Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
OhHello - #everycannes who is TIM?
75 Plays2 years ago
Tracking. Identity. Measurement. A lively discussion with Jeromy Sonne (Daypart.ai), Mark Donatelli (Cimply LLC), Olivia Kory (Haus), and Jeremy Bloom (OhHello.io)
Closing down 5 great days and sessions for 2023's #everyonecannes
Recommended
Transcript
Introduction to the Session
00:00:00
Speaker
Oh, good morning. I am Jeremy Bloom. I am chief mentor and founder of Oh Hello. We're going to do a fun little video montage here of an intro for a few seconds because it's
Inclusivity in Arts and Advertising Discussions
00:00:19
Speaker
just fun. Why not? Why not? You guys don't need the break. We don't need the break. I'm Jeremy Bloom with us today. We have Jeremy Sonny, Jeremy with an O versus Jeremy with an E.
00:00:30
Speaker
Olivia Corey and Mark Donatelli. We're going to go around the horn. Why don't you each, so let me take a step back. This is day five of hashtag everyone cam.
00:00:39
Speaker
The purpose of these sessions are to make, uh, when we had so many friends and colleagues, uh, go and participate in the South of France and the Colcet, uh, last week talking about the arts and advertising marketing. What we decided to do was make this inclusive for everybody everywhere, as there are, uh, a lot of sessions and we thought we can make this fun and virtual and something that has been incredible, incredibly pressing in our industry, trapping.
00:01:09
Speaker
Identity, measurement, T-I-M. What is Tim? Who is Tim? So with that said, we're going to go around
Expert Introductions and Backgrounds
00:01:18
Speaker
the horn. We're going to talk about tracking. We're going to talk about identity. We're going to talk about measurement. Jeremy, welcome. Hello. Hey.
00:01:26
Speaker
Thank you. And I do want it noted that Freewheel invited me to Cannes, among others. So I chose not to go. I'm above it, but I am not one of you PBS. So anyways, all right. So what track do I do? Who are you? What do you do? Oh, who am I and what do I do? I'm a guy that tweets too much. I'm sorry.
00:01:53
Speaker
I mean, you know what? It works. I run a company called Daypart AI. We are collapsing the analytics and media buying stack and are creating the media OS for modern brands and agencies. Awesome. Awesome. Olivia, Olivia Corey. Hello. Hello.
00:02:18
Speaker
And Jeremy's right. I would never have thought that Twitter is a place where business happens, but confirmed. It does. And it's great that I met you all, except you, Jeremy. I met Jeremy and Mark on Twitter. And it's been wonderful. Little on me. I have a background in growth marketing. I spent some time at Sonos. And prior to that, at Netflix. And in that role, saw the way privacy was moving, saw how challenged
00:02:48
Speaker
tracking and attribution was becoming. And in Netflix, I kind of learned like there's another way here, something called incrementality testing, running kind of high quality control geo experiments as a solution around all these privacy issues. And so we're building house to make that whole process really easy from like the moment you decide you wanna run more experiments through design, planning, execution and analysis kind of automating away a lot of the heavy lifting of data science.
00:03:17
Speaker
Amazing amazing and Olivia you join us from Detroit Jeremy You're joining us from Austin, Texas and mark you are joining us from Cleveland Akron mark Donatelli Thank you for joining us. Who are you? Oh Hello, hello I'm currently running a business called simply with a C We're a strategy and operations consulting firm. We find ourselves a lot of times working with a
00:03:45
Speaker
ad ops, marketing ops, rev ops, like where those things tend to, I won't say intersect, because they tend to collide, like at high speeds. The CMO and the CRO are not on the same page, et cetera.
Challenges in Modern Marketing Tracking
00:03:57
Speaker
So we find ourselves in those situations. My background, I started a really long time ago. I was in army intelligence. So put a pin in that when it comes to tracking, identity, type of stuff. So I've been in this game of,
00:04:13
Speaker
Who's who and where are they and what do they care about? I've been in this game for a really long time, worked at Axiom, I worked at IPG, WPP, been around the whole pond, as they say. Amazing, amazing. So what we plan on doing for the next 20 minutes or so is just talking about tracking, talking about identity, talking about measurement. Olivia, in your experience as a growth marketer, as someone who's been in ad tech, in martech and
00:04:42
Speaker
a co-founder of House who is rolling up her sleeves and truly getting shit done. Why don't you tell us just a bit of how you look at tracking and just with all the iOS changes, what does this holistically mean to you? You had an amazing thread that you posted on Twitter. Why don't you just kind of peel back a couple layers of the onion and tell those of us a little bit more? Yeah, I can give the high level. I think what I kind of argue in the post is that
00:05:12
Speaker
Uh...
00:05:14
Speaker
It's kind of not worth our time to even try and keep track of this stuff anymore. Really the central argument is like, it's not my job to care anymore. And so I hope that folks in these roles and in growth don't need to keep spending their precious time understanding what every single iOS update means because it's just things are moving in one direction. But at a high level iOS 17, two big changes, one around click IDs.
Shift Towards Privacy in Marketing
00:05:44
Speaker
What this means for brands and for advertisers is that when you, you know, everyone who's in the ad platforms, you know, and you're trying to understand where your clicks are coming from, you'll implement a click ID. Apple is saying that now for private browsing for messages and for email, they're going to strip that click ID out. And so UTM parameters kind of more broadly should be okay, we think, but like click IDs are going to be very challenged.
00:06:12
Speaker
I try and really separate out, like once a user has volunteered some PII, like once they've purchased and given you an email address, you're probably fine and in the clear. But before they take that action of giving you something, that's going to be very challenging in terms of being able to track those clicks back to an ad platform. Everyone's already kind of thinking through workarounds, but again, my kind of argument is like,
00:06:36
Speaker
you really want to keep playing whack-a-mole here and figuring out a solution that might not be relevant in a year or two. The second big change is fingerprinting. Now, I was at Quibi as a mobile app, so I got really familiar with this stuff where these mobile measurement partners will have an SDK inside your app and they'll be doing measurement and tracking and attribution back to the app platforms.
00:07:01
Speaker
And what iOS 17 is doing is saying, hey, what those guys are doing now, you as the advertiser and the brand, you're responsible for what is happening inside of your app.
Value of Incrementality Testing in Marketing
00:07:12
Speaker
So the MMPs and some of these SDK partners might say they're not doing anything, but they might call it something different than fingerprinting. They might say they're not fingerprinting, but you should really, really understand what they're doing and how they're doing that attribution because you could be on the hook for it as a brand.
00:07:42
Speaker
Give it to us. We've known for the last few years that the end is coming. Nobody's gonna say, you know what, I want less privacy. That moment isn't coming despite the absolute fever dreams of a lot of media buyers I know and DSP owners and stuff. It's done.
00:07:56
Speaker
Thank you, Olivia. Jeremy, you're a man of many words, and I saw you smirking a few times during that.
00:08:03
Speaker
right it's done it's been done it's only gonna get more done right and anytime somebody comes up with i don't know like unified id iteration fifty seven or whatever it is whack-a-mole it's putting a band-aid i'm like you know
00:08:19
Speaker
like cutting off somebody's arm and slapping a band-aid on it saying we fixed it, right? And frankly, too, I'll even go a step further. I've seen a ton of quote-unquote solutions that are just almost borderline griffs, right? Where they say, oh, we solved this. And you pull off the Scooby-Doo mask, and it's like IP matching again, right? But none of it makes any sense. And the only way forward is to actually look at durable solutions, right? Which means holistic, which means
00:08:48
Speaker
Right. And so I think that if you're not, if you are an agency, you're an in-house growth market or whatever, and you're not like seriously in the process of evaluating or implementing a probabilistic solution at some point, you're doing your organization a deep disservice at
00:09:14
Speaker
So that's my take.
Deterministic vs Probabilistic Tracking Strategies
00:09:16
Speaker
Mark, you've been handed. This is precisely where I think that the marketers, like the emperor has no clothes. Like we're going to find that the vast majority of people calling themselves marketers are not marketing anything. Like they're not even, they missed the first, they missed the
00:09:39
Speaker
on all this data leakage to be able to target and identify and like they're doing it all wrong. Like, do you mean that you're eating like a bunch of ads out on Meta and letting it do all of the algorithm, the optimization work? Does it make me a great partner? Let's pay max it. Let me Google it. Number one recommendation, spend more. Like every one of those recommendations. More and let the out, let it run longer.
00:10:09
Speaker
The algo. Every day, light the candles for the algo shrine and let the algo take over. Sorry. I'm of the mind, again, remember my background. To me, I believe firmly that every single person sleeps somewhere. Sure. Any marketable universe, I mean, obviously, you look at a marketable universe and people that aren't in prison or dead or under bridges or whatever.
00:10:39
Speaker
Most people who have money to spend, consumers who are spending money with brands, they live somewhere. And back in the old days, before digital got out of control, you know, before digital got out of control, we did things a lot in a more prescriptive kind of way. We used public record data. We would be very careful about how we did that because it would cost five or six dollars to reach someone.
00:11:04
Speaker
whether it be with a direct mail piece or a catalog, like this was real, like old school blocking and tackling. This is real marketing. Like, and we planned things and we thought things out and we did a lot of analytics. Now we're letting the data just, and the algorithms just decide and it's super wasteful. It's not good for the consumers. Like I'm again, I know Jeremy and I have gone back and forth on online a lot about the role of identity and
00:11:31
Speaker
I'm not, again, I think there's a role for deterministic identity. I work with some healthcare clients. I work with some financial clients. You know, there's no your customer, no your patient. Like there's a lot of validity to having deterministic identity at some point in the funnel.
00:11:48
Speaker
And to Olivia's point, look, if you have a great product and you price it right and you provide good support,
Measurement Challenges and Solutions
00:11:54
Speaker
guess what? It's going to be successful. You're going to have customers. You're going to be able to look at your current customers, do some data work, do some modeling and some, you know, they would call it a custom audience at Facebook, right? But it's all in a black box.
00:12:11
Speaker
There's ways that you can do this responsibly, understand from your current customers who you should target, and then go use things like geography, time of day, channel, contextual publishers. There's a million ways to reach people and be effective. And I think a lot of marketers in the last 10 years have gotten super lazy. We've brought up an entire generation of marketers that really don't know how to do it.
00:12:40
Speaker
Most of our time clients hire us to come in and fix the stack, like, because they can't tell the truth. They don't even know what it is. The CRO, the CRO is like, wait a minute. I gave you $2 million last quarter. Show me in the pipeline. Like show me where, show me which leads came from which campaign. Olivia, jump in. Fire.
00:13:03
Speaker
I have a hundred conversions from Facebook and I have a hundred conversions from Google and the CRO saying, well, I don't see 25 units in our business level data. So it's even worse than that Mark, because like you're right, like marketing probably used to be more incremental. My founder and I talk about this all the time where it's like,
00:13:21
Speaker
You're buying TV,
Complexities in Ad Tech and Marketing Evolution
00:13:22
Speaker
it's more reach frequency based, you're actually influencing people. And the way the ad servers have moved, like the sophistication of the server is such that they are all meta, Google, TikTok, they are all trying to time this so that they're the last touch before you come in.
00:13:38
Speaker
and they can take that credit. They know a lot more than we think. In that way, I still would probably bet on PMACS and ASC. I don't know if I would pull the plug on that, but incentives matter. And having them be the ones to tell you to spend more is a little off-putting. And then also, just the way the ad servers are moving is really problematic for folks who have to actually create demand. How is it problematic, in what sense?
00:14:07
Speaker
Oh, they're all, they're all kind of attribution stealing, if you will, and trying to be that last touch before you kind of come in and purchase. That's the way in which it's problematic where like more traditional advertising didn't have that luxury and had to actually do the work of. And there are, there are some bigger brands, like brands that are investing in their own identity graph internally, like their own, there are big brands who are probably doing better at managing and fighting this because they're, they know what they can, they know what they know. And they're like, yeah, yeah, thanks. We got it.
00:14:37
Speaker
I can build a data layer, but the smaller mid-market companies are completely at the mercy of these platforms. And Mark, I'm going to admit a terrible secret. I use deterministic tracking, and I use custom audiences, and I use device graphs, and everything. Yeah, you have to. So yeah. So it's not, despite my very inflammatory engagement bait,
00:15:03
Speaker
tweets It's a combination like you're saying, right? I'm actually super bullish on Deterministic in sandbox environments right stuff like retail media, right? Like I think deterministic is a hundred percent appropriate with retail media, right? But I'm talking when I say probabilistic and blah blah blah
00:15:22
Speaker
I'm mostly talking about like, you know, your CTV, your audio, your, your open internet. On the fraud, we don't even touch on fraud. I mean, there's, there's the consumer, consumer preferences and all that's one thing. Government legislation's another, like, is it unfair in the market? I mean, full disclosure, I'm an Apple shareholder. I'm a Google shareholder. I'm a Microsoft shareholder. I'm an Amazon shareholder.
00:15:47
Speaker
At the end of the day, look, I'm going to keep getting richer, like, as long as they're doing what they're doing. I sold out of Facebook at 180 bucks because I just got so disgusted with them as a thing. Like, I was just so mad at them. I mean, I could have made more money, I guess, if I had stayed in. But I will keep on, let's pivot to where you're going with this, Mark. So like Facebook to me, it's just a mess. So again, I think that there's people out there that
00:16:16
Speaker
they thrive on the fear, uncertainty and doubt. Whether you look at the Martek landscape, like the, or the Loom Escape or the, one of my big talk tracks in a previous company I founded was, look, it's more hype than it has helped. Like those people that are talking about the complexity, they're part of the problem. Because I think the reality is that as a business consultant, we like to say, okay, what are the objectives? What are you trying to achieve?
00:16:42
Speaker
budgets, like you start to go down the process and you realize that that entire complex landscape really doesn't matter to the vast majority of marketers. They need to be in one or two channels that work for them. They need to incrementally improve.
00:16:54
Speaker
and then expand their channel mix as their audience dictates that, as their audience moves or expands, then you go meet the audience where they're at, but it takes time. I love that last point, because it actually hits on both what Jeremy and Olivia respectively are both building in terms of lift, in terms of measurement, in terms of attribution. Jeremy, what were you about to say? And then we can hit on my next question.
00:17:20
Speaker
All I was going to point out too is that we've been talking a lot about like we've been ragging on performance brands, right? Which is fine. I'm down to ragging on performance brands. I love them. They're my people. It's where I came up. Like my first ad gig was on accident in 2011 running Facebook ads, right? So I am one of those people that very much so had to figure out how to do, quote unquote, real marketing.
00:17:47
Speaker
It's one that's eternally ongoing, but started like six, seven years ago when I was like, oh, I can't just rely on Facebook's algorithm forever. I'm going to get automated out of a job. And that's what kind of started me down this whole path of letting me hear. But that said, the other side of it that I think is worth mentioning when it comes to deterministic tracking and probabilistic tracking is like I'm seeing a lot of the same conversations, but from a different
00:18:15
Speaker
with legacy brands as well which is hey we can't just buy reach and frequency anymore we have to actually like the CFO is all over me we got to know exactly like what is the return what's ROI he's kind of like yeah what is the ROI what is the ROI even just per channel right like you know and how do we
00:18:39
Speaker
fund that owns us or whatever. And we're not going to get that next tranche of capital if we don't just measure it. Yeah, exactly. So what we're seeing I think is performance brands moving this way and legacy brands moving this way towards this like center of, you know, a messier probabilistic kind of ROI measurement. But I think it's going to become like the
00:19:08
Speaker
All very increment and not actually incremental is probably not the best word here because I'm mostly defending incrementality. But they're all just very like they don't solve. They're not looking at the seismic shift that's happening in the industry. They're all just like patchwork. Right. And I think that like
00:19:25
Speaker
the people that win are the ones that are looking inside the ship, whether you're like a senior performance. People are building solutions. And the word solution to me implies that there's some, I think a lot of people believe that as, oh, it's going to be a software thing. Yeah. That's the solution. Well, part of the solution is having the proper services framework and the right connectivity and even terminology and taxonomy amongst departments, right?
00:19:52
Speaker
You know, we had a client that the CRO was up to, the CMO got fired, right? And we came in kind of like on the heels of the CMO on the way out because
00:20:01
Speaker
she just couldn't answer any questions that the CRO was trying to answer. And so what we found was that, look, the case study is actually on our website,
Connecting Marketing Systems for Better Insights
00:20:11
Speaker
right? But there's a bunch of connectivity that when you have multiple systems, you have a CRM system, you have a Salesforce automation system, you have a DMP, you have Facebook, you have all these people out there. One of the things that they thought is they said, oh, well,
00:20:28
Speaker
This publisher that does these webcasts or educational sessions that we partner with, that's the best source of our leads because those are the ones that are in the pipeline. What we found out, once we went in and connected the data flows, attending a webinar is not the lead
00:20:48
Speaker
like qualifier, that's what gets them over the hump. The ones who are highly qualified then go to, they've already been influenced by our nurturing sequence, they've been influenced by our LinkedIn media buys, they've already been influenced and have decided to attend a webinar, then they show up in the pipeline. So thinking that you spend more with the webinar provider is not the answer. You need to go up the stream and figure out, well, how did they get to the webinar?
00:21:19
Speaker
What, it was the webinar the first time we saw them or was it the fourth time we saw them? What point in the journey did the webinar happen? And those are the kind of things that you, there's no right answer for any company, right? It's always different, but that's the onion that we try to appeal is to say, okay, well, you may think that this is a leading indicator, but really it's a lagging indicator, right?
00:21:40
Speaker
I appreciate that, Mark. Olivia, you spend a lot of time with incrementality testing. What are you looking for? Walk us through just your mindset. When you're helping a marketer just figure out their ROI, what is the methodology of your use case?
00:22:02
Speaker
Yeah, well, Mark just nailed it in terms of what we do. Like that webinar example is perfect of if you're on the surface, invest more in the webinar. Like that's what we see that all the time. The two most common offenders are brand search.
00:22:16
Speaker
and affiliate marketing. And like within affiliates, there's a lot of affiliate stuff you can do. And so I am more specifically targeting like discount sites where you like know you want to buy something and you go search and get your discount code. Based on a Google Analytics last click report, which is
00:22:34
Speaker
Um, sadly still, like pretty common is, uh, the way to kind of understand marketing's contribution. Um, you would say, I'm going to invest all my money in these two channels. And we see that with like a bulk of performance dollars being invested there. You run an incrementality test and this gets into the methodology where like you do a controlled experiment where the only thing you're isolating is the brand search.
00:22:58
Speaker
So you have a group of regions where you're not spending in brand search, a group of regions where you are. Statistically, they're indistinguishable from each other. We see brand search sometimes does work, but most of the time it's much less incremental than you thought. And you do need to kind of discount what you're seeing in the platform reported attribution. Sometimes it can still work, like it's cheap enough that it still has a role in the mix.
00:23:22
Speaker
But these are the places where we find brands are typically over-investing and then areas like YouTube, like this one, everyone's running a YouTube test right now because everyone says like, well, it doesn't work. I'm not seeing any conversions come in through the GA or platform reporting. And when you actually look at it the right way of like looking at the business level, what is this actually driving in my data warehouse where I'm counting these new orders and this revenue?
00:23:49
Speaker
it actually is like a pretty important channel.
Experimentation and Simplification in Marketing
00:23:51
Speaker
So those are just a couple of examples of like how you can actually run experiments to answer these questions versus to Mark's example, just continuing to pour money into things that maybe we're gonna drive your business anyway. Jeremy, were you about to say something second ago? Well, beyond just Mark and Olivia being much smarter than me, which is true, you should listen to them mostly.
00:24:16
Speaker
I guess the only thing I want to preface what you all just said, which is completely right with is I think one of the things that's keeping marketers from jumping away from their little pixel is it's safe. And what we're talking about seems really complicated and hard.
00:24:35
Speaker
And Olivia is probably going to physically cringe when I say this, but like if you're getting started, it's okay to do some bro math. It's okay to just like estimate it and it's okay to use like some low confidence intervals and like just dip your toes and like, yeah, your probabilistic is less probabilistic, but like you're still
00:24:56
Speaker
going to probably mostly have a pretty good idea, right? So like this can be as simple as pick a city, try something, say what were the sales last month, I tried the thing for a month, what happened to my sales, did they go up or down, and then make a decision as to whether or not you want to apply that, maybe try after that three more cities the next month, and if that holds
00:25:25
Speaker
that simple and you can run a bunch of these in parallel. It doesn't like Olivia, Mark are completely right about this, right? But to like somebody that has just like grown up on the Facebook pixel doing everything, like this is a terrifying conversation and not a printable. So just keep it simple start. And then as you learn the skills, you can get it more and more like data correct. I see so many data scientists that are brilliant.
00:25:52
Speaker
that scare away marketers from probabilistic arguments. This is why we name the company simply. This is why we name the company simply because everything is just too complicated and we think there's a basic approach. It's different for everybody. One of the things that you're, you know, what you're describing, you know, a previous company I founded was called Science of Bacon.
00:26:18
Speaker
SOB, SOB, right. Exactly. It later became hive. And I still actually have that it's, it's the predecessor to simply really was when I was on my own. I have a business partner now. But ultimately, science of bacon was founded on the principles of the scientific method. And, and step one, if you recall, in experimentation, is you just do something, right? And then like, to your point, Jeremy, just do it. And then track the results, change your inputs,
00:26:47
Speaker
track again like this is just basic scientific this is basic like 10th grade biology yeah yeah i learned this is 10th grade like like experiment measure outcome and it doesn't have to be pure because on the one hand you have the people that are you know basically throwing out like yet another identity resolution product that's like really just
00:27:12
Speaker
You have people that are like data scientists purist that will not accept anything less than a 99% confidence interval and just muddy the water to the point that every meeting becomes a headache. Like just just stake yourself in the middle ground, say I'm not going to be perfect, but I'm going to
00:27:33
Speaker
Oh, I was just gonna say, I'm not cringing at all. I love that. I think most businesses are like, if you're not at a certain size, you should do exactly what you just described. And yeah, I mean, I think just doing it and you'll see it in the business level, like, and I think that's
00:27:55
Speaker
That's the most important thing. So I'm, I'm, I'm fully supportive that the data science P value thing is something that we talk about a lot where it's like, what if you could just money ball it and make like a bunch of 70% confident decisions? Like if you can just increase your velocity of testing, do more, you're right. 70 or 80% of the time, isn't that better than running an eight week, 99% statistical powered experiment and potentially running eight weeks of like a losing variant.
00:28:24
Speaker
So we really try and reinforce that idea of not every experiment needs to be statistically powered to that 99%. And if you could just get your velocity up. And for businesses of a certain size, you should be able to see it at the business level. That's something that I think, to Mark and Jeremy's point,
00:28:45
Speaker
you don't have to overthink it. And like, this is not like- Keep it simple. Not magic. This is, we actually pride ourselves on not being like a magic pixel or anything like that. Like it's pure scientific method. And that's what I love about it.
Conclusion and Key Takeaways
00:28:58
Speaker
Amazing. So to wrap it all up, I think where the group is going is Tim, whomever he is, wherever he is,
00:29:09
Speaker
It should just be a lot more simple. It shouldn't be as complicated. It shouldn't be as convoluted. There's always going to be evolution and just to be able to test. Going around the horn to wrap things up. Any? Jeremy, real quick. Did you come up with Tim? Is that your idea? No. Oh, him. Sorry, wrong, Jeremy. He's honestly going to have it at some point. Is that you? Jeremy Bloom, yes. Yeah. Good one. I like it. Thanks.
00:29:39
Speaker
Thank you, I appreciate it. But having three panelists and guests that know the topic, topics plural, well better than I do, makes it a lot more enjoyable. So I appreciate all three of you being on here. I appreciate you guys taking the time to discuss who is Tim, what is Tim, where is Tim? Any parting words of wisdom and advice for those that are watching and since this is being recorded, just as
00:30:07
Speaker
as experts in our field, just one or two sentences each. Olivia, see you nodding, go. It shouldn't be that hard. That's easy, love it. I took that from Mark and Jeremy, I like it. Mark. I would say to build on that, start with the basics. I think that understanding who your current customers are, how they use your product,
00:30:35
Speaker
Those are good places to start when you start thinking about how do I get more customers? What are the experiments I want to run to get those customers? Start from the basis of facts. Keep it simple with a C. Jeremy. Don't be a midwit. Just run ads and see what happens. Nice. Well, thank you, the three of you, for taking the time. To everyone watching,
00:31:04
Speaker
whether you're watching live or recorded, we appreciate it. Oh, hello.io is going to be launching this summer. Very fortunate to have so many wonderful mentors and experts that are part of this community that we're building. Thank you, everyone. Have a wonderful day. And thank you for joining hashtag everyone can till the next one. Thank you. Thanks.