Wednesday Mar 02, 2022

Google Privacy Sandbox (and how it could kill Facebook ads) EP - 20

Listen on:



Greg Marshall  0:02  
Okay, we're good. All right.

Blake Beus  0:06  
So it's funny you You and I both had an idea to talk about something today. We hadn't talked with discuss before. And it was the exact same topic

Greg Marshall  0:16  
as the same top 100%, the same topic. And I was curious, because I was going to actually ask you if you heard about this, right, yeah. So why don't you reveal what this topic was?

Blake Beus  0:28  
Google announced, what they call privacy sandboxes. And and

Greg Marshall  0:34  
what's the translation for that translation log interactive?

Blake Beus  0:40  
Well, they're kind of light on detail. But let me let me just talk about sandboxes really quick, because this is not a new concept. This is a concept that's used quite a bit from, like a security standpoint. So quick, super simple example is there's there are people out there security experts that research computer viruses, but they don't want to infect the computer with the viruses, but they need to run that virus on a computer to see what it's doing and see how to write their software to remove it or detect it or whatever like that. And so they put it in what's called a sandbox, which is basically a walled off area of the computer, that when something is inside, that it can't escape and get out. So that's like what a sandbox is. So Google, Google's privacy sandboxes would be would be something along those lines where things would run in isolation. And it sounds like I mean, they're kind of light on details, but it sounds like, they're going to either give people the option or enable the option by default to make sure that your apps are running in isolation. And so what you're doing in the app is stays within the app. And then the tracking doesn't bleed over out outside of the app or data aggregation or gathering data doesn't, doesn't, you know, can't get into the the app from outside the sandbox or whatever. So yeah, that's, that's the topic well, and

Greg Marshall  2:03  
essentially, what that means, you know, the what caught my attention, because, you know, always good headlines always fear tactics work. And I saw the picture of the head of Mark Zuckerberg face. And I said, you know, something along the lines of like, you know, Facebook's in serious trouble, because now Android is doing what Apple did, which is taking away. What is the third party tracking? Is that what it's technically called? What is it called?

Blake Beus  2:31  
Yeah, it would it would be called third party tracking, basically, Facebook's ability to track information outside of the Facebook app. Yep. And and apply that information to advertising strategies inside the Facebook app.

Greg Marshall  2:44  
And what do you think? So what are your thoughts? I have a couple of thoughts, but I'd like to hear yours. What are your thoughts on how this will impact digital advertising? Particularly? Facebook? I feel like it impacts everyone. Yeah, like every platform no matter what. Yeah. It seems like Facebook, for whatever reason, is like the target.

Blake Beus  3:03  
Yes. Well, I mean, a report came out last week or the week before that talked about how iOS his privacy changes, impacted Facebook's bottom line by about 20 billion in the last 18 months. Wow. So clearly, I mean, that's a number big enough that everybody, every executive, Facebook is worried about it. Yes. So this could also have another massive impact. Because what is something like 95% of people that use Facebook, or the time spent on Facebook happens inside of a mobile app? Yeah. And the two platforms are iOS and Android. Yeah. Right. And so this clearly is going to be causing them concerns. What it means from a practical standpoint from advertisers and things like that is that most likely Facebook is only going to be able to use reliably use data gathered within the platform for for targeting data interest based targeting data and things along those lines. So historically, if I browse something on my computer, Facebook could track that back and then deliver me ads on my mobile device. So things like that are going to be less reliable. Yep. I think moving forward. But we'll start see some some additional things happen. So one thing that a lot of people don't really notice. You see him inside of Facebook, and I click on a link inside the Facebook app on my Android phone. It opens it up in a browser. And you probably notice that's not your web browser. Yeah, that's a Facebook web browser. So you're still inside the Facebook app. And so they're still gathering that data, being able to attribute things but if I close Facebook on my phone and open up Google Chrome or Firefox on my on my phone or Safari, if you're on iOS Do some browsing there. Facebook's not going to have any any knowledge about what he did what I'm doing, what if I'm looking at, you know, dog beds? Yeah. On him.

Greg Marshall  5:11  
Now, let me ask you this. How much of an impact does that really make? So meaning is it mostly due to what I've been hearing a lot of just like within like forums or Facebook groups or whatever, Twitter, Twitter's a place that I actually see a lot of information from advertisers, because I follow them. And what I'm hearing is they're they're basically saying that you can no longer really rely purely on the algorithm. And so you actually have to like be a marketer. Right? And, in my mind, that's kind of good luck. Do you believe that's true? Is it most because like, I always felt like I saw courses all the time about hacks. Yeah. Right. And I don't see those anymore. Yeah, my guess is those are going to be completely gone. Yeah. So what are your thoughts about that?

Blake Beus  6:04  
Yeah, absolutely. I was gonna mention hacks. And like back in the day, you probably you probably remember this more than 90, but four or five years ago, a winning Facebook ads strategy was to have have have your ad with a couple of catchy phrases a catchy image, and then you just duplicate that ad set, like a bajillion times targeting different very tiny, tiny audiences? Because because that would that would essentially trick the algorithm into delivering your ads, more that those types of things are just not going to work? They probably I mean, they don't really work now. Yeah. And they're probably going to work less. So what people are going to have to start focusing on is product market fit. Yep. And the language and things will stuff read the language needed to to get people inside that target demographic, to take action to take that next step to click on the link. And, and it's going to go back to traditional marketing strategies. I personally think we're gonna see a lot of these small Facebook ads, ad agencies kind of go away. Yeah, right. And when I say small agencies, I'm talking like a single person agency that watched $1,000 course and his business has been target has been running Facebook ads for people, I just don't think many of those people have enough passion or drive to really dive into the core principles of marketing. And and to put in all of that extra effort. They were relying on the tricks, right, you set up your campaign like this, use this objective. You you you do this, and this and this, and it all happens inside the Facebook ad platform, and you win. Yep. I don't think that's going to work and video

Greg Marshall  7:46  
game has passed. Yes. Yeah. Well, and what am I here? My thoughts? So if there's less tracking? I mean, theoretically, it shouldn't necessarily matter, as long as you have good messaging and good ad creatives. And a good product. Good. You know, just like the fundamentals. Right? Yeah. What do you what do you think about? How well do you think do you think it shifts? As far as how Facebook and these other ad platforms show an ad? I've been hearing a lot about what your ad copy and ad creative becomes your targeting? Is Yes, what I've been hearing a lot of explain to me, my understanding of that is essentially like, if I write an ad, let's say I'm targeting people who own dogs, yeah, it would still be somewhat broad targeting. But then you're saying, Hey, if you own a dog, I got this dog food is the greatest thing ever. Blah, blah, blah. And that somehow, the algorithm or Facebook or Google, whatever, figures out who to put that in front of is that? Is there any accuracy to that?

Blake Beus  8:55  
I would, I would hazard a guess. And it's honestly, it's all kind of guesswork at this point. But in my mind, I would say yes, in and that's not a new concept either, right? Like, interest based targeting has not always had as the number of interests that are available in there, right? That's always kind of fluctuated, so. So you probably remember, and I know I've done this before, especially, this is especially useful for small geographic areas. You can only target so small geographically, but if you want, like if I'm running some ads for dentist's office, yeah. Or something like that, you would target as close as you could. And then and then you would say something like, you know, hey, Hill City resident Yep, target to call them out, because then you're the last bit of targeting is happening in your ad copy. You're basically telling people to self select, oh, they're speaking to me. Right. And that's, that's a totally entirely not creative way to do the targeting the ads, but, you know, with dog owners, you could really speak speak to the dog owners. Talk about the very First line could be something catchy that resonates wholeheartedly with someone who has owned a dog. I'm not a dog. Yeah, so I don't know what that would be I own cats. Yeah. So probably the very first thing for me would be like, you know how that cat keeps scratching your favorite piece of furniture? Yep. Right? I'd be like, yes, yes, I do know, that frustrates me that that gives me an emotional response. And it makes me want to read further.

Greg Marshall  10:24  
Well, here's a question for you then. So as far as the delivery, and maybe I'm looking at this different, or incorrectly? So as far as delivery of the ad, though, how would you would you still rely heavier on interest targeting to guide it? Or would you just say broad, and then somehow the platform would have to figure out how to put that in front of someone? Because hypothetically, if I launch an ad, and you launch an ad, and there's no targeting, how does the platform decipher what pocket of people to show that?

Blake Beus  10:58  
Yeah. So in my eyes, this is where you, you got to just test, I think this is going to be different based on on the the product, the industry, the market, whatever, because some some industries or interest interests are going to be easier for and, you know, a Facebook ad platform to identify than others, right? Like, if it's something along the lines of does this person like dogs? Yeah, Facebook could determine that based on how many dog Facebook groups they're in. Yeah. Right. That's the pictures they post. Yep. They're scanning the pictures to see if those are pictures of dogs. And yep, like, they have access to all of that information. But if if it's something along the lines of, you know, is this person interested in? I can't think of something right now. But something there's not like a clear face like roof or like, like fixing my roof. Yeah, if I'm, if I want, yeah, that's a good example, something that I need right now. Like, I there was a hailstorm and I need my roof fixed. Face. I'm not following roofing groups on I have a followed roofing group for years and years. And Facebook, this is something that I kind of need right now. And so Facebook is probably not going to know anything about that. Right? And so it's going to be one of those things you'll have to target. But when you say if I target something broad, how does Facebook know who put in who to who to put it in front of this is where historical ad account data kind of comes into play in my eyes? Yeah. So what they do at first is say I have $100 daily budget, and my target audience is 50 million people. That's super broad, right? Obviously, they can't show it to all 50 million people to get a good sampling of data. Yep, on my $100 a day budget. So they're going to pick a pocket of people to show it to. And then And then based on how those that pocket of people react, whether they're clicking through whatever, and who clicks through and what attributes they have, they can then expand that to larger pockets of people. That that pocket of people is probably picked based on historically who has clicked on things in your ad account. And they're going to try to show it to the best people first, to get kind of a good idea on that. But again, Facebook doesn't publish this stuff. But yeah, if I were to write algorithms, and if I were to write a platform to try to figure this stuff out, it's not just me, other people that are in data science and everything. This is how it would probably most likely work. Yep.

Greg Marshall  13:29  
So here's my thought to that, then, if you had something like roofing, right, that you would have to then you made, my thought is you may have to rely more on research, and then plug in the kind of the demos that would match that. So one thing that I know from selling, you know, even like some T shirts and specific areas, yeah, remember that company that would sell shirts that were based off of where you're from. So that that's a good example to like, in the past. If you were to try to just target like, let's say, Baltimore, or Atlanta, or New Jersey, whatever. It seemed like that audience was always like, too small for Facebook to like, really scale to anything, right. Only spent so much money. Yeah. But I wonder now, if you were to do targeting based off cities like that, if it would start to work, because at least it gives you a better direction. You know what I mean? Like, there's not as much interest targeting, for example, like, if you're doing roofing, you would go like, well, what places have the most storms? Yeah, right. And you and you would plug in all these cities? Yeah. And then target the cities. And then you would say, well, homeowners usually are like 30 Plus, right, right. And you'd have to use more like logic like that. Is that what you think would happen or is that oh, yeah, the wrong direction?

Blake Beus  14:57  
Definitely. I think I think the market is Make it are going to be the marketers that are good at high quality research of identifying those things and figuring those things out. And then taking the extra step of doing that extra targeting inside the ad copy, which you and I both know, those are two totally different skill sets. And not a ton of marketers have both

Greg Marshall  15:17  
both either really strong and one. Right. The other but both,

Blake Beus  15:21  
not both. And it's hard to be an expert at both. Because there there is, I don't know, using different parts of your brain I

Greg Marshall  15:28  
like it's just, it's just different. It's different patterns. Like it's like doing math and painting. Yes, that's the way you think about that is not the same. No, not to come up with some No, not the

Blake Beus  15:41  
same. And so if you're doing it as like, you know, a career or whatever, you're spending some dedicated time focusing on the one piece, while at the same time, you're getting worse at the other piece, because you're ignoring exactly you're ignoring working that muscle. So I mean, now that we're talking about I think a good combo for an agency would be to kind of basically pair up the analytical research person with a an emotion based copywriter. That's flexible. Yep, that would be in my eyes, a pretty killer combination.

Greg Marshall  16:12  
That's essentially what you need. Right? And, you know, I feel like with the question, I have them moving forward, as, as I guess, targeting becomes weaker? Right? A lot of the optimizations were done, because there's so much data being fed back into it. Mm hmm. How do you think that's going to impact like? Or would you even need to be using conversion objectives? If you're doing because the point of conversion objectives was to like, get the algorithm use all this data it was getting Yeah, or to get a result? And then it needs a certain amount events in order to optimize? Yeah, well, if it's not getting all that data, you may never reach that unless you spend like $10,000 a day.

Blake Beus  16:53  
Yeah, I think I think we're going to need longer runways, when you first launch an ad for conversion objectives to get the right kind of data or you're using conversion objectives for that are super simple conversions. So for example, a conversion objective on a $500 purchase May in the past may have been fine. But now, that's that's too far down the road, people have to think about it, because what would happen in the past is people click on the ad, that maybe 500 bucks is not something a lot of people just buy on a whim. So they maybe come back and think about it, maybe bookmark the page and come back in a couple of days and then buy, well that that conversion, a couple of days later probably won't get communicated back and count towards your ads. Whereas historically used to if so, I would say we would like to use conversion objectives that can happen in a single session. Yeah. So So you're still within Facebook's browser window, you click on the ad, you're in the browser window, I make a gut decision right now in the first 30 seconds, I'm going to go for it. Yep. And I make that purchase, then that conversion objectives will probably get communicated back just fine. Because you're still within the ecosystem, you're still inside that same sandbox. Yep. And so the data in there can be shared. And that's, that's what I would think the the other thing to think about is you can use conversion objectives for like, percentage of video views and things like that, and then retarget, and follow those people up with just a general, you know, general targeting that just blast it out to everybody in that group. Because, again, that information is still all within the sandbox, if someone's watching a video on inside the Facebook app, and they're watching two minutes of this four minute video, and you're targeting people that watch two minutes or more. Facebook already knows that like that, that data is really available to the algorithm and you can retarget based on that very easily.

Greg Marshall  18:47  
Well, here's two things that I'm seeing currently running ads. Number one, there's been two ad accounts, where I almost shut the campaigns off, and then they picked up. But both of them had the same exact characteristics. It took 14 days before actually started to opt interesting. And then it picked up interest and then it's become consistent. What kind of a daily budget. Were you looking at? 100. So one is 150. And the other ones 200?

Blake Beus  19:17  
Were they at least like break? They were making some Yeah, it was breaking even so it wasn't like you're losing $150 a day for

Greg Marshall  19:24  
now. So it wasn't like $0 Yeah, it's just you were you were probably like at a point eight row ads.

Blake Beus  19:31  
So you lose a little bit of money. You're, you're losing about 20 bucks a day.

Greg Marshall  19:35  
Yeah. And so you're like, well, if it's not hitting you the temptations to relaunch. Yeah, but then after the 14 day period, now it's been consistent. Actually one of the other accounts the one account name talking about has actually jumped to 3.6 and the recharge 3.6x rose on cold traffic. Dan five on retargeting, what, but it's taking literally weeks weeks like enemy he was not interested in at all

Blake Beus  20:04  

Greg Marshall  20:05  
And so what as I saw that I'm thinking, I'm wondering if and same thing, same budget $100 a day. Yeah. I don't see any of that with anything less than that. Yeah. So if you spent 2050 bucks there's there's literally no stability whatsoever. Yeah, it's almost like now before a lot of people that would get in like smaller companies, they could spend $20 or 10 or whatever. And in my experience, try testing those budgets is has an hour

Blake Beus  20:31  
interest. You have to be spending a budget or you putting into retargeting.

Greg Marshall  20:36  
The retarget on this one, I

Blake Beus  20:38  
believe is only 10. Okay, so targeting was pretty small. Yes. Okay.

Greg Marshall  20:43  
So yeah, sorry, I left that part. So the retargeting budgets. I always think of cold traffic. But yeah, the retargeting budgets, our website visitors, they haven't bought last 30 days using $10 A day budget,

Blake Beus  20:54  
interesting website visits. So I would say in my eyes I'm guessing retargeting website visitors is probably going to get worse and worse. Yeah, over time. Yep. And retargeting based on in app events is going to get is

Greg Marshall  21:10  
going to be the agent like Facebook or Instagram and gays in our video views. And so

Blake Beus  21:15  
one of the things I wish they would do, you can do this with a video you can do with a post is I wish you could retarget people from a specific post yes, that have interacted with a specific post? Because then you could you could build up some some things with image ads, you could do that with a video people have watched a portion of the video ad or whatever, but not a specific post

Greg Marshall  21:36  
or even like people who hovered over specific posts. Yeah, a certain amount of time. Yeah. Is that were expanded the comments. Yeah. Because that takes at least give you like, an insight on how qualified they are. But the other challenge you always run into is like the size of the audiences, right? Like, if you have something where let's say you're in an ultra niche market, it's a lot harder to build a retargeting audience than it is for like fitness or yeah, how to make money or more generalized topics. And so that might be where Facebook becomes more challenging is for highly niche audiences. That just, there's just not enough people. Yeah, to actually like, reach. That's where well and hit before I even move on to the next thing. Here's the other thing that I've been running a lot of tick tock ads. Oh, yeah. Early testing of the same. Like. So here's, here's something. So we spent let's see what campaigns $100 a day and others at the next Rama test is $200 a day to see a budget determines anything. Here's something that's super interesting about tick tock. You know, on Facebook, when you run, like, optimize for add to cart or view content, it tends to not do that well. So far on tick tock, if I optimize for purchase, it does horrible. Uh huh. But then if I optimize for Add to Cart, it does great, really. And it's very odd, because you're like, that's an easier conversion to get to add to cart. But I get more purchases optimizing for Add to Cart than I do for payment. And here's the thing, I changed no variables, same audience, I didn't have any competing. So I turned off the add to cart one. So like there's no overlap. I had multiple the same exact ads, everything the same, the targeting was the same, same budget, but no purchases on the one optimizing for purchase after three day period. And then the other one is $100 a day. And then the other one had I want to say was six purchases within the same timeframe. Same budget, same audience, but I was optimized for Add to Cart

Blake Beus  23:46  
interesting. In my eyes, that probably is an indicator that their algorithm is still young, it needs to be refined some more, because it's easy. Like that's an easier thing to capture algorithmically. I guess you help them it

Greg Marshall  24:06  
says add to cart. So I was I was curious. The other thing that was very interesting was, I like to look back at the data as far as accuracy. So so far, I'm still in like the early stages of optimizing for purchase, right? So I've only had around 30 purchases or so in the last few weeks. What I'm seeing though, is the accuracy of the ad, ads manager and the conversions are 100%. Really, so every single purchase that came in through Tik Tok shows on the ads manager 100% accuracy. So it was there was none off because I went to went and I'm using a Shopify store. I went into Shopify, look through all the traffic, the traffic actually says tick tock, this person converted from there, and the exact number of tick tock buyers showed up on the ads manager, which I thought was. That's interesting, really unique because i How are they getting?

Blake Beus  25:08  
The same with Facebook and Google? It's always all over the place. And I might just be because coincidence. I mean, maybe I guess it's gonna say might just be because traffic from Tik Tok is very specific, right? Google's this massive ecosystem with all of these different platforms and everything. So it's, it's probably, there, there's so many different ways people can get to a particular landing page inside the Google ecosystem, sort of that that it's it's probably hard to identify even for Google or Shopify or whatever, where you know how to attribute that. Whereas with tick tock, it's like one path. Yep. So so that was my that would be my guess. But I really don't know.

Greg Marshall  25:53  
Because I was, I mean, well, to me, it sounds like a great guest. Because I was looking at it because I don't understand the the data science part behind how algorithms all work. And I was just looking at it like, wow, this is incredible. It is 100% accurate, like, how is and it tells you exactly how many events it separates it out to how many events have happened that the pixel has captured. And how many events came specifically from ads, which is different from the Facebook events manager? Yeah. Which clumps them all together? Yeah. And you have to use the ads manager to see yeah, here. It actually has a lineup, it almost feels like Tik Tok is looking at all the weaknesses of Facebook and what Facebook advertisers complain about. And are giving you those.

Blake Beus  26:36  
Yeah, I would guess, and this is all just a guest. But I would guess over time, that's going to get worse. So one of the things tick tock does, and this actually drives me nuts get like absolutely nuts. Excuse me, if if you click on a link in tic toc and go somewhere else, you're open, you get opened up in tic TOCs. Browser. Yeah, there is no way to get out of that browser. You you can't like all the other browsers, you have to you can click on a button, say, you know, open in my my Firefox browser or chrome browser or whatever. Or and I'm talking inside the app on web yet, but very few people use tick tock on the web, like so few. I can't even copy a link inside of the TIC tock browser by long pressing on it, copying it and opening it elsewhere, which, which I do that's part of my regular flow, because I like to keep certain tabs open to come back to them later. But within tick tock, tick tock, if you open up a link, you can't do anything other than go straight to that link. You are stuck inside their app. So maybe that's how they try and they're forcing that but I'm guessing over time due to pressure of people saying This sucks. Yep, that's going to go away. And you're probably going to start seeing less

Greg Marshall  27:53  
got it as the time to get accurate database. Yeah,

Blake Beus  27:57  
I guess. It tick tock tic TOCs an interesting one. Yeah, sure. So anyway, well, I don't have anything else to talk about on the subject. I think we've covered really well. Did you have anything else?

Greg Marshall  28:08  
No, I think that's okay, that covers it. That's where we're at. All right. Well, Greg,

Blake Beus  28:13  
how can people how can people chat with you? Well,

Greg Marshall  28:15  
if you want to chat, go ahead and go to Greg Marshall co book a call. You can talk strategy. And what about you

Blake Beus  28:20  
just Blake Beus calm? I kind of have everything on there.

Greg Marshall  28:24  
All right. Well, thanks, guys for listening, and we'll we'll talk to you next time. All right.


Comments (0)

To leave or reply to comments, please download free Podbean or

No Comments

Copyright 2021 All rights reserved.

Podcast Powered By Podbean