous, and AI can look at a lot of different unstructured data and it can help be more instead of being just deterministic, it can be probabilistic, and it can help be predictive and take that data and make insights for marketers. And there's a lot of ways to do that. You can

use it by combining other data. So in our case, we can take our weather data and combine it with other shoppable data like from Nielsen to create look alike segments that then we can be able to target people that way without ever having to know anything about that user other than their behavior on our platforms, that relationship again that they have between us as a publisher and them using our platform, and then you can use it in social So social influencers are really hot right now

for marketers UM and the marketers are benefiting a lot from having influencers showcase their products. The challenge there with the brands is finding that right influencer. So you might know a handful of influencers, but with a I, you can look at all the influencers. You can look at their tone, you can look at the content that they supply, and you can decide which one of these influencers is

the best to represent my brand. That's going to be brand safe, because that's critically important as as you're fixing that picking that right influencer for your brand. So AI can comb through all of that mass amount of data which would take you know, humans days and weeks to be able to do, and it can really augment that human process to get a better result and a faster result, and then it can adapt over time and improve over time. So there's just two examples of how AI can help

brands better target um and better. Um, you know have targeted type advertising. So could you give a hypothetical example of how data about whether might be used to target

a specific add experience to a user. Sure, So we know from research that weather and mood have a connection, so people's moods and what they do when they feel a certain way, and so when we look at data around whether, and if you can match that with buying behavior data, then you can predict what somebody might be interested in purchasing based on the weather because maybe their their mood is indicative of them wanting to spend at

a certain time. So kind of a really simple example is when it's chili outside, we know that people are in the mood for soup for dinner. Right, super simple example, UM, But then there's more examples like that that maybe aren't so simple. So we have a partner that was, you know, a big retailer, and we did a lot of analysis on their purchasing data and what people were purchasing in

their store. We combine that with the weather data and we could see that when certain weather conditions existed, for example, when humidity was at a certain level in a certain region of the country, that strawberries sold so well compared to other weather times, and so what they did is help that retailer be more prepared from the supply chain perspective to say, when this weather condition is approaching, we should put strawberries out, you know, on the end cap

of the counter, because we know that they're going to sell and we have to make sure that we have enough supply. Since one simple example that wouldn't be as obvious as the cold weather and the soup example. And so by looking at those insights, you can use AI to gather all of these insights. Then a retailer can make sure that they have enough stock on hand, they can make sure they have enough workers on hand, or it can help them better promote products within their store

to get them to move off the shelf. Could you tell us a bit about what's in the ads. How does that fit into the role of AI in future advertising. So that's a really great advertising product that we've provided our marketers, and essentially what that is, it's it's a chat bought technology that leverages AI. So this allows a user on our platform to directly interact with the marketer on our site. And it can be many examples different

you know partners that we've had. But the chat bought would provide information to the user and the user could ask it questions and using AI, the AI anticipates the questions, provides the answer, and provides additional information. So it's a really great interactive marketing tool that can be used on our site. And then that marketer gets all of that information from that interaction with the user with their permission, that then they can use to provide insights and help

them to better market their products. And so it could be an insurance company that could be given quotes right on the side. It could be a company that's um telling you how to use their products in certain recipes based on what your likes are. So there's a lot of uses for it UM that we've seen between CpG brands and other retailers insurance um and it's worked really really well for them as a way to not just display their ad but to actually have that interaction with

the potential customer right there on our platform. Now, of course, today one of the most obvious natural world of what's going on is the pandemic. How is that played into these operations? When the pandemic started and everyone was looking for information around the risk of COVID in your area, whether it was the number of cases or the number of people hospitalized. We saw that that data was coming in at a state level well being the Weather Company

and people coming to our site. It's people need local information, right and and the same with COVID. While it was interesting to know what was happening in your state, people really just want to know what's happening around me or around where people that I love live. And so we set out to aggregate county level COVID data from multiple sources, and these were all approved sources, government sources, and the sources that were official sources. The problem with that data

it came in all forms. One site may have data in a PDF, one site may have it in a graph, someone else had it in a map. And so the power of Watson was able to take all of that data, no matter what structure it was in, pull it together and aggregate it so then we could put it on our platform and so on weather dot Com and a weather Channel app, we had a COVID section that provided this information down to a county level. So from that perspective, that was what we initially started and how we wanted

to use that data to provide that public service. As we started to continue to collect the data, our marketers were interested in also having that data to help them drive their business and to help manage their business, because what a marketer did not want to do was to market in areas where the pandemic was really really high.

They wanted to be sensitive to their messaging, uh, to you know, their potential customers and be sensitive to the pandemic going on, and so they didn't want to make that mistake and to you know, target advertising in places, um, you know, where they were really seeing this influx of the illness. And so we were able to take that data, similar to using weather data to help build triggers, so they could appropriately market in areas where they were starting

to see recovery. They could appropriately market, you know, depending on what their products were, two areas that maybe needed some of their products or services that they offered. So that was just another way of using AI to build a trigger for marketers, um you know, so what they were doing was relevant and what they were doing, you know,

was sensitive to the situation that was going on. Now that's that's that's very interesting because I think in anyone who's ever worked in newspapers or knows anything about the newspaper business of old, you know that you always have to be careful about where where an advertisement is popping up next two ways, say a problematic story and so forth.

And so this this is kind of like a geographic version of that to a certain extent, making sure that that that the advertising is is sensitive in the way it targets individuals. Uh, that's interesting. Absolutely, that's important. And then if you look at other illnesses, so we're coming upon flu season, and so we also then have data

around flu that's actually predictive. We were not able to get there with COVID uh just yet um, but with flu, because you have so much historical data, you can actually put together an algorithm using AI that actually can predict it. And when I talk about the risk of flu, there is a correlation between flu and weather, and so we can look at weather data points and see where flu trends based on this weather data points and tell people that there's a risk of the flu maybe in that area.

And because our forecast goes out fifteen days, we can predict out fifteen days, and so that's really benefits shoal, you know, in order for people to perhaps go get flu shots if that's you know, what they choose, or at least to have medications on hand that can help in case, you know, the flu you know, does hit their family um and they've be able to, uh, you know, at least be prepared. So, looking towards the future, what challenges do you see concerning digital privacy in the future.

I'm not sure I would say there's challenges with digital privacy. I think that brands, publishers like ourselves are going to become more and more transparent and think of privacy first. So privacy by design and consumer privacy is going to become even more important because it's really important as a publisher or brand to establish trust with their user and their consumer. So that part of it, I think is going to change, but I think it's a very positive change.

As a result of that, some of the tactics and some of the processes that we've had to use around advertising are just simply going to have to evolve. And fortunately there's great technology out there like AI that can help move us into a new era of advertising, which is really where we need to go. So I think this change is positive because it will force marketers and publishers to use these incredible new technology like AI to drive their business. And actually I find it to be

very exciting. The products that are coming out, and we've been building a suite of products for this, they're very exciting, and they can do so much more because they can be more probabilistic. It's more than just automation that programmatic provides. It's more insightful, it's more intelligent UM And so I think it's going to be as exciting time as the advertising industry adjust to the privacy changes. And do you think the trend is going to be towards more personal

lation or less. I think there's opportunities for equal to more personalization. I think we will go back to doing some of the things that we did years ago, especially around like contextual advertising as an example, UM to how to target. But I don't believe that again targeting is going to way. It's just a different way that we

will do that. And again, the most valuable user is one that a marketer can have a relationship with and they can serve relevant ads to, and so we have to keep solving for that problem as identifiers go away. But again there's there's technology to do that that's actually smarter and more privacy safe. So really I feel it's

going to be a win win for everyone. What are the opportunities you see in advertising today and especially looking toward the future, What what would be uh, you know, when in your wildest dreams the industry could go in whatever direction you chose, what would be the kinds of changes we would see come about, and what would be the kinds of opportunities that we would embrace. So the changes that we're on the path for that I really strongly believe have to continue is that consumer privacy and

consumers feeling safe with brands, So that's really important. So I feel that the changes that we're seeing with identifiers going away, it's not a crisis in the advertising industry. I see it as a tremendous opportunity for us to evolve and to really use some amazing technology that can help serve the same purpose of underwriting content so it remains free for users, but also providing marketers and brands a safe and trusted way to reach their potential customers.

So it's really about this you know, evolution as well. That's happening in the industry and we should embrace it. Sherry Backstein, thank you so much for joining us today. It's been great talking. Thank you. I've enjoyed it. Thanks again to Sherry Bastein for chatting with us today. Again.

If you want to hear more from this series, just look up the episodes of our show as well as episodes of tech Stuff labeled smart Talks, and if you want to learn more about the series itself, you can go to IBM dot com slash smart Talks to read more about IBM's work with AI and advertising. You can check out IBM dot com slash Watson Dash Advertising and if you'd like to check out more episodes of our show.

You can find Stuff to Blow your Mind wherever you find your podcasts, and we just asked that you rate, review and subscribe. Huge thanks as always to our excellent audio producer Seth Nicholas Johnson. If you would like to get in touch with us with feedback on this episode or any other, to suggest a topic for the future, or just to say hi, you can email us at contact at Stuff to Blow your Mind dot com. Stuff to Blow your Mind is production of I Heart Radio.

For more podcasts for my heart Radio, visit the I heart Radio app, Apple podcasts, or wherever you're listening to your favorite shows.