Dan McCarthy | Professor - University of MD | The Unfairness & Disparate Impact of Privacy Policy

0:00The CMO Confidential Podcast is a proud member of the I Hear Everything Podcast Network. Looking to launch or scale your podcast? I Hear Everything delivers podcast production, growth, and monetization solutions that transform your words into profit. Ready to give your brand a voice? Then visit IHearEverything.com. Welcome to CMO Confidential, the podcast that takes you inside the drama, decisions, and choices that go with being the head of marketing. Hosted by five-time CMO Mike Linton. Typeface helps the world's biggest brands move from business brief to fully personalized campaigns in hours, not months, with its agentic AI marketing platform.

0:48They are the first enterprise platform with agentic AI marketing workflows designed to instantly automate work that used to take weeks. With Typeface, one campaign scales into thousands of personalized experiences across ads, email, and video while staying true to your brand. The company's AI-native platform integrates seamlessly into your MarTech stack and marketing workflows and includes enterprise-grade security. Adweek named Typeface AI Company of the Year. Time Magazine featured them as a best invention, and Fast Company called them the next big thing in tech. See how major brands like ASICS and Microsoft are transforming marketing with Typeface.

1:37Learn more at typeface.ai slash CMO. Welcome marketers, advertisers, and those who love them to Chief Marketing Officer Confidential. CMO Confidential is a program that takes you inside the drama, the decisions, and the politics that go with being the head of marketing at any company in what is one of the most scrutinized jobs in the executive suite. I'm Mike Linton, the former chief marketing officer of Best Buy, eBay, Farmers Insurance, and Ancestry.com, and I'm here today with my guest, Dr. Dan McCarthy. Today's topic, the unfairness and disparate impact of privacy policy. Now, Dan is a professor at the University of Maryland, which the business school is in his background right now if you're watching us on video.

2:29Previously, he was an assistant marketing professor at Emory University, where he created what may be the first customer lifetime value course. His specialty is the application of statistical methodology to contemporary marketing problems. He has been one of our most popular guests in spite of all the titles we have put with him over the past few years of the show. So it's great to have him back to discuss his latest research. Welcome, Dan. It's great to be back. And yeah, it's a fun topic. So yeah, excited to dive into it. It's a lot more fun than it sounds. And let's discuss this research, which was titled Evaluating the Impact of Privacy Regulation on E-Commerce Firms, Evidence from Apple's App Tracking Transparency.

3:22Now, I know this is a lot more exciting than it sounds, and I can probably hook you up with some marketing creative folks for your next research title. But for now, let's just tell our listeners what this is all about. Yeah, so the basic story is that, as a lot of you may be familiar, a lot of people have not been happy with how they feel they've been tracked. And they kind of talk to their friends. Next thing they know, they see a Google ad about that thing, even though they're like, I never, you know, I never even went on the web about that thing. How are they collecting this data on me? and what Apple had done with one of their previous iOS updates was this they instituted this app tracking transparency policy and all that means is you download a new app you know the Nike Run Club app and you open it up and the first question that it asks is do you want this app to track you across you know across devices etc and 80-90 % of people they they just they have no idea really what exactly that means but they say well this kind of sounds like something that's bad i don't want that to happen to me you know so they ask ask app not to track and and all that means to to a company like nike is that before if you're kind of scrolling through your facebook feed you see that new pair of you know nike vapor flies you click through on the link well now suddenly you're not on facebook anymore and if the person then made a purchase at the nike website previously nike would have known that the facebook ad is what caused that conversion to occur but now we don't have that identifier anymore so you might make that purchase but now nike won't know that it came from the facebook ad so that's really the big thing they call it this idfa you it's just kind of it's basically a customer id but the customer id was shared um you kind of across all the different domains that people would go to.

5:29And so companies like Facebook, you know, they kind of rely on these ads for, you know, for their ad revenue. You know, they kind of took a big hit because now they can't really, you know, say definitively this ad caused this or that to occur. So it's like a big source of data that effectively went away for them. And when did this get put in? And I'm thinking back, like if I look for like, you know, if I'm cooking and I need a new pan, I look in there and I look for the pan and then suddenly I get like 10 ,000 ads for pans or pots. Tell me, like, when did this come in? And then, you know, what was it like?

6:14How did the ads get served beforehand? Good question. So this policy went into place in kind of late April 2021. Yep. But interestingly, you know, like some of the other iOS updates, they kind of slow roll it at the beginning. Yes, they don't kind of push people aggressively to do the update. But it was kind of in June where they started to really nudge people to update to the new iOS. I think it was 14.5. And so it was really at that point that it went into effect. So, yeah, so before there was this little kind of meta pixel, you know, that was kind of put on Nike's website. And so Nike, you know, Adidas, Puma, you know, just kind of all of these different advertisers.

7:03And that's really what allowed Meta to know, well, this person came from a Facebook ad, you know, and that kind of was all that info about the purchases that happened on Nike's website. It was all being then sent back to Meta so that Meta could say, well, this customer made this purchase at Nike on this day. And I know that they also, that same person using that same identifier, did all of this other stuff on these other websites on these other days. So it's really kind of this unified identifier that Meta had that allowed it to basically serve up more effective ads for all of the people who wanted to advertise on Meta.

7:45So this has blown up attribution. And this is only iOS, right? Did everyone else have to do this too, or it's just iOS? Just iOS. And that kind of is one of the other big things. So you have, you know, everyone's heard of GDPR. You've heard of CCPA. Why don't you tell everybody what those mean just because everybody might not know? Yeah, those are basically other kind of government instituted privacy initiatives. You know, so, you know, in Europe, you know, they institute GDPR. Now all these companies have to effectively do other things to be more privacy compliant. You know, Apple's ATT is kind of an example of what they call self-regulation.

8:29Yeah. where it's kind of the companies themselves that are instituting these privacy policies. And in some sense, it's a really good setting for an academic paper like ours, because Apple, they had full control and kind of it applied to everybody all at the same time, you know, and so it made for a pretty clean analysis. But because it was only Apple, they did it on the iOS device, but it's not like, you know, Android devices would have been subject to the same thing. Right. Or desktop devices. It really was just the iOS devices that were directly affected. So I think that there are kind of spillovers that were happening.

9:11Like a lot of people, they get a lot of info about their audience, you know, who they should be targeting. Like, what is my best audience from that sort of metadata? And so when that went away, it would also reduce the effectiveness of, you know, your Android spending and your desktop spending. because suddenly you just have a less good view of how good your base is. But yeah, the direct effects were strictly with iOS. So we're going to go to the direct effects in a minute. Just a reminder to our listeners, ATT is not the phone company. It is app tracking transparency. So when we shorthand it and talk about ATT, it is about app tracking transparency.

9:54And you kind of talked about why you wanted to research this. But there's an economic and consumer and marketing and financial story behind this research, right? You were like, hey, how's this going to affect everybody? I think that's it. Here is a word from our sponsor, Typeface.

10:18To orchestrate workflows across channels, meet ArcGents, your AI teammates handling complexity so you can focus on what matters. All with a reimagined workspace where you and AI create as one. Spaces works the way you do. Create anything. From documents to videos to entire campaigns. At scale. Our agents help you craft campaign storyboards powered by your brand hub. Giving you the perfect starting point. Transform one asset into countless variations for every audience and market.

10:59Join industry leaders already reimagining their content lifecycle with Typeface. Welcome to Marketing's Next Chapter, where every story finds its voice.

11:12Now back to our discussion with University of Maryland professor Dan McCarthy. Yeah, and you know, kind of this speaks to some of our previous interviews. I'm all about like the big company effects. Yeah, I think that that's it's an area that I just I'm really excited about. And you look at a lot of the other literature and and it's just not about like what effect did this have on companies that are advertising at the company level? Did it hurt their revenue? Did it help their revenue? And so that was really the focus that, you know, that we had kind of brought to this paper. It was to say, well, it has these effects on consumers.

11:50It might have other effects, but what companies really care about is they did this privacy thing. How did it hurt my revenue? That's where the pedal hits the metal. Well, this is the thing. One of the reasons I wanted to have you on the show talking about this is you look at this and you think, oh, it's just a privacy thing. You don't think about the financial impact or the targeting impact or your ability to read results impact on it. you think oh it's just one more thing i have to do and what you're saying is this is actually blowing up a lot of the attribution in the funnel if if i'm reading it right yes it means you're the amount that you spend on your marketing you're going to get less revenue for that spending and that means you're you know for the same amount of marketing budget your revenue is going to go down you know so so let's talk about you said they slow rolled it let's talk about how it rolled out and then any specifics you saw in what happened.

12:56And then we'll talk about the disparate impact on businesses. What we found was that the click-through rate effectively got 37 % worse after ATT went into effect. So you spend your money on those Facebook ads, 37 % fewer people are clicking through you know so that would kind of translate into you know that many fewer purchases you know and also this will make your cac go up your your customer acquisition cost explode right for the same amount of money uh you're getting a lot less click through so by definition your cac is rising right yep and so then the big question is all right imagine that you're one of these companies, you say, oh, shoot, you know.

13:49We're getting so worse at marketing right now. Like our CAC has declined by a huge amount. And gosh, this stuff is so much more expensive to buy. You were just going into that. I didn't mean to jump the question on you, but you were going to explain like that. It's kind of the other thing that the firm might be thinking then is no one's putting a gun to this company's head to say you have to spend this much amount of money on facebook ads you know if if the effectiveness goes down a lot you would think they'd reallocate a bunch of budget away from facebook to other channels right and i think that's much easier to do if you're a nike you know because you have all these channels you've got a very well built out and diverse marketing function but if you're a direct to consumer brand a lot of them they only invest in like Facebook and Google.

14:43You know, that's kind of like the main. And so effectively what we found was there was some amount of kind of reallocation of marketing budget away from Facebook, but it wasn't very much. So it's maybe just to take a step back. There's kind of like two data sets that we had. Yeah, let's go over the data sets and then talk about Facebook, all the different ways you could do this. And then why direct to consumer companies used Facebook so much? So, yeah. So the two data sets, one is the advertising data set. And it was kind of this I'll show you mine if you show me yours data co-op type model. Yeah.

15:25So companies they would supply. This is how much we spent on all of our different channels over time. And if you provide that data to this company, then the company will also provide you aggregated insights about all of the other companies that they're tracking. So they have thousands of companies in this data set. And so we can directly observe from the companies themselves. They spent this much on Facebook, this much on Google, this much on TikTok, et cetera. And we had that before until after ATT. So we can just empirically observe, like, how did the budget allocation change? And then we had this second data set that gave us really nice insight into the revenue impacts.

16:09And it was this company called Grips Intelligence. I don't know if you've heard of them before, but Superdata, again, thousands of websites, thousands of companies that they track. and for each of those companies, they can see how much revenue, how many clicks, how many visits to the site. They can also see the source of purchases. So did these purchases come from Facebook? Did they come from other sources? And they can see the device that was used to make the purchase. So if someone was on their iOS device and they made a purchase on that device, then grips would directly observe that in their data.

16:52Yeah. Effectively through grips, we could, we could see all the data that companies would make available in their Google analytics pages. So this is how you get to the minus 37, right? Exactly. Yeah. Okay. And then, then, so tell me why the direct to consumer smaller, I think it's mostly smaller companies to lean so heavily on Facebook and not other channels. It was easy to use and it was relatively quite effective. It was also a pretty deep channel. So you could put a whole bunch of money in it and it wouldn't get capacity constrained. So, yeah, so a lot of companies that had kind of great success, especially these direct to consumer companies like, you know, Warby Parker back in the early days and all those different Shopify businesses that sell this wide range of doodads.

17:46They would all pretty heavily rely on Facebook ads. And there was really good tracking of the effectiveness of that channel. Yeah. So they could know I put in this amount of money. This is how much revenue I got back for that spending. And so they could very easily see, you know, we seem to be doing pretty well in this channel. So, yeah, so it's a channel that they they leaned on very, very heavily for for demand generation. And so I so I think this is super interesting because suddenly you are driving all this marketing. You have a really clear view of your acquisition, a really clear ROI look.

18:26and suddenly this privacy thing rolls out actually probably slowly over time. So it's not like you hit a CAC cliff. You probably are, your CAC is decreasing as the adoption goes. And then, then boom, suddenly if the average is 37, I'm sure there's a bunch of people that just completely fell off the cliff. And, oh, go ahead. Yeah, the people where it completely fell off the cliff were, well, So there's the effectiveness of that form of advertising. And then there's like the downstream consequence for revenue. The downstream consequence for revenue was like super bad for the smaller companies. They were the ones who took almost all of the hit.

19:08So one thing that we had done was we said, imagine that we kind of rank sorted all the companies that we're tracking. And then just get like that middle 50 % customer. the ones bigger are the big companies the ones that are smaller are the smaller companies you know so just kind of split it up by the median in that way the big companies actually were unaffected by att because i can move i can move all my money to anywhere and i i have practice with all those other channels as well but if i'm a little company i only have practice with one channel, I'm kind of, this is why you think it's really unfair, right?

19:49Yeah. So for the small companies, so for the overall average, the impact on revenue was again, about 37%. Although it depends on exactly how you do the measurement. But for the smaller companies, it was like north of 60%. Yeah. So they really got, they got thwacked. So yeah, so this was, very tough for them. And are they, are the little companies recovering or did this put just a lot of them out of business? I think a bunch of them likely went out of business. Now I would say that you hear 60 plus percent revenue decline. I should kind of clarify it. It's like a relative revenue decline. Right.

20:32Within the channel, right? It's, it's only the part you can see. well so it's actually overall revenue for these companies but it's kind of like relative to other companies that did not have exposure to the facebook ads so what we effectively found was um if you were primarily investing like google and these other marketing channels your revenue kind of continued to kind of move up the companies that invested very heavily they were very very reliant on the facebook ads the revenue had been going up before att and then it kind of went flat And so it wasn't like the revenue fell by 60%. That would be catastrophic.

21:10That would be catastrophic. Hey, I have a question. So when stuff like this gets done, how come your research is not done in advance? Or how come Apple or the government or whoever's putting this in doesn't think about this kind of research done in advance to see what would happen? Why is it after the fact? Well, from a researcher standpoint, now that the dust has settled, we've seen all these, we've seen multiple years kind of post implementation of ATT. That gives us like really good data to be able to say, oof, this had this effect. And we can kind of do all these different analyses. Imagine that it was, you know, imagine that it was kind of like April 1st, 2021, like right before they were planning to roll it out.

22:01we wouldn't have had any data to directly observe to be able to see the effects. So I'd say now that we have a paper like this that's out there, one could imagine if another privacy initiative was being contemplated, they should certainly look to papers like ours for guidance. You know, like, wow, this might hurt small firms, you know? let's let's um let's talk about what marketers around the world should be thinking about this particularly if there was smaller companies or or facebook heavy you got uh also ai creating a lot of zero click search uh so maybe some of your google google search is is declining as well or maybe more expensive how what advice do you have for uh the brands looking down the barrel at gosh this attribution is getting harder and harder.

22:59Yeah, I'd say, you know, for one, you want to make sure each of these, you know, kind of key quantities that I talk about a lot, CAC, you know, customer lifetime value, you know, you name it, they're all channel specific, and they're all evolving over time, you know, they're changing. And, you know, I'd say that this kind of further reiterates the importance of first having channel specific metrics, So it's not just how much am I spending on marketing? What is my overall blended CAC? But you really want to break it down channel by channel. And then the second big thing is, as we've seen with this example, your CAC today is going to be different from your CAC a year from now.

23:42And so you really want to be paying very close attention to how these figures are tracking over time. The same goes for customer lifetime value. You know, so I'd say, you know, one of the other implications of these privacy measures is that you're not going to be as effective at getting exactly who you might have wanted to acquire, you know. And so it could mean now that you kind of have more of a spray and pray type of acquisition approach. You might bring in a different type of customer than you did before. And so it really pays to be tracking post-acquisition value and CAC in a channel-specific way over time.

24:22I hear in your comment there also, if you are averaging CAC across all your channels and you're averaging customers across all your channels, you will be lost in this world. I used to have saying averaging is killing us, and I think that's what you just said here. Yeah, you're leaving so much money on the table. You wouldn't be able to see, oh, this channel is doing really well, and this other one seems to be doing terribly. I think that's a sort of insight. It all gets glossed over into the average. And so, yes, again, Theta, my company, that's kind of what we do day to day. But even though this was not a CLV paper, it really kind of highlighted the importance of having those sorts of metrics and tracking them religiously.

25:11I think that effectively what it all boils down to is different channels have different returns on investment at different points in time. And if you're an investor or you're a company marketing manager, the key is you want to be making as much money as you can. You want the highest return that you can. And so if you have CAC and you have post-acquisition value, that gives you the basis for the ROI calculation. So just you really want to make sure that you're kind of doing all of your investing in the most intelligent way. And having kind of the most up-to-date granular numbers is what really makes that possible.

25:49Yes. Another reason to not average your marketing results. I want to go back because you mentioned the Pellet we did, our first show with you, the Rise and Fall Peloton has seen through the lens of CLTV. We talked about DoorDash and subscription models. You know, what has transpired in those pieces of research, if anything, that is new and newsworthy to our listeners? Good question. So Peloton continuing to tread water big time. Yeah. So I think that all the things that we talked about, I think they've kind of continued to hold. And it doesn't seem like the company has yet been able to really crack that.

26:37I know that they brought in new leadership. So it's going to be interesting to see whether they're able to do something that the previous leadership was not able to do. But, you know, I would say I was kind of optimistic that Barry McCarthy, not just because he shares my name, but because of his really good track record. Yeah, look, he had a great subscription model track record. Yeah, Netflix, Spotify, CFO. I mean, this guy, he is solid. yeah so he he i won't say that he kind of exited with his tail between his legs but they're still sucking wind big time so and they they rolled through a couple cmos i believe as well so i think it's probably a tough model to make money on how about you know uh we talked about the power of subscription models for not just the company but also the category any updates on that or is that just still rock solid actually we just had a big update on that paper just a week or two ago.

27:42And in fact, a lot of the things that we added to the paper were things that we talked about, Mike. So I was kind of thinking of you as we were putting it together. If you want to credit me in one of your papers, I would love to be credited in an academic paper, and we would love to have CMO Conventional called out in the paper. So the acknowledgement section, I think one of the things that we talked about was when a category matures, do the effects change? You know, if you're the last one in to create a subscription program, what does that mean? And I'd say one of the things that we found, so we really did a deeper dive into kind of how things were different when Postmates launched their subscription versus when DoorDash launched their subscription.

28:33And the good thing about DoorDash was when they launched DashPass, Postmates Unlimited had been in the market for a while. And so you have some people that are signing up for DashPass for the very first time, but you had some people who they signed up for DashPass, but they actually previously had a Postmates Unlimited subscription and then they canceled it. And so you've got these people who've got kind of varying degrees of prior experience with these delivery subscriptions. and what we found was the the people who had prior experience with a competing subscription you still saw a big bump you know they were spending a lot more at doordash but there was there was no category expansion you know so i think there tends to be if you sign up for a subscription like this for the very first time there's like a novelty effect you know that oh i've got uber one right now you know i got to use this thing you got to get a lot of food delivered yeah so you see this big increase in restaurant delivery spending um not only with uber but with uber eats but you know just in the category as a whole um when you're on your like third you know or fourth subscription you're you kind of know what you're doing you know you see you're like i've got this amount of spending i'm going to be doing yeah and you're you can do arbitrage and all kinds of stuff.

29:58Yeah, it becomes more strategic. So you still will spend more at the company that you have the subscription with, but you're not going to, you don't have that honeymoon effect of just going bonanzas in the category. It's going to be, I've got$100 to spend over the next couple of weeks. I'm just going to reallocate more of that towards this company where I get free delivery from now. So that was really interesting. and there's actually there's two other things i thought are just super cool so i'm really i i i don't know that i've ever been so excited about a revision um it's palpable it's coming right over the zoom yeah either that or the coffee that i had before um the second thing that was really neat was you sign up for dash pass you cancel dash pass now suddenly, yeah, you, there's no, you're not getting a monetary benefit for placing orders at DashPass anymore.

30:57But what we found was, um, after you cancel your, those people still subsequently spent significantly more at DoorDash than the people who never had a subscription to begin with. And so it has this kind of long lasting effect that even extends beyond like the duration of the subscription itself so i think a lot of people when they think about am i going to be am i making money on my subscription typically they'll say well how long did the people stay and how did the spending change while they were subscribed but here what we're saying is well you want to go even further than that yeah that's what you're saying is the subscription has brand power yeah even after it's gone and that's exactly the sort thing you want, you know, that it doesn't even need to be there anymore.

31:49And there's still, their behavior has changed. Yeah, you can raise the price. And when people disappear, you can still maybe get some of the benefit. So that's interesting. I have to ask before we get to our traditional last question, you also wrote a piece called Everyone is Cheating Their Way Through College. Tell me, I'm sure that's AI based, but tell me, tell me about it. well certainly professor yeah not even close to your league but i want to hear this i didn't write well so i didn't write that piece but i was commenting on that piece um yeah i do think that you know we're kind of entering this weird era where um you know students have just super easy availability to you know to these llms and um and these llms are now very they're very smart no so So, you know, Ethan Mollick, he's this Wharton professor that's done a lot of great work in the space.