← The Marketing Architects
The Marketing Architects
Nerd Alert: Targeting Without Tracking
00:08:14
Welcome to Nerd Alert, a series of special episodes bridging the gap between marketing academia and practitioners. We're breaking down highly involved, complex research into plain language and takeaways any marketer can use. In this episode, Elena...
Loading summary
Transcript18 lines
- [00:00]
A
Nerd Alert. Learning is important, right?
- [00:02]
B
Yes, exactly. What a bunch of nerds.
- [00:05]
A
Nerd alert.
- [00:06]
B
Marketing Architects. Hello and welcome to the Marketing Architects, a research first podcast dedicated to answering your toughest marketing questions. I'm Elena Jasper on the marketing team here at Marketing Architects, and I'm joined by my co host, Rob demars, the chief product architect of misfits and machines. Hello. Hello. We're back with your weekly Nerd Alert. Every week, I'll take a deep dive into academic marketing research and translate its complex ideas into simple, understandable language for Rob, and of course, for all of you. Are you ready to nerd out, Rob?
- [00:35]
A
I'm feeling so nerdy. I just set my password to supercalifragilisticexpialidocious with an exclamation point, which is great because I can never access my account again.
- [00:46]
B
I was going to say I would not be able to spell that. All right. This week I read Reach Measurement Optimization and Frequency Capping and Targeted Online Advertising under K Anonymity by Juan Gao and Mu Kwai from LinkedIn. The study looks at what happens to reach frequency capping and ad performance when platforms can no longer track individual users and instead have to rely on privacy safe user groups instead. But before we get too far, Rob, when you hear the phrase privacy first advertising, do you think, all right, this is a future, this is where we're going, or do you worry that it's going to break a lot of what currently works in advertising?
- [01:19]
A
I don't worry about it at all. It makes, it makes me think less tracking, more thinking. I think it forces us to stop, you know, pretending precision equals persuasion. I think it just continues to elevate. Creativity matters. We're shifting away from a strategy of surveillance and focusing on what really matters.
- [01:41]
B
And you work at a television agency that doesn't. I do on this type of. So you're not that worried about. It doesn't keep you up at night, it keeps a lot of marketers up at night because they're dealing with it, with their digital channel. So that tension is what this paper tries to resolve. What happens to our advertising effectiveness when individual tracking is gone and everything has to work at the group level instead? So the study, it centers on a privacy framework called K anonymity. Instead of identifying individuals, platforms place users into groups of at least K people based on shared traits like geography or interests. So instead of saying Rob specifically has already seen this ad twice, the system only knows someone in group A has seen this ad. So that's a big shift in how Digital advertising works once individual tracking disappears. If it does, frequency capping becomes probabilistic instead of precise. So a platform won't be able to say this exact person has seen this ad three times, stop showing them the ad. It can only estimate the likelihood that the next impression will reach someone who hasn't hit the cap yet. So frequency capping doesn't go away, but it gets more difficult. The authors call this probabilistic frequency capping, and it sounds a little odd, but the goal is still the same. Avoid overexposing people while maximizing reach. The difference is that instead of certainty, you're working with more of a probability. So, Rob, does that sound like any other traditional media channel we know exactly right. So that's how traditional media like TV has always worked. You never know who saw the ad, just the odds that someone in a household or an audience segment did. So digital could be being pulled back towards that same model. But let's talk about our favorite topic, reach under K anonymity. Reach is no longer a clean count of unique individuals. Instead, it becomes an expected value. So platforms, they can still report reach, but it's an estimate, not a headcount. The more privacy you add, the larger the groups and the more uncertainty there is in that estimate. The study also looked at optimization, specifically how platforms should bid when they don't know who the individual user is. So traditionally, frequency capping was binary. Either the user was eligible for another impression or they won't or they weren't. Under privacy, that system discounts bids based on probability. So if there's a high chance the impression will hit someone already over the cap, the platform bids less aggressively. If they think they could reach someone new, it's going to be more aggressive. This creates this decline in bidding intensity instead of a hard stop. So one other interesting insight is that a small number of highly active users can distort everything. So. So if a few people are generating the most impressions in a group, the system's going to become overly cautious, even though most people in the group haven't seen the ad yet. Which again, feels familiar if you've ever planned around heavy TV viewers. So we're kind of getting back to a little bit of traditional planning. The researchers also ran simulations to understand the trade off between privacy and performance. So, Rob, what would you expect to happen to add performance as privacy increases? Do you think there's going to be some total collapse, A great gradual trade off? What do you think would happen?
- [04:33]
A
I don't think it'll be a total collapse. But I definitely think that the stuff that's gonna break the fastest has always been on borrowed time. Considering the scrutiny around privacy, we've known this is a topic, we've known that it's coming, that stuff's gonna take the hit fastest. Hypergranular targeting is obviously gonna be the first domino to fall. But the upside again is just more honest signals, cleaner experiments and better learning over time. You know, this idea of the precision theater is just going to go away.
- [05:04]
B
You are pretty spot on. So when they simulated moving from individual tracking to these more smaller anonymous groups, it caused about a 1/3 drop in efficiency. But increasing privacy much beyond that caused much smaller incremental losses. So I think you're right, like that hyper targeting was what really took the biggest hit. They also found that in the extreme case, like if everyone just had to become one giant group, performance started to look more like traditional mass media. So probably just harder to measure. Which means privacy is not going to destroy advertising. It might just make digital behave more like tv. And you might need to look at it more like a channel like television. So let's talk about a few takeaways. First, perfect precision was always a luxury. Individual level frequency capping was definitely nice, but it wasn't essential to effectiveness. Advertising still works when outcomes are probabilistic. Second, privacy pushes us back towards reach first thinking. Because as these platforms lose their precision, broad reach becomes more important than trying to squeeze your efficiency out of tiny segments. And finally, expect some messier measurement. Reach numbers will come with more uncertainty. That doesn't mean all campaigns will fail. It means dashboards will be less exact. We, I think, know firsthand how difficult measurement is with a channel like tv. So there's a chance that digital could become a bit like that too. Time for a Rob GPT. Think of digital advertising like weather forecasting. No one expects a forecast to tell them with certainty whether it'll rain at exactly 3:17pm on their street. What we want is guidance. Even without perfect certainty, certain forecasts can be incredibly useful. They shape our real decisions every day. And privacy first advertising works the same way. You might not know exactly who saw your ad. You can understand the conditions. The numbers might be less exact, but they're still good enough to make strategic choices. And that's what this research is saying. Uncertainty doesn't make advertising effective, it just makes it probabilistic.
- [06:46]
A
It's funny, we do, we just, we actually want precision so badly that we get angry when we don't get it. Because your example with the weather, it's like, well, now, you know, you start getting weather apps saying it's going to rain in one minute, and you're like, oh, that's so amazing. It's telling me it's going to rain in one minute, but then it doesn't rain in one minute. And you're like, that was actually. The precision was actually useless.
- [07:09]
B
Yeah, it's just. It feels more comfortable to have it sometimes than to not have it. But, yeah, that was interesting. It's fun to read a paper like that because all this privacy stuff has been odd. Like, at one point it was a huge topic as everyone thinks is going to hit now. I think people are way more concerned with AI than privacy. And there's. I don't know, there's still a long way to go to know how it's going to affect things, but that it's good for people to be aware of what things could look like if they did have to move to that model. So that's it for this episode of the Marketing Architects. We'd like to thank Taylor de Los Reyes for producing the show. You can connect with us on LinkedIn and if you like the podcast, please leave us a review. Now go forth and. And build great marketing under K. Anamin Anim. Oh, no. Animity. Is that how you say it?
- [07:55]
A
Animity.
- [07:56]
B
Animity.
- [07:57]
A
Anonymity.
- [07:59]
B
Anonymity. Is that right or are you just teasing me?
- [08:01]
A
Anonymity, Anonymity.
- [08:03]
B
Is that, is that actually how you say it? Because I'm going to say it a lot. Marketing Architects.