811 | The Godfather of AI attribution says Your Marketing Dashboard Is a Lie | Jeff Greenfield

Most marketers believe that we invest dollars to buy clicks and clicks lead to leads or sale, which is wrong. That’s Jeff Greenfield, co-founder of Provalytics, an AI-powered media agnostic attribution company that’s rewriting the rules of marketing measurement. He built one of the first multi-touch attribution platforms back when cookies were king, then watched the whole thing collapse. Now, he’s back with a new model that doesn’t need clicks, it doesn’t need cookies, yet still proves ROI in a world where nobody clicks. But what’s the secret? The greatest thing about AI is it enables us to see what we can’t see ourselves.

Marketers spending $2 million a year can definitely get ROI with this approach, but it will also work for a florist spending $300 a month. In this episode, you’ll learn why cookie-based tracking is collapsing and what replaces clicks in a new attribution model. If you’re measuring using clicks, CTV will not never, ever work for you. Digital out of home will never work for you. Meta, you’ll never be able to scale. We actually want to be focusing on… Next, you’ll learn how AI is towering a smarter cross-channel measurement system. AI is able to run hundreds of thousands of different what-if scenarios. What if you did this? What if you did this?

That’s why brand building is a new performance marketing and ultimate attribution blueprint for agencies. It’s an upper funnel that actually builds the awareness in that mid funnel that reminds people, hey, remember, we saw each other a while back.

Clicks aren’t giving us the full picture now. So how do you tell people that we should be thinking about measuring marketing impact? Let’s say, Jeff, that you’ve got this great new hose that you’ve created and you decide to put it online, you get a Shopify store to start selling it and you run ads on Meta, so Facebook and Instagram. And then you’ve got an ad in Google running. So when people search for it, they see it. So someone comes along and they’re scrolling through the Facebook feed and they see the ad and it looks cool, but they don’t do anything. They keep scrolling because that’s what people do. A couple of weeks later, they’re out doing some gardening, turn on the hose and it’s leaking like crazy.

And they say, oh my God, should have gotten that hose. What is the name of that hose? And they pull out their phone, they Google it and click on that ad and they make a purchase. Let’s say at the end of the month, you had a thousand sales. So if you were to log into Google, it would tell you you had a thousand sales. And if you log into Meta, it would tell you that you probably had 800 sales. So that’s 1800 sales versus the actual thousand. So that’s the famous duplication issue. And what ends up happening is that if you’re advertising, let’s say across five or six different separate platforms, none of them know about the other. So they’re all vying to get as much credit as possible.

So what ends up happening to the market or at the end of the month who has to make sense of all of this, somehow they have to have a magical formula to get the numbers to actually add up. And they never do. The other problem that happens is that due to people utilizing Google Analytics 4, which is the free analytics that 99 percent of companies around the world utilize, that is a click based focus, which means that you’re you’re looking down at that lower part of the funnel towards that action part of things. And so as a result over this last 20 years or 30 years of this digital journey, what’s occurred and what’s being left off is this concept of branding or upper funnel.

And it’s an upper funnel that actually builds the awareness in that mid funnel that reminds people, hey, remember, we saw each other a while back, maybe you want to check us out again. But most companies are way overinvested down at the bottom. Most marketers believe that we invest dollars to buy clicks and clicks lead to leads or sales, which is wrong. Marketing, we know the way marketing works is we invest dollars to buy awareness, which is attention. We actually want to be focusing on impressions. Impressions are that awareness, that attention metric. That’s where we want to be looking versus the clicks. The clicks are the the after effect of the awareness.

We want to be looking at the awareness that’s actually driving those clicks. Impressions are what drive performance. It’s what builds brands and it’s what grows your sales. You want to use a metric so that when you finish through with your equations and you say, hey, this is what you should do with your marketing buy in order to improve things, you want to make sure it’s something that they can actually go out and duplicate. And regardless of what type of media you’re buying, whether it’s connected television, digital out of home, podcast advertising, even paid search or YouTube or any of those, it’s all impression based, all of it.

Even paid per click is based upon impression and impression share. So that’s that’s the first piece of it is that it’s something that you can go off and do and replicate. The second is is that it allows us to take not only digital marketing but also align it with all forms of traditional marketing. It’s a unifying metric that brings everything together, whereas clicks, the problem with clicks is that if you’re measuring using clicks, that means that CTV will not never ever work for you. Digital out of home will never work for you. Meta, you’ll never be able to scale meta like you’ve heard your competitors are. So you’re you’re you’re holding yourself back.

That’s why we want to live with impressions. You know how a lot of agencies will, you know, build reporting and trying to understand things. They’re pulling data from, you know, multiple different platforms. How is AI now helping, you know, sort of cut through that clutter and create a better picture? That’s the greatest thing about AI is it enables us to see what we can’t see ourselves. And it’s kind of like the example of like a car accident that happens at a corner and the police officer goes and asks one person what they saw and they go across the street. And that person saw something completely different. And so then the police officer has to kind of merge the versions together.

But the traffic camera up above sees everything. And that kind of AI, AI comes in and is able to make sense out of all of these different perspectives, primarily because it is it is living in the world of saying we don’t we’re no longer at that user level. That’s the kind of the first truth is that we can’t track people across everything. And like you said, in the beginning, we’ve been led down a path to believe that this buyer’s journey is this linear thing and it’s not. And so when you accept the truth that we need to pan the camera back and you train AI accordingly based upon it, it’s able to make really good sense of these different divergent views of what’s actually going on.

Determine where where should I be scaling the dollars? Where what piece is having the biggest impact? Where should I be investing more and cutting back on other? Well, that’s that’s the coolest thing about it. So going back to something you said earlier about how marketers will download a report from their platform that they’re in charge of and they’ll notice impressions, but they’ll look at clicks. And if it’s not performing, they’ll cut it. And what they don’t realize is that these impressions don’t live in their own world. You need to look at everything altogether. And so when you look at everything altogether and you look at it on a granular basis, not just channel, but channel, campaign, ad group, creative, you look at it that level and you look at it daily across everything you’re doing.

What you will notice is that even in the same campaign throughout the month, the impressions are never the same every single day. They’re always a little bit different. And you’ll also notice that the results, your sales or leads are never the same. They’re always a little bit different. Your website visits are never the same. They’re always a little bit different. And the gold standard for advertising measurement these days is to do a holdout test where you pick a pick a geo like a Dayton, Ohio. And you decide in Dayton, Ohio, it’s going to be the only city and you’re going to turn on CTV there. And and then you look at the results for sales just in Dayton.

And if you see sales go up a certain percent, then you know, CTV worked and that’s how we do it as human beings as we do these A.B. tests. The problem is, is you start to scale and you’re working across multiple channels. It’s impossible to do that at scale. But with AI, when we look at these marketing plans at that level of granularity over a course of a month, what we realize is you as the marketer are actually doing hundreds, if not thousands of micro experiments every single month. You just don’t realize it. Just those little changes in impression volumes every single day are enough of a signal for a sophisticated AI machine learning to pick up on and look at those differences to figure out what is actually having an incremental impact.

And that’s the key is which one is actually doing something to produce sales or leads that you would not have had if you hadn’t have actually done that. And that’s really the big secret with this is that there’s never been a marketing plan throughout the month where the spend is exactly the same. Well, the spend may be the same, but the impression levels are always a little bit different every single day. And so it’s that little bit of signal that the AI is able to pick up on. And then from that, it’s able to then run hundreds of thousands of different what if scenarios. What if you did this? So what if you did this in order to determine what is the next best thing to do in order to show you how to take that same amount of spend and squeeze out more ROI?

And that’s really the secret is that you can take your same marketing budget and get more leads or more sales. Then you can march into finance and get more marketing budget to play more. Now, I know a lot of people will say, well, well, you know, I’m only spending three hundred dollars a month. I can’t go out and spend thousands on a sophisticated measurement system. And I always say to them, that that’s right, you probably can. But you could take that Google sheet that you have, the one that has every day with the amount of money you’re spending with the clicks and the website visits and your cost per click. And you could go back and recreate that for the last 12 months and add that column for impressions.

And then once you do that, you could create a graph of impressions and then overlay it with clicks and then kind of stand away from your computer and pan back a little bit and start to look at those relationships, meaning when clicks went up, go back a couple of days to a couple of weeks to see when impressions were higher and then dig in on those days. It’s really that simple. You know, figure out what happened on those days and do more of that because you’ve been missing the big picture. You feel that branding is more important than ever. And I want to unpack that a little bit. We have a good 20 years of research. And what we’ve seen is that his brands have invested more dollars in that bottom of the funnel or what we call short termism, if you will.

We’ve seen the effectiveness of advertising start to decline dramatically. And there’s a couple reasons for that. One is when you spend more on the bottom of the funnel, you are decreasing the number of people that see your ads. Reach, which is incredibly important. Reach is one of the fundamentals. But more importantly, the messaging is different. The messaging at the bottom of the funnel, like you talk about these performance agencies is all about benefits, buy now, why you need to have it now. You know, here’s a discount. Whereas the branding at the top of the funnel tends to be more storytelling, tends to be more emotional.

And what we know is that when we have a chance to put kind of an emotional hook, if you will, where someone feels something about our brand, they may not be in market, but they get a good feeling about it. They remember us later on when they’re ready to buy. And so having that branding at the top has a massive impact on the memory in someone’s mind about how they feel about us. And so what we’ve seen, especially because of these TikTok style ads and the way that people scroll is we’ve seen even large brands have moved towards these really fast ads where you can’t even keep track of them. And but I I pray that we start to see a return towards storytelling because the brands that start to tell stories that focus on emotion.

Those are the brands that are going to win later on because those are the brands that people will remember when they’re in a buying mode. And so if you’re taking a look at branding campaigns, and most of the impact is actually happening happening before someone’s ready to click or before they’re ready to buy. What does that process look like of communicating the value to a client? Well, that’s the best thing about measuring via impressions versus clicks, because when you’re looking at impressions, you actually can incorporate branding into the whole picture and you can do something that all of digital marketing has not yet been able to do up to this point, which is instead of talking about branding in terms of, hey, it makes people feel good.

It lifts your brand awareness, which, you know, the truth is, if I’m a CFO of a company, I don’t care about brand awareness. I care about money in the bank account. That’s all they care about. And with incorporating impressions into your measurement approach, you can actually connect branding to finance. And that’s what folks really want. Also, what you’re able to demonstrate as well as you’re able to demonstrate that lag, if you remember when we talked about the hose earlier, there’s a lag between awareness, a timeframe, a time to conversion for each channel, for each campaign, for each product that it takes from awareness all the way through.

So imagine not only when you’re able to connect branding to money, we were also able to demonstrate that says, hey, campaigns on TV with television is taking 38 days, but it’s also lifting brand search. It’s lifting display. It’s lifting all of these other channels as well, too. That’s where the value comes in as being able to combine those together. So it’s less about saying, hey, this is performance and this is branding. And it’s more about having an overall healthy marketing mix because you want to fill your funnel and keep it constantly full. And so you want to go up funnel. The advantage as well of being at that top of the funnel. And here’s the big secret.

The lower you are down on the funnel, you are in this full time bidding environment. You are bidding against your competitors. And if you notice the cost per clicks, keep going up every year. They will because meta and Google, their profits keep going up. But the more you get to that upper funnel, the less competition there is. You still may be paying on a bid basis, but there’s fewer people up there. There always are fewer people, which means you can get more bang for your buck, if you will. I remember a client that we had that they were following the numbers. The marketing team was very experienced knew exactly what they were doing. But 92 percent of the budget was on non-brand search.

So on generic terms, category terms, because that looked like that’s what was working and they had tried connected television, I think nine months before they did it for a month and a half and it didn’t work according to them. Of course not. They were using Google Analytics 4. I don’t know about you, Jeff, but I’ve clicked around in Google Analytics 4 and there’s no tab in there for connected television advertising. It just it doesn’t show up. We were able to show them because when we go in and do essentially we start with an audit, we audit the last 12 months. So there’s great insights for customers to be able to look over the last year.

And for them, we were actually able to demonstrate to them that their their cost of acquisition for a new customer with connected television when they ran it was so much less expensive than it was for Google non-brand search. And we also were able to demonstrate for them that their cost of acquisition for Google non-brand was much, much higher than they ever anticipated.