The Cookie Collapse: How Are Savvy Marketers Adjusting?

The “cookie apocalypse” will mean the end of an era for digital marketing as we know it. We’ll discuss with digital analytics expert & CEO of Provalytics Jeff Greenfield how savvy marketers are adjusting and what strategies and marketing technologies these organizations are implementing right now to thrive in the new normal.

 

Hugh Macken:
And good afternoon everyone. My name is Hugh Macken. I’m with VMR Communications and I’m joined by my co-host Danielle Milliken.

Danielle Milliken:
Hello.

Hugh Macken:
And we are here with a very special guest who I’ll introduce momentarily. And our topic today is the Future of Digital Advertising in a Cookie-Less World. And we’ll be looking at what the future is going to look like and how digital marketers can and should adjust to what is really a rapidly changing landscape for digital advertising. As I said, we do have a special guest. His name is Jeff Greenfield.

Jeff is an expert in all things digital measurement and he is an acclaimed expert. He’s been featured in many, many mainstream news outlets for his expertise. I recently watched an interview, you did Jeff, on Bloomberg and was just really impressed. We’ve known each other for several years. Jeff is the CEO of Provalytics. And Jeff, if you wouldn’t mind, I mean your resume is so long, I think you’d be better at summarizing your background. Would you, first of all welcome, and please tell us more about yourself and your background in digital measurement and analytics?

Jeff Greenfield:
Thank you so much Hugh, and Danielle as well.  In terms of my background that’s applicable for the audience today is I started on the brand side specifically doing branded entertainment, these large scale programs for brands that really get the word out. But that was back in the early 2000s and back then there was no way to measure anything. The measurement was really lost. People started moving towards digital because you could actually measure clicks and that was really cool. But then digital advertising got so complicated and it was really tough to figure out what was working and what wasn’t working.  So in 2008 I started a company called C3 Metrics. C3 was the leading multi-touch attribution company for enterprise clients. So we’re talking folks like JP Morgan, US Bank, a ton of folks in the pharma and the financial services space who spend a substantial sum each year on marketing.

And for them it’s very difficult to figure out what was working and what wasn’t. And it was an interesting time because we were able to get the entire digital trail back then we had tags that were on Facebook, even on Amazon for a period of time, even YouTube, we could collect all of this data and it was really incredible. So we had a full deterministic methodology of figuring out where you should spend your next marketing dollar to get the biggest bang for your buck, if you will. And then all of a sudden things, the world changed and Facebook stopped allowing tags, YouTube stopped allowing tags and that wasn’t a big deal. But now we’re living in a world where there’s so many different channels that don’t allow you to measure, not clicks, but the impression, which is that when somebody gets exposed or sees an ad.  And so you don’t have independent ways of measuring that. So I exited C3 Metrics at the end of 2019 and started looking at the landscape to see what was going on out there. And that’s when I started Provalytics. And Provalytics is the next generation of attribution that works incredibly well and was designed for this new privacy centric cookie-less future that we have. And that’s part of the reason why cookie-less is coming up is it all revolves around privacy and stuff like that. So it’s coming… 2023 and 2024 are going to be very interesting years with all of the changes that are coming up. And I’ll leave it right there with that, Hugh.

Hugh Macken:
No, sounds good. So Jeff, if you wouldn’t mind just elaborate a little bit on at a very high level what some of those privacy changes are from a technical standpoint. And I want to look at the browser level at the OS level. Apple for example, with its privacy policies. So cookies, third party cookies being deprecated, identity solutions being proposed by the likes of the trade desk and being adopted, unified ID 2.0. So tell us a little bit more about what changes are going to be happening on the technical side. And then I’d love to talk about of ways that marketers can adjust on the technical side, but also ways that marketers can adjust on the strategic side. So how might marketers align more closely with media? But let’s just start off with what are all of the technical changes that are happening with, just give us a bird’s eye view.

Jeff Greenfield:
Sure. I’m going to give a little history with this and go back to something that was called GDPR, which was the start of all of this. And GDPR is the European version of privacy controls. And the way the whole internet has always worked is that the internet was free and the reason it was free was because of advertising and everyone was automatically opted in. And GDPR came out and said, “Guess what?” People are not automatically opted in. And if you have a list of people that have said, “Hey, I want to get your newsletter,” you have to assume that none of them are on your list and you have to email them and ask them for permission. So that caused a ripple effect because there were a lot of companies that had built businesses on identity stitching, meaning connecting someone who was at their work computer with their iPad at home and their mobile device.

And they had stitched together all of these different identities. So when GDPR came out, it ripple effect occurred where a bunch of companies that were operating internationally shut down their European operation because it was so stringent it would be starting brand new again even though they had been operating for a half a dozen years. So the next legislative thing that came about was in California, the CCPA. Now if you notice and if you have a website, you should already have this, there’s a button or a link down at the bottom that says, “Do not sell my personal information.” So California came out with this law that says you have to have a link at the bottom of your website for California residents where they can access their information, they can ask you to amend it or they can ask you to delete it.

And the regulation states that you have to have a form that people can fill out and also allow for an 800 number. So when this first came out, a lot of people thought it wasn’t that big of a deal, but California just recently fined a very large retailer for this. So they’re getting very serious about that. But shortly after that, Nevada came out with another law and Maryland came out with another one. So from an internet regulation standpoint of view, there’s pending legislation in at least 40 states for how you deal with people’s information. And that ties right in to how the internet has always operated. Now it and the operations have to do with cookies. Now if you notice when you go to Amazon, you log right into Amazon if you’re not using your phone and you go there, they know you, they know all your stuff.

That’s called a first party cookie. A first party cookie is the type of cookie that the site you’re on, the one that shows up in that URL bar where you type in whatever site that is, a first party cookie is written by them. So if you go to the New York Times and you log in, New York Times can read and write their own cookies. But the internet and all of advertising was built on third party cookies. And what that means is that when you’re on the New York Times, there’s a bunch of other vendors and ads that are there that are dropping cookies that are from different domains. And the reason they’re doing that is if I’m an advertising company and you’re on the New York Times and then you go to the Wall Street Journal, I want to know that you’re also at the Wall Street Journal.

If I show one of my ads so I can control things like reach and frequency. And that’s how the whole internet has always functioned. Now, Apple has always operated in their own ecosystem and privacy has been a very important component to the entire company. So in their Safari browser, they’ve shut off third party cookies for a very long time. They’ve been gone for a very long time. And to marketers, most marketers just said, “That’s okay,” because there’s more Android people than there are Safari, so it was no big deal. And that occurred both on the browser and on your cell phone as well too. Firefox, which is a small percentage of the browser community, they also shut off third party cookies. And then the big announcement is Chrome, was supposed to do it next year, but it’s being pushed to 2024, Chrome has about 65-70% of the market.

So what that means is that the way advertising has worked, the way you have targeted, the way you’ve measured is going to be completely different in the next year to year and a half. And we always talk about, and Hugh and I have talked about this for years, the concept of digital transformation. And most companies have gone through a digital transformation. Well guess what? I hate to say this. There’s another one coming and it involves how you target, but also how you measure. So the other bad news that goes along with this is that the number one metrics that everyone uses is Google Analytics. And everyone has years and years of data in there. And it’s great how you can go in and see your historical data. Well next July 1st, July 1st, 2023, Google has said that platform will stop taking in data. It will no longer accept new data and you have to use their new product, which is GA4.
Now, I don’t understand why a company the size of Google can’t just change the code and have it all secured in there, but they are unable to do that. So what that means is that you have to change a code on your website, which is no big deal for some people. But the other big issue is that the reports in GA4 look completely different. You don’t know where your conversion report is and all of these things. And so now there’s a learning curve. So if you have a small company or even a medium size or large company, not only do you have to pay to get the code change, but you also have to go through a whole new training. So the next year to year and a half is going to be a major transformation that’s going to have to go forward.

Hugh Macken:
And so Jeff, what I hear you talking about really are two different aspects in a way of digital marketing. On the one hand measurement. So how do we do measurement? The way we do measurement going forward is going to change. And then the issue of audience targeting. And really there’s the identity resolution issue that relates to both of those issues. So on the one hand, we want to be able to measure the effectiveness of our advertising and the ability to do identity resolution effectively has aided us in doing that. At the same time, audience targeting cookies, third party cookies have enabled that. So that too will be impacted. Is that fair to say that it’s really those two areas primarily that will be impacted? And really the challenge before us is putting forth technology infrastructure that will allow us to adjust to the changing environment and then also strategies that will help us to adjust as well.

I’m just struck by how many articles I read about how to solve the problem of audience targeting and reaching audiences. Do you think it makes sense in terms of adjusting to the changes from a strategy standpoint for organizations to be maybe asking a fundamentally different question, which is what are different ways in which, what are different models we can use to do more effective advertising? Because at the end of the day, what matters for the CMO say at an institution, say a higher ed institution or an e-commerce provider, isn’t the ability ultimately to target audiences? Ultimately it’s the effectiveness of their advertising that really matters. So are there other ways of going about that? And we’ve spoken quite a bit about the idea of aligning with media. What do you see as the future from a strategy standpoint and a modeling standpoint to adjust?

Jeff Greenfield:
Well, and that’s part of the big adjustment that’s going to go on Hugh, which is that most marketers today in the digital realm are addicted to the granularity of data that they’ve been brought up on. A great example is my daughter. My daughter ran marketing for a large auto dealer group. And I remember her telling me a couple years ago, “It’s pretty amazing. I can go into Facebook, I can target Ford F-150 leaseholders whose lease is going to expire in six months.” I’m like, “That’s pretty granular.” Well that’s all gone. And so for a lot of digital natives who have never done planning and large scale, both in traditional and digital media, a lot of them are feeling like they’re being choked a bit because it’s like, “How am I supposed to work with this broadness?” But the reality is that the research has shown that this level of granularity that we’ve gotten so addicted to, it’s like we’re in the middle of the forest and we can’t see the trees, that old analogy. We’re in too deep and we’re too addicted as marketers to this granularity of data.

Now remember, third party cookies are going away, first party are not. So what that means is, in the example I gave of the New York Times and the Wall Street Journal, and there’s all sorts of other specialty publications out there, they have paywalls up that require people to sign up or people will sign up on the email. And what they’re doing is that they are mining that data and building up really great segments internally. So I think one of the things that’s going to happen is that when we had this explosion of ad tech, we had a move towards exchanges like the trade desk, but some of the earlier ones were all of the buys were done on these exchanges where if I’m the New York Times, I don’t have any people selling ads. What I’ve got is I’ve got ads that are all going through an exchange, but now we’re moving into a world where doing direct buys have a huge benefit because of the level of data that they have.

So for example, if you’re like a college or university and you’re focused on applications and enrollments, you can go and do a buy on Google and Facebook, but you start to dig into your data and you start to look at, you may have 20 different degree programs and there may be one on nursing that’s very, very popular. Well, there’s a lot of specialty publications out there that you can go to and do direct buys on that specialize in nursing and nursing careers. That would be great to flow into this strategy. So I think that’s part of it from the targeting aspect of where folks are going to have to start thinking because that granularity of data is, it’s not even available now and it’s going to be even less over the next year or so.

Danielle Milliken:
Yeah. So Jeff, you made a comment about direct buys and how that’s going to become the future of digital marketing. So can you speak a little bit more to that in the sense that how companies can maybe adjust what they’re doing? Like their team. How they might need to adjust their teams and their roles moving forward to make sure that they’re ready for this landscape change, whatever that might be.

Jeff Greenfield:
Well, I think one of the best things that a company can do is the difference between direct buys and the exchange buys all has to do when you’re negotiating. So sending your head buyer or the person who’s handling it out for a course on negotiation and researching online, what are the types of things that you can get with a direct buy? Because with a direct buy, you can negotiate a package where you’ve got ads on the site, they push out a couple of newsletters, maybe they do an advertorial for you as well, customized content.

And Hugh and I have talked before, and he’s even showed me examples of some of the work that you guys have done with really, really cool email capture that goes right into the client’s CRM, which is amazing and these types of deals are all out there. But I think the key is trying to figure out what’s available and then also negotiating the best deal because it’s a whole different world versus just buying ads and spending money. This you have to plan. That’s the other thing, is the time that’s involved too. And there’s a lot of work that goes into planning these types of campaigns.

Danielle Milliken:
For sure.

Hugh Macken:
For sure. Yeah, no, that makes sense. So Jeff, if you would speak a little bit about the key components of a marketing technology stack. So again, imagine you’re speaking to a CMO at a university say, and they’re trying to figure out, “Okay, what are the key components that we need marketing?” So I’m thinking like CDP, DMP, data warehouse, like Snowflake. What are a platform like Provalytics or there’s, I believe Oracle has a product in relation to measurement and analytics as well. So what are the key components that you would recommend to help marketers adjust to this new landscape?

Jeff Greenfield:
So the first thing is that I would set with the CMO, the understanding that Google Analytics, which 99% of marketers use, is an awesome web analytics platform. Meaning it does a really good job of once people end up at your site, you can see what they do on your site and you can see where they came from. Now anyone who’s dug into GA or even GA four, you will notice that 80% of the people come in, it’s called organic, they just show up. And the problem is from a marketer’s standpoint of view is that if you’re using Google Analytics to plan your media buys you’re doing the wrong thing. There’s no ifs, ands, or buts about it. Most marketers know that you have to fill the top of your funnel. There’s that AIDA, awareness, interest, desire, and action. And in order to fill that funnel, that awareness, you have to put messages out there, they can be targeted, but the broader, the better, as long as it’s within your target or your geo that you want to be at.

And what you really want to know is when you look in Google Analytics, what is driving, how did those people find out about my site to come directly in? Another way to look at it is that most larger brands will spend money in Google and what they call brand search so that when you type in the name of the company or the university, you’ll see them listed there first. Well how are people coming through brand search? Because Google Analytics will show you, “Hey, brand search is converting really well. Wow, you’re getting a lot of leads from brand search.” But it’s because brand search people already know about you in order to search. So you have to ask yourself, I really want to know who’s driving, what is driving brand search? What is behind that? And now for smaller marketers, one of the best ways to do it, so I’ll give a solution for smaller marketers first.

For smaller marketers, you want to do a little research into something called regression analysis. And regression is all about trying to figure out what’s the causation behind these things. And what you would do is you would take all of your channels that you’re spending on and you would look at how many impressions you’re producing and how many clicks each hour. And then you would run regression analysis between that and your brand search click. So you can see who’s actually driving brand search because that’s your top of the funnel drivers and that’s where you want to spend. Now for larger folks and larger players like a university, you need a platform that can automatically grab that impression data, not only analyze it, but have the proof behind it that it actually works. And you see this is a huge distinction. One of the biggest problems with most attribution and all of these platforms is they do a really good job of reporting about what happened last week, yesterday or last month.
But they don’t do a really good job of planning, which is what we’ve been talking about a lot. And planning is all about forecasting and saying, if you do this, you’re going to get this much money in return. And what’s really important with that is to be able to predict what the return will be on investment and then putting your money where your mouth is and showing confidence scores for that, which is what we do at Provalytics. And very few analytics platforms, if any, actually do that. And that’s important because remember at the end of the day you want to know how are people finding out about me?

That’s where you want to spend your dollars. It’s very easy to spend close to the bottom of the funnel. Google search is easy, Facebook retargeting is a great place to be. And then of course, as we talked about specialty publications that you can go to, these are great, but how do you fill that top of the funnel? In order to do that, you have to have a platform that can do it or you have to do a hand analysis to figure out what is driving your brand search, what is driving your organic traffic.

Hugh Macken:
Yeah, no makes sense. And I mean you’ve mentioned branding quite a bit and I wonder, are we going to see a pivot here on the digital side to where digital marketers start actually focusing on branding as opposed to just constantly focusing on direct response? I just read an interesting article recently on Airbnb and how they invested in branding and it really paid off with respect to direct response. So what’s your thought on that? Are we going to see more of a shift toward branding on the digital side?

Jeff Greenfield:
I sure hope so. And I’ve mentioned this book before, I’ll just hold it up right here. The book is called Lemon, It’s by Orlando Wood from the IPA in the UK. It’s available on Amazon, full color. It’s amazing. And what the author has done is he has researched the effectiveness of ads and what they’ve seen is they’ve seen an increase since 2006 in what they call short-termism. And short-termism is spending closer to where the transaction is. So it started, because in 2005, most brands, all they had was a brand. There was no performance divisions or anything like that. But when digital started coming out, they started testing and then they built up teams. And now there’s for large brands, they spent a lot on branding, but they spend even more on performance marketing. And look at like CPG, how close to the sale can you get?

Well now not only can I be on Amazon, but I can also be on Walmart. And if I’m tied when someone puts a competitive product in their basket, I can push out an ad that will compel them to buy mine as well or to switch it out for a discount. That’s a lot of money being spent almost at the point of purchase. And that’s what you call a rise in short-termism. And it’s directly correlated to a decrease in ad effectiveness since 2006. Brands definitely need to invest more money, especially now as they say the R word, the recession. And at the ANA masters a couple weeks ago down in Florida, the big thing that everyone was talking about is that marketers and CMOs are going to have more pressure than ever on their marketing budgets to prove to their CFO that if I do this, I’m going to get this. And to be able to plan effectively and forecast and then of course actually hit those numbers. That’s the other problem is that if you don’t execute, you’re definitely not going to hit your forecasts.

Hugh Macken:
For sure. Yeah. So you’ve talked about this model, Jeff, in terms of demonstrating the effectiveness of the advertising, this new model that is a hybrid between media mix modeling and multi-touch attribution. Would you mind just explaining that in simple terms in terms of, well first of all, what is multi-touch attribution? What’s media mix modeling or marketing mix modeling and what’s this hybrid that you’ve been talking that you’ve been talking with me about?

Jeff Greenfield:
Yeah, absolutely. So we’ll start with marketing mix modeling, MMM, because it’s been around since the sixties and seventies and M is typically uses regression modeling that I talked about earlier. And it looks at all of your channels. Now remember this was developed before digital. So companies would do an annual marketing mix modeling where they would look at their TV, radio, print, direct mail, out of home and they would do an analysis and regression towards sales numbers. And then the output of that MMA model was to say, here’s how much budget you should allocate to each of these channels for your sales to continue to move up. And that was great. Now of course what happened is that companies didn’t do them every year. They would skip a couple years and they would just focus on the budget allocations from years prior. And then digital marketing came along and when they integrated digital marketing as a channel into it, since they didn’t understand digital marketing, the output would say spend more on brand search and on affiliate.

Now of course we all know that as we talked about before, brand search is only from people who are already knowing about you. You can’t spend more on brand search to get more and you can’t spend more on affiliate. They all happen as a result of all the other activity. So multi-touch attribution started around 2007 because digital marketing had become, at that time somewhat complicated because you had display, you had search, you had search across multiple platforms and how do you figure things out? And Multitouch attribution said, we’re going to collect all of the data, every bit of granular touchpoints, whereas MMM only took in aggregated data, sometimes monthly, sometimes daily, but primarily large scale privacy centric data, whereas MTA was all user level data, it all model it instead of regression use machine learning. Because if you do regression with that amount of data, it will take forever, forever to get it done.

But machine learning, you can teach the machines and we’ve got great technology. One of them has been around for hundreds of years called Bayesian, which is just incredible because it thinks, and it works the same way that human beings think because we are constantly updating the way we think. So when you think about self-driving cars that aren’t here yet, self-driving cars are always getting smarter, the more obstacles that they face. And that’s the same way with machine learning for marketing attribution. So MTA would take all of that data, put it into a Bayesian model, and it would constantly be updated every single day, sometimes throughout the day with the most updated result of what was going on right at the moment and historical. And that worked great. And the greatest thing about MTA versus MMA is that you weren’t just limited to sales. You could have multiple KPIs at C3 metrics.
We had some clients with 25 or 30 different conversion events because there were different teams that were interested in different things. Some teams were interested in traffic to a particular site that was their KPI, other folks were leads, other folks were actual enrollments when talking about EDU clients here. So that was really great and also it did really well with digital. But since MTA was built on being a deterministic model, MTA fell flat with TV and radio and then podcasts as they started to explode, it was like there’s nothing for anyone to click on. How do you measure this?

Hugh Macken:
Right, right.

Jeff Greenfield:
Yeah. And that was the big problem. The other big issue with multi-touch attribution is that there was no incrementality, so there was no measurement of, it would tell you where to spend more, but it didn’t tell you how much more business you would get.
And marketers always had this question of, “Well, what happened if I didn’t do that? Wouldn’t I get those sales anyway?” And that’s the basis of incrementality. Luckily marketing mix modeling is all based about the contribution that each tactic, but an MMM word, it’s all about the channel. So there is incrementality in MMM, but it’s at a channel level. And so what I wanted to do is I wanted to merge these two together. I wanted to take something that was always on, used some of the MMM background so that you could incorporate not just digital media, but also TV, radio, CTV, podcast, print, direct mail, bring it all together, and not just to sales but to multiple KPIs. And so now what Provalytics is the Venn diagram, if you will, of MTA and MMM. The basis is MMM, but it’s a whole new way of modeling that uses machine learning at scale.

We’re able to take in aggregated daily data from Facebook and from the platforms down to a creative level. And we’re not limited to just pushing out channel level recommendations, but we’re pushing out granular recommendations so that channel managers will know exactly how to allocate based upon their goals. And the forecasting we’re doing is incredible. If I had a team of 500 statisticians running through, because our models, when clients will give us what are their goals for a forecast, how far out do you want to go? Do you want to go 30, 60, 90, 180 days out? What is your risk quotient? How much are you actually willing to increase or decrease a tactic? And what do you want your budget to be? Do you want it to be the same? Do you want to go up 10%, 25% or do you want to go down? And then what our platform will do is go through and run hundreds of thousands of simulations and we’ll export out the optimum plan, meaning, “Hey, this is the amount of money you’re spending, here’s the ultimate, ultimate, this is going to get your biggest bang for your buck.”

Now, in all the years of me doing this, I have never seen any marketer take on all recommendations. They never, ever do. So this forecast is based on a perfect scenario. And so what ends up happening is that clients will take one or two of them, but you don’t actually know what day they put it in at and what the spend is every day. So what Provalytics does, and right now we’re starting to get a lot of the October data from our clients. Most clients send us data monthly. We have a few that send it to us weekly. But what we’ll do is we’ll borrow from MMM and we’ll do a technique called a holdout. So we will hold out and remove all of the sales data and all of the KPI numbers for each day for the last 30 days. And all we will give to our model, to Provalytics is the impressions and the clicks and the spend.

That’s all we’ll give. And then what marketers are able to look at is see how well the model predicted the outcome. And all models I should say are wrong, some are useful. And what clients are seeing is it’s for certain KPIs, the model is spot on. For others it’s not because it’s not perfect. But what happens over time is that since it’s machine learning, it gets smarter and smarter over time. And that’s another thing I’ll mention is that traditionally marketing mix modeling, you need to have three years worth of data in order to build out a model. And this is one of the other problems with MMM is that it becomes a whole project. It’s a lot of work to put this whole thing together. And so that’s why folks don’t do it more than once a year and sometimes skip years.
With Provalytics since it was built on the basis of MMM, but we borrow from the world of MTA, we can start with as little as the last three months. The more data we have, the better. So for customers of ours right now that we’re moving forward, what are we? Eight or nine business days away from when the Black Friday sales start. When we have the historical data from last year, we’ve been able to give them plans for how to execute to get the biggest bang for their buck for this upcoming Black Friday. So the more data we have, the better, but we don’t need three years worth of data. That can be overkill for our models.

Hugh Macken:
Interesting. Wow. And just in the last couple of minutes that we have, Jeff, how would you say a CDP would tie into all of this? A CDP customer data platform, obviously very useful from the standpoint of audience targeting, but with respect to measurement, is there a role that a CDP can play? Does a platform like Provalytics integrate with CDPs? How does that factor into all this?

Jeff Greenfield:
Yeah, and you stated it, One of the best advantages of a CDP is that it allows you to leverage that first party data you have, your first party cookie data. So similar to Google Analytics, which looks at how people traverse your site, a CDP is taking that and putting it on steroids, and it allows you to create unique segments within them. Now the cool thing about segments is that obviously you have a segment of people who have purchased, you have people that are multiple purchasers. If you’re in retail and then you have people that show up, maybe fill out a popup, but don’t go all the way through, let’s say in the case of EDU, the lead process, they don’t want to just subscribe to a newsletter or you just have a ton of people who just show up and then leave.
And so one of the advantages with the CDP is you can create unique customer segments, and then what you can do is you can link those segments to your media buys so that you’re buying the multi shoppers, you’re buying the existing customers to bring them back for retargeting. You’re buying the folks who showed up but didn’t actually fill out a form or anything. And then you can do it on the conversion side to see how many of those people who showed up but didn’t fill out a form, who then were exposed to later advertising ended up purchasing. And what we do with those segments is that becomes a dimension within our model. So now when you think about multiple KPIs, you can look at, let’s say those folks who show up and don’t do anything, but let’s say the segment is they have to spend at least three and a half minutes on the site and you’ve tracked how they’ve converted.

Well, with Provalytics, we can track which media is driving them and what levers to push if you want to increase their conversion rate. So those segments are really important. The other type of dimension that’s also available in Provalytics is geo. So a lot of folks just like to look at the entire, let’s say the continental US, but there’s a lot of brands that target differently. The best one that comes to mind is Southwest Airlines. They are regionally based. All of their ads are regionally purchased. Even their television, they do not buy national TV. For them, they would want to look at things in a region by region basis, and Provalytics is built from the ground up to include that.

Hugh Macken:
Interesting. Wow. So lots to consider for those of us who are trying to make sense of all of this. And Jeff, your perspective definitely is invaluable. I’d love to have you back on again soon to talk.

Jeff Greenfield:
Oh, it’d be my pleasure. Hugh. Absolutely. Anytime.

Hugh Macken:
Yeah, for sure. Just to talk more about this new model that you still haven’t come up with a name for it. I’ve been going back and forth with different ideas.

Jeff Greenfield:
Well, the name that we’re using is, it’s Agile, so it’s not locked in like MMM. And I like the word impact versus incrementality and it’s modeling. So we’re going forward with the concept of AIM, “Agile Impact Modeling”. How do you like that?

Hugh Macken:
Agile Impact Modeling. All right, there you go.

Jeff Greenfield:
You like that, Hugh?

Hugh Macken:
Yeah.

Jeff Greenfield:
Good.

Hugh Macken:
I like it. Because honestly, I was going with something that was going to be MMM.

Jeff Greenfield:
It’s ad tech. We have to have an acronym, so we’ve got it now, AIM. AIM is the acronym.

Hugh Macken:
There you go. There you go. Oh my gosh, Jeff, it was just such a pleasure. Thank you so much for joining us and definitely thank you Danielle as well for joining us. And I want to thank Elijah Medcore, our producer who’s in behind the scenes. Thank you Elijah. And thanks be to God that we were able to make this happen. This is great.

Danielle Milliken:
Yeah.

Jeff Greenfield:
Absolutely. Thanks so much for having me.

Hugh Macken:
Hopefully we’ll be able to do this again. Yeah, thanks so much. And so just final thoughts would be for those who are interested in learning more about you, Jeff or Provalytics, so your website and the website for Provalytics?

Jeff Greenfield:
Yeah. So best website to go to is Provalytics, that’s P-R-O-V-A, prova, which is proof in Italian, lytics. So P-R-O-V-A-L-Y-T-I-C-S.com. And you can go there, fill out some information and get in contact with us. And if you want to chat with me, just go to jeffgreenfield.com is my personal website, jeffgreenfield.com.

Danielle Milliken:
Awesome.

Hugh Macken:
All right. Awesome. Great. All right, well thank you. Thank you so much again, Jeff.

Danielle Milliken:
Thanks Jeff.

Hugh Macken:
And have a great every week. Thanks everyone for joining us. Thank you.

Jeff Greenfield:
Thank you so much. Have a great day.

Danielle Milliken:
Bye.

Hugh Macken:
You too.

Jeff Greenfield:
Bye.

Adobe uses AI to speed up marketing mix modeling

“Adobe’s new tool likely will appeal to existing Adobe analytics users, but it’s not the only way to overcome the privacy restrictions hampering MTA or the speed and cost issues that long have hampered MMM, said Jeff Greenfield, a pioneer in MTA who now is CEO of Provalytics. His company takes an alternate approach, combining elements of attribution and marketing mix modeling to measure “incrementality,” or the incremental impact of changing spending across a variety of online and offline media.”

“AI, while it’s a hot buzzword in marketing, is probably a misnomer here, Greenfield said. Few if any systems in marketing analytics truly show signs of human cognition or original thinking that are hallmarks of AI, he said. “Machine learning” is likely a more accurate description, he said, of the series of algorithms that could go into automating a marketing mix modeling process.”


Process that takes months is reduced to weeks, becoming practical for more marketers and potentially helping prove their worth

Adobe is launching a marketing mix modeling service that uses artificial intelligence to assess return on investment in weeks rather than the months it typically takes for such analytics.

The upshot is that marketers can use the tool, which will be generally available as part of the Adobe Experience Cloud, to adjust media and marketing plans on the fly, or at least within a month or quarter, rather than taking a retrospective look at what happened in the past to adjust future spending.

The tool, powered by Adobe’s Sensei AI engine, also appears practical for smaller and mid-size marketers that don’t have teams of data scientists on staff and often can’t afford the price tag for an outside marketing mix modeling project. One early pilot user was AAA Northeast.

Adobe’s move comes as marketing mix modeling (MMM) enjoys a renaissance after years of losing ground to multi-touch attribution (MTA). Making the AI tool available across an already huge base of Adobe could fuel the growth of MMM further.

Attribution has taken a hit because it relies heavily on tracking the online behavior and purchases of individuals, which has grown more challenging amid the dwindling availability of cookies and digital identifiers and will become even harder once Google follows through on plans to eliminate third-party cookies from its Chrome browser. Google has delayed the move multiple times, with plans to now eliminate cookies by the end of 2024.

MMM uses modeling based on a variety of data streams but doesn’t require individual identifiers. And the approach long has focused on offline media—where it had its roots—in addition to digital. But it also generally has been costly and time-consuming, requiring six- or seven-figure price tags and months or work.

Big impact in one month for AAA

An early pilot user of Adobe’s “AI as a service” approach is AAA Northeast, which used it in March to deliver a 28% increase in lead generation for its auto insurance business while reducing advertising spending 16% for the month.

“Our tech stack is primarily Adobe,” said Lisa Melton, senior VP of marketing at AAA Northeast. “And so when they came to us and asked if we wanted to be part of the pilot, we were like, absolutely. It’s every marketer’s goal to make sure your dollars are going where they have the biggest return.”

The AAA Northeast insurance business that was in the pilot uses digital display, search, connected TV, linear TV and a small amount of postcard direct mail, Melton said. Adobe’s AI MMM analysis led her to shift money out of linear TV into display and search.

Melton hopes to continue to testing the AI MMM tool on other lines of business, perhaps including membership acquisition, which has more direct mail in the mix. And she’s now incorporating the Adobe Experience Cloud into her tech stack.

Adobe previously had an AI attribution tool, but in conversations with marketers found that they wanted ROI analytics solutions that cover other media that attribution can’t, including offline cookieless media and social media walled gardens, said Monica Lay, principal product marketing manager for digital experience at Adobe.

Marketing mix modeling can analyze results from those media, but historically took six to 12 months to set up initially and three months to deliver reports on an ongoing basis after that, Lay said. That doesn’t help marketers who want to know how to spend incremental dollars—or cut budgets—within a current budget cycle, she said.

Pressure to deliver more for less

“We’re noticing in customer conversations pressure on senior leadership to deliver more for less,” she said. “We’re seeing budget cuts across marketing spend, but there’s still pressure to deliver the same revenue targets.”

Liz Miller, VP and principal analyst of Constellation Research, who’s tried the AI MMM tool, said she really likes it as “a way of accounting for all that data that can come in as a giant tsunami that no one can really manage. This starts to bring a new layer of what I refer to as decision velocity, which is about making not only good decisions, but making great decisions faster.”

Miller also believes the tool, as part of an Adobe service suite, could open marketing mix modeling to a much wider range of marketers who simply couldn’t afford it before. “Media mix modeling has been cost prohibitive for a lot of organizations,” she said.

Gerry Murray, research director for marketing and sales technologies at IDC, sees Adobe’s move as “an inflection point” for intuitiveness and simplifying how marketers can use models that drive decision-making.

‘Dawning of a new day’

“I think it’s a bit of a dawning of a new day for marketers to have a more holistic ability to show how they drive the revenue and customer lifetime value metrics that the C suite is really interested in, not just the clicks and engagement metrics,” Murray said.

Adobe’s new tool likely will appeal to existing Adobe analytics users, but it’s not the only way to overcome the privacy restrictions hampering MTA or the speed and cost issues that long have hampered MMM, said Jeff Greenfield, a pioneer in MTA who now is CEO of Provalytics. His company takes an alternate approach, combining elements of attribution and marketing mix modeling to measure “incrementality,” or the incremental impact of changing spending across a variety of online and offline media.

AI, while it’s a hot buzzword in marketing, is probably a misnomer here, Greenfield said. Few if any systems in marketing analytics truly show signs of human cognition or original thinking that are hallmarks of AI, he said. “Machine learning” is likely a more accurate description, he said, of the series of algorithms that could go into automating a marketing mix modeling process.

Read the original article here

Provalytics CEO Featured in Marketing X Analytics Podcast

Provalytics CEO Interviewed on Friday Fireside

Why Radio Needs To Invest In Data

Radioinfo’s Wayne Stamm spoke to Jeff Greenfield at the NAB Show in Las Vegas who tells radio “It’s time for us to put on our big boy pants, radio has to grow up…”.

“Radio’s reach, especially during the day, is becoming more attractive for advertising. If radio want to be around decades from now, they are going to have to invest in data, and the right kind of data.”

– Jeff Greenfield

Wayne Stamm for RadioInfo
I was really interested in one of the things you were saying today. So there’s been a real change in what’s happened. We’re seeing people coming back to radio now?

Jeff Greenfield
That’s absolutely true. They found that digital doesn’t have the full ‘opportunity to view’ that traditional media has.  In traditional media (Radio, TV, Print) you buy something, and you pay for it when it runs. And if it doesn’t achieve the ratings, or the promises that are there, there’s make goods that are involved with that.  That’s been the history of all traditional media. If you pay for a radio spot, and they forget to run it, or there’s a big announcement, and it doesn’t run, you don’t get charged for that.  But in the digital realm, what happens is that just because of how the internet is designed, you have to scroll to get content.  And so if you buy a million ads, half those ads are going to be down below the fold. And you still have to pay for those.  Now, digital has kind of shifted around and fixed their problem because now the currency of digital is based upon an ad has to be in view. So they have a make good system – but that just came into being in the year to year and a half.  And so digital is really getting its footing.  But what that has done is given brands time to kind of play around in the digital realm.  And what they found is that digital tends to be and of course, digital doesn’t want to be this, but digital tends to be a lower funnel medium.  So digital is very good at getting a message across right at that moment of purchase or right before that purchase. It’s kind of like the end cap at a supermarket aisle – they’re very, very good at that.  But they’re not good at the upper funnel messages.  I always like to to remind clients who get all excited about YouTube and how great it is and how it’s a replacement for TV that Judge Judy has more reach than all of YouTube.  Network radio and radio in general has a tremendous amount of reach.  And that’s why advertisers are going back to traditional media – because what they have found is that they are moving back to TV at night when people are at home.  But during the daytime, the way to reach people is that more people are streaming radio while at work.  You notice it every day you go into work. You notice people have got these ear pods in these earphones on and what are they doing?  They’re listening to radio.  And what P&G has found, because P&G is one of the biggest buyers of network radio. In fact, several quarters ago, they bought up everything that was available, because the best way to get out their message during the daytime hours is radio.

Wayne Stamm for RadioInfo
That’s an interesting story on its own. So you’ve you’ve had situations where some of those companies that bought digital early in the piece and we’re spending big on it decided that they’d take a closer look at what they were spending and how that was working.

Jeff Greenfield
Yeah, they did. They found that a couple of things. One is that they were buying digital across several hundred thousand sites.  So they wondered what would happen if they eliminated all except for 5000 sites – it obviously would have an impact, right?  So they eliminated all except for the 5000 most important, cut their spend dramatically, and there was no impact.  So then they cut their cut their digital budget in half, and still no impact.  And it’s because for a lot of brands that are spending on radio and on traditional media, you have these long ‘consumer journeys’. Meaning the timeframe that a consumer is in market is very, very long, and the ‘in-market purchase frequency’, how often a consumer makes that purchase, is not something that they do every week, not something they do every two weeks, it’s maybe something they do only every couple of years.  And radio is a perfect element to do that.  Digital not so much.  Digital is great when they’re searching for something, need a quick answer, but it’s not good for these long consumer consumer journeys.

Wayne Stamm for RadioInfo
Now the biggest problem with radio at the moment is I don’t think that they really understand data particularly well, and what sort of data that they should be supplying? We’re still stuck, I guess on the rating situation. And how many people listened last week? It’s not all that helpful, I guess.

Jeff Greenfield
Data can be your friend, and it can be your foe.  The biggest problem that radio has is that and the one reason they have yet to really go all in on, on investing in data is the fear about what happens if we do this study, for customers to find out the impact that we have? And we find out that it’s not as great as we once thought it was, and we’ve done all these studies in the past, we know that there is an impact.  But when we start going in and we start providing this type of data results at Facebook and Google have that a real time what happens if we find that all of a sudden it worked a moment ago, but it’s not working right now?  How do we respond to that?  Are we able to bob and weave the way digital has?  So that’s Radio’s big problem.  It’s time for us to put on our big boy pants, radio!  Radio has to grow up and realize that if they want to be around decades from now they have to invest in data, they have to spend millions of dollars.  And we’ve seen this already happened with some of the big networks.  But what they’re doing is, in some cases, I hate to say it, they are deceiving advertisers. They’re using data in the wrong way. They’re looking at some correlation. And correlation does not always equal causation.  You’ve got some big radio networks that are trying to take credit for everything that happens on a website from point A to 20 minutes and in some cases, 30 minutes after, unless a radio monologue.  Comcast right now is out with a product where they’re trying to convince advertisers that 30 minutes after one of their TV spots airs on Spotlight, that they should get credit for everything that happens.  And that’s just garbage!  What you need to realize is that data will help you, but you can’t just take credit for everything because this is no longer the days of taking credit for everything.  Credit now has to be shared.  This is the days of fractional credit.  That’s what attribution is all about.  You have to attribute credit and you get a little bit of piece at a time  That’s the way it works.

This interview was at the NAB in Las Vegas and originally appeared in Radio Info

Provalytics CEO Interview on Bloomberg Technology

Shery Ahn
Great record sales for Alibaba and yet investors don’t seem to be impressed. The stock actually fell and Baird now saying that sales saw a meaningful deceleration from 2018 levels. What’s your take?

Jeff Greenfield
It’s definitely a deceleration compared to previous years, but the key is that if you look at the growth since they started pushing on Singles Day, it’s absolutely incredible.  We shouldn’t be too impressed with the dollars, the big story here is the data: how they’re able to do it and when you look on comparison – when you look at the number of people that Alibaba shoppers have and you compare that to the US population, Prime Day and Black Friday it is actually not that far behind. The numbers with Alibaba are much much larger, but there’s so many more people.  The big story here is the data and what happens with the data for retailers in the US.

Kurt Wagner
I’m curious when you talk about that data. Give us a sense of why that’s so valuable to Alibaba especially in the growing field here of e-commerce players. What does that kind of data give them that someone like an Amazon for example might not get through prime or do they get that through Prime Day?

Jeff Greenfield
They do get it through Prime Day. The key here is understanding how that data has an impact for advertisers. There’s a lot of advertisers here in the US and a lot of big companies that traditionally don’t have access to that data. For example, CPG companies – like a cleaner or soap company – they don’t have access to that data and advertisers are finally waking up and that’s why you’re seeing that a lot of advertisers are now jumping onboard and realizing that in order to get that data they need to have a relationship with the consumer. If you look at Tide for example – they have gone and started acquiring cleaners around the country so they can actually have that ‘one-to-one’ relationship.

 

For a lot of people in this country, their local cleaner is now called Tide Cleaners.  The other side to that data and why it’s so important, is that consumers today want a curated experience. Consumers don’t want to just walk into a mall and have everything that’s for sale there and that’s why we’ve seen companies like Stitch Fix start to explode like crazy.  This idea that you can get exactly what fits you, exactly what you need, what’s made for you, to come to your home every single month – that’s what consumers today want and that’s where retailers really seem to kind of miss the message.  If you walk into a mall today – you’re really walking into a graveyard of brands that have not caught up and are not taking advantage and looking forward with business intelligence.

Shery Ahn
When it comes to Alibaba though, there’s always some questions about what the Single’s Day sale means for their bottom line. There has been some accounting scrutiny over Alibaba here in the US, so what do we know about how much this actually contributes to their business?

Jeff Greenfield
That’s a great question – because at the end of the day in order to get that many consumers to purchase that large number of goods, there’s a lot of deep discounting that goes on. Across the world, consumers all still want one thing – they want a huge bargain and that’s why sales are so big. That’s a big question that we’re not going to find out today in terms of what it’s going to impact their bottom line – but believe me there’s a lot of discounting that went on there.

Kurt Wagner
You talked a lot about making sure that this experience for consumers is moving towards this idea of being personalized. How do you kind of align that with what we’re seeing from social platforms like an Instagram or a Pinterest that’s trying to be both spontaneous and also personalized at the same time?

Jeff Greenfield
Social platforms are a whole different aspect in terms of data. The big problem these platforms have is they can personalize as much as they want but they’re not actually selling anything so they’re missing that link between commerce. Think about it in terms of Facebook – the only thing they’re selling are ads. They’re really just selling your information. Google themselves are also not selling anything at all – all they’re selling is your information. They’re missing that link that Amazon actually has.  Amazon is in an amazing and powerful situation because not only do they have that data, not only are they selling ads, but they also have goods that they’re able to sell.  It wouldn’t surprise me when you start to look at Facebook and Google with the amount of cash they have, that they would be looking to make an acquisition for some shopping platform. They need to have a direct connection to commerce in order to really take full advantage of all the data they have.

Shery Ahn
How are we expecting the Singles Day performance to help when it comes to that Hong Kong IPO share sale of 15 billion dollars that could come any day now for Alibaba?

Jeff Greenfield
Investors should be somewhat impressed with those numbers. They probably aren’t that impressed that the growth wasn’t as high as the years before, but I think what’s going to be interesting more than even what happens with Alibaba is how does this impressive number with Singles Day – what does it do for the economy here in the US and what what’s going to happen in Q4 as we move towards the shopping season here?  Are we going to see a 25 percent jump? We should.  If we don’t see as much of a jump what that tells us is that Alibaba is really taking better advantage of their data than the retailers here in the US.

Shery Ahn
It’s all about that data and analytics and Jeff thank you so much. That was Jeff Greenfield joining us from Provalytics. This is Bloomberg.

 

What is the impact of losing 3rd party cookies on MTA?

Google’s announcement about the removal of 3rd-party cookies initially sent shockwaves through the advertising world and Wall Street.  Is this an Attribution Apocalypse or a necessary cleanse to the advertising ecosystem?

The changes and potential impact was the topic of an industry conference call held with Shyam Patil, Internet Analyst, at Susquehanna Securities and Provalytics CEO, Jeff Greenfield.

Shyam Patil:
I’m Shyam Patil, the Internet Analyst at Susquehanna. Welcome to our conference call with Jeff Greenfield, CEO of Provalytics and the former Chief Attribution Officer at C3 Metrics. If anyone in the audience has any questions, please feel free to email them to me during the call. What we’re going to try to do with the structure of this call, we’re going to try to just go through what Google is doing, the time frame, the near term impact to companies. Then we’re going to talk about what happens in a couple of years, talk about some of the specific changes that are going to happen then. And then the impact to companies in the ecosystem at that point. Jeff, to start out, can you talk about exactly what Google is doing and what the time frame is?

Jeff Greenfield:
The big announcement that Google made, is that they will be blocking third-party cookies. This is very similar to the change that’s been made both in Safari and Firefox previously. To really understand it, let’s step back a little bit and talk a little bit about the difference between a first-party cookie and a third-party cookie, because that’ll help put a frame of reference on things.  Cookies are part of computer code. They’re how computers and computer language talk to each other. It’s part of JavaScript, it’s part of PHP. This is part of how things operate. So, it’s fascinating that these browsers are actually getting involved in blocking stuff like this.  A first-party cookie is when you go to a website and there’s a certain level of personalization that’s there. You go to Amazon and you’re automatically logged in and they say, “Hi Jeff.” Or you go to the New York Times and you’re automatically logged in. So first-party cookies are cookies that are set by the page that you’re actually on. And those are what make the digital experience an ‘actual’ experience. It’s what makes it comfortable to you and it’s not like you’re starting from ground zero.  A third-party cookie is a cookie that’s set by someone that is not the owner of that website. So if you go to the New York Times and there happens to be an ad there for a shoe company, there may be a cookie that’s set by that shoe company or that is set by that ad server. That’s a third-party cookie, because it’s different than the site that you’re actually on.  The whole digital experience has been about this combination of first-party cookies and third-party cookies and the whole advertising ecosystem, including the ability to target, has been built on third-party cookies. That ecosystem is extended by what you would call a cookie sync. Where, somebody who has first-party cookies syncs with somebody else who has first-party cookies and then they come up with a ‘secondary’ cookie. This secondary cookie would allow them to find individuals based upon their preferences, what type of stuff they’ve bought, maybe their age, maybe some other demographic information. And that’s been the entire basis of the advertising ecosystem.  When Safari and Firefox did away with third-party cookies, it was a hit, but it wasn’t that big of a hit. Primarily because most people around the world tend to access the digital landscape using a Chrome browser. Most companies just shifted most of their targeting over to Chrome, advertisers accepted the fact that Safari was tough to target, but they didn’t worry about it because the vast majority of people were using Chrome.  So this announcement is a big deal.  It’s a real big deal to a lot of companies, and it’s also a huge deal to brands and advertisers. Because in their world, they don’t know what’s going to happen next. It’s forced a lot of questions, and it’s causing a lot of people to start to rethink strategies.

Shyam Patil:
You talked about this a little bit, but maybe to go into more detail. What are the big things that happen next? Maybe the specifics there. Maybe you can weave this into your response is just how will user tracking and targeted advertising be done at that point?

Jeff Greenfield:
That’s where things get really interesting. So remember, the entire digital ecosystem has been based upon these third-party cookies. And the reason that they’ve done this is because, let’s say my company is jeff.com. And I’ve got ads that appear across all of these different websites. So if you see an ad on the New York Times and it’s from the jeff.com ad network, I’m going to set a jeff.com cookie that’s a third-party cookie. Now when you go to boston.com I’m going to be able to see that I just saw you on the New York Times because I’m going to be able to read and write that jeff.com cookie. I can’t see any New York Times cookie because I’m not on the New York Times page.  So right there is the importance of that third-party cookie – it allows me to see you across all of these different sites. You may read about it where it’s called ‘cross-site tracking’, but it’s part of the nature of third-party cookies. That aspect is going to go away. Companies have all tested and tried different things. Going back in the early days, people were using flash cookies because everybody in the world had Adobe Flash on their systems. And the greatest thing about Flash is that you can put information in Flash and it would always be there. You could always find it. And it was cookie-less, which was great.  Of course Flash is not supported anymore and it stopped working and automatically getting called in the browsers. So that was gone. So then what companies did, is they did what a lot of the fraud companies do, which is digital fingerprinting. And a digital fingerprint looks at a combination of different aspects. So your operating system, your browser and versions, all of that data, all the way up to how many different fonts you have installed on your system.  And what they found is that if they could pull like eight or nine of these different aspects out, they could do a very good job at identifying you on a statistical basis. They wouldn’t know it was absolutely you but could do a good job of identifying a unique computer. Safari and Firefox in particular have done a really good job at stopping that practice and Chrome is also doing that as well, so digital fingerprinting is gone too.  There’s other methods that companies utilize. One of the big building blocks of the internet is the addressable aspect, so that every computer has to be coming from some sort of address, what they would call an IP or an internet protocol address. All IP Addresses are unique and you can link these back to specific households and to specific companies as well. Where it gets difficult is that, if you’re part of a company, everyone at the company appears to becoming from the same IP address. So that makes targeting a little bit difficult.  A lot of companies in the advertising ecosystem are going to start utilizing some sort of proprietary identification, which will be based upon location with some aspects of early digital fingerprinting – such as operating system, browser, etc. In addition, many companies have direct relationships with the sites in which the ads are being served. Criteo is a great example. Criteo works with almost 20,000 publishers. One of the unique selling prospects of Criteo is their direct relationships.  If you go back to the birth of this whole digital ecosystem, it was all about the publishers, the Yahoos of the world. They controlled the ad buying experience for marketers. All of a sudden we got into the programmatic world where it didn’t matter who you were, because everybody was providing inventory for programmatic and we moved away from direct relationships. Back in the early days, every publisher had their own sales force to sell directly to advertisers. As programmatic grew, direct sales (also know as IO sales) started to shrink.  We’re going to see a change back to where publishers are saying, the inventory I have is unique. And the reason for that is because, I know who my readers are, I know what they read, I know where they live, I have all this demographic information on them. And it’s more valuable to me to have a direct relationship with people with advertisers and a direct sales force, versus selling everything via a programmatic exchange.  It doesn’t mean the exchanges are going away, but we’re going to see a move more towards more private exchanges and a move back to direct relationships. With a direct relationship, Criteo can get data from a tracking perspective pushed back to them that will bypass Chrome. And it does that through an API where the publisher themselves will send information back directly to Criteo servers and that’s what’s going to happen.  And what that allows both the publisher and someone like a Criteo to do is to say, “Hey, we’re going to live in our own kind of world, in our own existence.” Because this is the first of many changes to come from these walled gardens. Chrome is not the way that we access the internet. Chrome is an extension of the Google and the Alphabet business strategy. Changes to Chrome at the end of the day, they are always going to be Google’s their best interests. Right now it’s being said that it has to do with privacy, but the reality is, is that this is part of a larger business strategy for Google. And so the key way for tracking for these companies is going to have direct relationships. That’s going to be their best method. Companies that don’t have direct relationship and don’t have scale, they’re going to struggle. They’re definitely going to have some problems.

Shyam Patil:
Cross-site cookie tracking. Can you talk about why that’s important now, and then what impact this will have, and who do you think is going to be impacted the most?

Jeff Greenfield:
Cross-site cookie tracking is the purpose of these third-party cookies. It allows me to track you across as many different sites as you go. As long as my ad is there and my code is there, I can track you all the way across the entire ecosystem, wherever you go. And what that allows me to do, is it allows me to make sure that I’m giving you the right frequency, that I’m not showing you 80 ads a day. It allows me to write a better experience from a brand perspective, which I know most of you on the call are probably laughing about, because that’s been one of the complaints the longest time about the web is that the experience should be incredible. With all of this technology, I should have such a positive experience with brands.  But typically what ends up happening, is a crappy experience.  I go to a place and I make a purchase – so they know that I made the purchase and their partners should know that I make the purchase, but now I get bombarded with ads for the next two weeks. And somebody is wasting somebody else’s money.  In the early days of the internet, efficiency was really, really important. And whenever someone would make a purchase, they used to call it an ‘unpixel’. It was a special piece of code that got fired to let all the partners know that this person made a purchase, stop showing them ads at least for 30 days. It depended upon the frequency of purchases. But don’t show them anymore ads because they just bought today. They’re not going to come back and buy again, but that doesn’t even happen these days.  So it’s funny because, this is going to impact the ad experience, which I think actually needs an overhaul and that’s what consumers really feel about it. That’s going to be the biggest impact is the ability to deliver ads in a targeted manner. And what’s happened is, is that you can get really micro targeted. And a lot of advertisers in a lot of brands, they get really turned on by this. Because I can actually go right now in the world of cross-site cookies, and I can go and I can find individuals who have purchased with the competitor within the last three months, and put it on a major credit card. Now none of that stuff is online, but what’s happened is companies like MasterCard and Visa, they are taking their data they have on consumers. They are matching it into a cookie pool to sync up. When you go and buy ads, you can buy segments of individuals.  And that sounds like wow, the more targeting you do. So you can find people whose lease is about to expire on their truck and this is the third truck that they’ve owned. You would think, “Wow, I’m going to target these people. My sales are going to go through the roof.” And yes, these people do buy. But the problem is this – not only do you have the cost of the ad itself, but now you have the cost of the data on top of it. Now all of this operates in this world of third-party cookies. And the cost of that data sometimes triples the cost of the actual advertising.  And what a lot of advertisers have found is that, this is a really cool thing to talk about. People get very, very excited about it, but the cost is so obscene that it makes profits kind of disappear. So yes, I can acquire customers, but it’s at a cost that is too costly for me to continue to do that strategy. And that’s part of the problem.  So that process is going to disappear. Those companies that are out there that have this data, they’re now going to have to find another way, in order to link that data up. It’s going to impact the ability to target. It’s going to be gone unless they find other ways and it’s going to impact the ability to deliver on the frequency. So all of a sudden, if these companies didn’t figure things out, you would start to see tons more ads or you wouldn’t see any ads at all. Obviously you’re going to see ads, so chances are you’re going to see a lot more ads.

Shyam Patil:
I’ve got a bunch of questions from the audience, but before I get to those, I wanted to just talk one more change just conversion measurement. You talked about this a little bit earlier. There’s also a change that Google made recently. Can you just talk about what’s happening with conversion measurement in two years, and what’s going to be the impact?

Jeff Greenfield:
Conversion measurement means the actual outcome. Advertisers buy ads because they want to lead to some sort of conversion. In the world of e-commerce, it’s an actual purchase, in the world of B2B and branding, it may be some other KPI (key performance indicator).  Most companies that have operated in this space, when they sell ads, they typically provide two tags. They provide a tag to put on every single page on the site, which is a retargeting tag. So I know that when people have been to the page, I can go and find them and I can find them on a cross-site cookie basis, so that’s going to go away. And the other piece is a tag that goes on the conversion page or the action page where whatever your KPI is and I provide a tag to go there. That also historically has been a third-party cookie as well too, so I can identify these people that have actually purchased.  A lot of companies have switched to first-party cookies.  Facebook has done that.  They primarily did that to get around the whole Safari change so they could still get conversion information. This is definitely going to impact the smaller companies out there – they are going to have to update to first-party cookies so they can actually get that information. I see them definitely doing that. Part of the larger picture of conversion tracking, is this movement of the digital landscape impacting sales and the retail world off of the digital world.  Because remember, most of commerce still happens in the brick and mortar. And we’re starting to see more and more brands move into brick and mortar. We saw Amazon move and purchase Whole Foods. We’ve seen a lot of strictly a digital brands do the same – Warby Parker is one of them. They keep opening retail shops. Indochino, a suit brand for men, they have retail shops as well. You don’t actually buy anything in the shop, but you try things on and then you make your order digitally.  So the ability to link those two together for outside third parties, that is typically been a link that’s been done by syncing a series of third-party cookies together, that’s going to be impacted. That’s incredibly important, because more and more you have companies like Facebook and Google that are wanting to demonstrate that buying ads in the digital world, the virtual world where they live can impact sales in store. You buy ads on Facebook, people are going to go buy furniture at the Ashley Store. Or if you advertise, we can send people into Home Depot. So being able to connect those two together historically has been done through third-party a lot of times utilizing LiveRamp or using some sort of proprietary tool without those third-party cookies, they’re going to have to try other ways to link together.  The other impact is a blockage into the app world. Think about it, if I’m a brand and I buy an ad on Google on page search, I’m usually driving people to my website. So when somebody clicks and they come from Google, I can see that they came from Google, so I know my clicks are working. When somebody goes and they click and I’m sending them to download my app to the app store and when they’re there, how do I know that click actually happened? And then how do I get information out of the app store?  So there’s certain app providers, one of the largest is AppsFlyer. Traditionally what they do with someone like a Google or an Apple is they say, “Hey, when that user clicks, send me a click at the same time. Either redirect through me. So when someone clicks, send them to AppsFlyer real quick and then we’ll redirect to where they’re supposed to go. Or send me a click at the same time.” And so what Google is now said is, “Hey, we’re not going to do that anymore. We’re not going to send you information on every single click. We’ll send you information for people who actually downloaded something. When somebody converts and they download an app, you send us the data you have on them and we’ll tell you what they clicked.”  So it’s moving to this model where these walls of the walled gardens are getting higher and higher, and they’re saying, trust us. It’s important in business to trust, but you also have to verify and you have to validate. And when those walls are up, it’s impossible to verify and validate. And that’s the problem.  And the app world is not going away. The app world is getting much, much bigger. It’s billions of dollars a year and lots and lots of money is being spent in the app world, to not only to get people to download the app but get people to utilize the app. And so this is another way for the walls to come up from Google.

Shyam Patil:
Can you talk about companies going from being third-party to being first-party cookies? Can you just talk how companies are able to do that? And then I have a few follow ups on that.

Jeff Greenfield:
It’s just a matter of how the computer code work. The key is, is that in order to write a first-party cookie, it can only happen through JavaScript that sits on the side as a piece of computer code, so you have to have JavaScript.  A lot of companies historically when they traffic out their code to go onsite, it’s typically what they call a one by one image tag. Image tags can only write third-party cookies. An image tag gets called, it doesn’t interact with the browser and it’ll call my jeff.com information and that’s it. But if I put a jeff.com piece of JavaScript code on the site, jeff.com can also, if I’m in New York Times, jeff.com can read and write New York Times cookies, and I can also read and write jeff.com cookies. So it lets me do both first-party and third-party cookies.  For years, Google analytics was third-party cookies, because that’s how the whole advertising ecosystem worked was on third-party cookies. They moved to first-party cookies a number of years ago. Facebook, when Safari did away with third-party cookies, Facebook quickly moved to first-party cookies. They still have the option from a privacy aspect I believe, where your pixel that you put on the page can be third-party cookies. I’m sure that’s going to be go away and it’s going to be all first-party.  So it’s just a matter of moving from a one by one image tag to a piece of JavaScript. and so it’s just a matter of re trafficking, which is a little bit of a workflow. But over the next two years, that’s not going to be an issue for a lot of folk. But they’re going to want all of the members of this ecosystem, will at least be wanting to get that first- party data themselves when they’re on a publisher site. So they’ll mostly be writing first-party cookies.

Shyam Patil:
Another question. So sites like Facebook, Snap and others use pixel for targeting remarketing attribution, however those guys impacted by the Facebook as a pixel on Zappos and they can use that today to remarket on their app or tell whether there was a conversion, will that still be possible in a couple of years? What will happen there?

Jeff Greenfield:
And that example with Facebook having a tag on Zappos, that tag will be a first-party cookie, which will send information back to Facebook on that user. So they’ll still be able to target people using first-party cookies because remember, they have that relationship with that site. Because remember, it’s Zappos that went and said, “Hey, I want to enable this behavior so that I want to retarget within my Facebook feed.” So they’ll be able to send that information back and forth.  Even if Google came along and decided to do away with that, then Facebook would set up an API, where Zappos would send information server to server. So that when someone landed on the Zappos site, there would be a little piece of computer code that would send information directly back in the background from Zappos to Facebook, essentially would bypass the whole Chrome and Google environment.

Shyam Patil:
Another question is, you talked about the direct relationship structure. You’ve talked about it quite a bit just now and then you talked about it earlier with Criteo. What’s the mechanics of that? Do you have to go one by one to the publishers? How many direct ties will retailers and publishers just generally have?

Jeff Greenfield:
Someone like a Criteo, when you start to see the shifting landscape of what’s happened, remember years ago I was a publisher and I would write great content that would get spidered and people would find me. And I would typically get found in the yahoos of the world and now the Googles of the world because that’s where people went to go find stuff. And Google because they’re constantly changing their business strategy over the years, when you now go to that Google page, 80 to 90% of what shows up on your screen at any one time is ads.  In fact, it was just a shift that was made with how now the ad display, the little picture next to the ad where it says ad is now smaller than it ever was. The ads aren’t a different color. It’s almost impossible to tell the difference between the ad and the organic content. And so now if I’m a publisher, the chances of me getting found in Google and counting on that strategy is nonexistent.  In fact, we just read about a local Boston company, TripAdvisor, because of a Google update they’ve been now moved to the second page. Publishers can’t rely upon a strategy anymore and they’re starting to realize that folks like Facebook, where five years ago publishers built up a whole Facebook strategy, they started working with them and then Facebook’s business strategy changed as well too. They’re realizing that they’re an island on their own. And they need to build content and make content that is right for their readers. They need to understand who the people are that are coming to visit them. That’s the whole purpose behind the paywall besides getting revenue from it, the real purpose is, that the more I can learn about you, the more I can create content for you and deliver the data that you want.  And then an extension of that is having a direct relationship, not only with the brands that I work with, but also the partners and having a sharing relationship so that we can both thrive in this world of wall gardens. Essentially creating my own little sort of walled garden, if you will, because I can’t count on the ever evolving business strategy of Google and Facebook to deliver the traffic that they used to. I need to find ways to create my own audience and curate it.  And Yes. To answer your question, you have to go one by one to each one of these. The larger ones is just like an advertising relationship. They’re meeting with them on a regular basis. There’s two sides to the house there at Criteo, there’s the side that’s selling the ad, and then there’s the side that’s developing these longterm relationships. I’ve long said for many years that Criteo has thought it was a one trick pony, because all they’re doing is they’re delivering these dads to advertisers. But as their strategy continues to grow and evolve and the strategies of Google and Facebook evolve, and these relationships that they have with publishers where they’re actually delivering revenue to these folks, I could see where they could expand even more so, and provide even better tools for publishers that help improve their revenue.  And I think that could be an interesting growth area for them, because that’s something that these publishers desperately need and they need a partner and Criteo’s proven to be a partner for them. So that’s a separate side of the house that most people on the marketing side don’t see, but they definitely have it in order to keep those relationships as strong as they are.

Shyam Patil:
It sounds like in the new framework with privacy sandbox on the Google that you don’t expect much impact to one-to-one retargeting. You don’t think guys like Booking.com or Criteo are going to be impacted, because the likes of Criteo will be able to get their first-party cookies embedded in publisher sites. Is that a fair summary?

Jeff Greenfield:
Yeah, and their ability to create their own unique proprietary identifier. They’re also sitting on top of massive amounts of data. They’ve got the smartest and the brightest minds in the world. And not only that, it’s not like Google is pulling the rug out from under them and saying, “Hey, this is happening on Monday, guys. Good luck.” Yeah. If that were to happen, that would definitely be a problem. But two years, two years in the world of digital, that’s like 25 years in the world of brick and mortar. That’s a lot of time. A huge amount of time.

Shyam Patil:
Maybe more specifically on Trade Desk, had another question come in. It doesn’t sound like you expect much impact at Trade Desk, because of the talent they have as well as their unified ID, their own identifiers. Specifically on Trade Desk, is that going to be their solution, the unified ID that they have? Is there something else? And do you think they could lose share to Google’s DSP? Could Google’s DSP be advantaged by in any way by this? Or do you think they could lose share to Amazon’s DSP? Can you just talk about what Trade Desk solution is going to be or what you think it might be? And then, if they could lose share in the landscape to either Google or Amazon?

Jeff Greenfield:
The biggest issue that Trade Desk is going to have is right now. And it’s the confusion this has created across brands and across advertisers. It’s forcing a lot of them to start rethinking their strategies.  The problem is for a lot of them, they’re utilizing the Google ad server and they’re part of that stack, so it’s always a question of do I go all in on this? Or is it better to use different components? Trade Desk over time could actually gain share, as long as they handle the PR aspect, which is to instill confidence in advertisers that, “Hey, we’ve got our unified ID. This is going to work.” You want to be independent. That’s the biggest thing is that you don’t want to be dependent on Google to deliver everything.  In terms of Amazon, Amazon has done a really good job of enabling that search. People tend to forget how powerful search is and why Google became the pipe that it is. It’s because their search functionality was so much better than Yahoo’s. So now start to think, when you go to search for a product on Google and the results that you get versus when you go to search on Amazon and the results that you get, what is more visual? What provides you better information? You only have to do it a couple times and then all of a sudden your behavior starts to change. Most people probably start to search first now and Amazon. And so as a result, that’s a pipe that Amazon hasn’t really started to monetize just yet. They’re just starting to do a little bit. It’s like the early days of Google when they started a first putting up their ads.  The difference that Amazon has and the value that Amazon has over Google is not only can they put up little ads and start to monetize it just a little bit, but once you’re using that pipe, anything you buy from there, they get a piece of it. And not only that, but sometimes they can move you towards, after you’ve clicked on some ads where they’ve made money, you may end up buying a product that they own where they even make a bigger share. And that’s the big piece that the Facebooks of the world and the Google are missing, which is that end piece of commerce. That’s a huge piece that they’re missing that one or more acquisitions that they can make, that would really be a game changer here for them.

Shyam Patil:
So we’ve talked about Trade Desk, we talked about Criteo. LiveRamp, what do you think is going to be the impact to them from this change?

Jeff Greenfield:
LiveRamp right now, they’ve got this identity graph which is pretty incredible. It’s incredible in the sense it’s kind of like Criteo in the aspect that Criteo has 20,000 publishers that are part of their network that believe in Criteo. And LiveRamp has every single member of the digital ecosystem contributing data. You can now take our identity that we establish for our clients and link it up with a million other things. It’s absolutely incredible. And the value that they’ve created is that they’ve gotten everybody together.  LiveRamp historically is operated based upon third-party cookies but they’re going to have to evolve that strategy. And again, you’re talking some of the brightest minds there. I don’t doubt them at all. Primarily because every single member of the ecosystem from every single advertising technology company, to every single major brand is contributing data into this.  What’s interesting is who’s not contributing data? It’s the other walled gardens. Google will not allow any tags if they’re carrying a LiveRamp tag. So the walls have been up against LiveRamp for a very, very long time. So most people see the ecosystem where LiveRamp is its own separate walled garden. But the difference is, is that they are a walled garden in that aspect that they’ve gotten everybody to buy into it and contribute data. So incredibly powerful.

Shyam Patil:
So you don’t really see much impact to Trade Desk, Criteo, Ramp, and Facebook, Google from the changes. Is that a fair summary of your assessment of what’s likely to happen in a couple of years from this change?

Jeff Greenfield:
Yes. I’ll add in one piece. One aspect that is going to be impacted has to do with, brands do a lot of what they call brand tracking, which has brands like to know based upon their large global advertising, how do you feel about us? What’s the likelihood that you would refer someone? What’s the likelihood that you would buy? Brands do this and they spend multi-millions of dollars each year. These brand trackers are typically done via surveys. You may see them where the survey pops up on a website. And the way that they do this is that they know whether you got exposed to an ad or you didn’t get exposed to the ad, because they’re tracking technology is there.  There’s going to be an impact in the ability of brands to track. And where are they going to be able to do that really well, is that there are companies out there that have people that have opted in where they’re allowing the survey company to look at everything that they’re doing on the web. Every bit of browsing behavior is being passed to them. Even though they may be using Chrome. That’s a separate thing that gets on and Chrome can’t have anything to do with it. And these individuals, they also have it on their TV sets, so you know what they’re watching and they’d have it on their phones too.  The company out there that has one of the trackers if you will, of people that have opted in for that has been Comscore. That’s going to become very valuable and there may be a lot of survey companies that are going to start utilizing their group of people, because that’s going to be incredibly important is to have a group of folks that have opted in that allow everything to be looked at so that they can check on what their impact to the brand has been if they were exposed versus that they were not.

Shyam Patil:
Awesome. If anyone in the audience has any further questions, send them over to me and I’ll make sure I get answers back to you and thank you Jeff so much for taking the time to join us today and share your thoughts. This is the very interesting topic that we all have a lot of questions on, and we’re all trying to figure out.