The Best SEO Podcast: Unlocking the Unknown Secrets of AI, Search Rankings & Digital Marketing

Data-Driven Decision Making: The Untold Story of Marketing Analytics with Jeff Greenfield

• bestseopodcast.com • Episode 637

Share episode

Jeff Greenfield of Provalytics joins us to reveal the challenges of modern marketing attribution and how AI-driven analytics can help marketers understand which efforts drive results.

  • Marketers still struggle to identify which half of their ad budget is wasted
  • The shift from brand to performance marketing has created an overemphasis on clicks, not impressions
  • Digital marketers often overlook that awareness through impressions must precede sales-driving clicks
  • Most businesses deal with 10–15 conflicting sources of "truth" across platforms and agencies
  • GA4 and legacy analytics often miss the value of TV, podcasts, and other non-digital media
  • Grasping “ad stock” – how campaigns influence sales over time – is key for accurate attribution
  • AI models that assess incrementality reveal which efforts actually drive more sales
  • Complex customer journeys across Meta, Google, and Amazon complicate attribution
  • Strong models evaluate three key elements: ad stock, saturation, and media synergy
  • The surge in Amazon search (35% of product searches) and AI tools is reshaping marketing

-------------

Guest Contact Information: 

https://www.linkedin.com/in/jeffgreenfield/

https://provalytics.com

—----------

More from EWR and Matt:

Leave a Review if it was content you enjoyed: https://g.page/r/CccGEk37CLosEB0/review

Free SEO Consultation: https://www.ewrdigital.com/discovery-call

One-on-One Consulting: https://www.ewrdigital.com/digital-strategy-consulting/private-consulting-session

—

The Unknown Secrets of Internet Marketing podcast is a podcast hosted by Internet marketing expert Matthew Bertram. The show provides insights and advice on digital marketing, SEO, and online business. 

Topics covered include keyword research, content optimization, link building, local SEO, and more. The show also features interviews with industry leaders and experts who share their experiences and tips. 

Additionally, Matt shares his own experiences and strategies, as well as his own successes and failures, to help listeners learn from his experiences and apply the same principles to their businesses. The show is designed to help entrepreneurs and business owners become successful online and get the most out of their digital marketing efforts.

Find more great episodes here: https://www.internetmarketingsecretspodcast.com/  

https://seo-podcast-the-unknown-secrets-of-internet-marketing.buzzsprout.com

Follow us on:
Facebook: @bestseopodcast
Instagram: @thebestseopodcast
Tiktok: @bestseopodcast
LinkedIn: @bestseopodcast

Powered by: ewrdigital.com

Hosts: Matt Bertram 

Disclaimer: For Educational and Entertainment purposes Only.

Support the show

Speaker 1:

This is the Unknown Secrets of Internet Marketing. Your insider guide to the strategies top marketers use to crush the competition. Ready to unlock your business full potential, let's get started.

Speaker 2:

Howdy, welcome back to another fun-filled episode of the Unknown Secrets of Internet Marketing. Hopefully, I will edit in the bumper and the front. I don't know why it's having issues, but thank you so much for coming back. Everyone, I have a great guest for you.

Speaker 2:

I know we talk a lot about digital marketing strategies, seo, paid campaigns, email marketing, all the different things and I thought it none other to bring on Jeff Greenfield with Prova Analytics Provalytics sorry, provalytics and it's an AI-driven analytics tool that helps you figure out where that marketing that you're spending is, where it's going, what you need to do with it, how to reorient it. A lot of people out there are using just the GA4, the Google Analytics, maybe a Supermetrics. There's so many configurations, jeff, so it's so good to have you. What I can tell you is each individual client that we've talked to over the last well, 26 years, every single person needs something a little bit different. So we've tried to use Google Data Studio to set stuff up. We've done different kind of agency analytics. We've used so many different tools to try to find what works, and what I've found is it really needs to be customized to the client on what works best and for those clients that are out there listening, or even in-house marketers.

Speaker 2:

There's attributions, an issue like conversions. It's getting a lot harder with AI oversights in Google and people are not clicking through now. So there's so much going on, and I think that leveraging AI to help you make decisions is absolutely where the direction is going, and so I really thought, from a marketing attribution standpoint, to bring you on to kind of talk about what you're seeing in the marketplace, what works. Would love to hear your origin story, because I'm sure this was an issue for you with clients, because, well, everybody needs something different and no one's tracking everything, and then there's different groups not talking to each other and it, you know analytics is a big focus for sure, so welcome.

Speaker 3:

No, definitely. And you know I'll tell you kind of, the dirty little secret of analytics and of marketing is that you know, if you're a new brand that's just launching and all of a sudden you start advertising and people start buying, that's great. You can see that cause and effect relationship. But if you're an existing brand that's been around for a long while and you're spending dollars, the question that always kind of haunts especially the folks in finance at night is how many of these people would have bought anyway if we weren't advertising. So, meaning, what impact did advertising really have on the bottom line?

Speaker 3:

And that kind of gets to that old statement, the John Wanamaker statement, that everyone repeats half the money I spend in advertising is wasted. The only problem is I don't know which half. Yeah, I mean people just say that and people see the quotes all over the internet, but the fact is it's a reality. There's a lot of people who have purchased from you in the past who would probably come back on their own without advertising, and that's why there's this big movement towards incrementality, which has always been a component of advertising. Marketers have always thought about what is actually moving the needle for me versus I'm just spending dollars and that's kind of really what this gets to is the heart of is how do I find out who you know, what dollars should actually be spent, what dollars are actually incremental and how do I find the ones that aren't and move them over so that I can get more ROI with the same budget?

Speaker 2:

Okay. So before we get into it, I'm going to ask you something, cause I know you talk about analytics all the time, but I want to, I want to take one step back and talk about the differences in your mind of brand. Okay, cause, because bigger companies brand drive sales right, marketing supports branding right, but brand drive sales like Nike, the brand drive sales. Then they got all the other stuff on top of it to encourage movement Right, and of course they have so many skews and they're trying to move everybody. They have to continue to keep selling. But but I well, I can tell you the reason people are buying and that reoccurring piece of it is on brand. So, if you took a step back of it, how do you look at brand versus marketing? And is there any kind of measurement on brand measurement versus marketing measurement, before we get into kind of like attribution and marketing analytics?

Speaker 3:

Yeah. Well, in order to answer that I want to take time travel a little bit before digital marketing, because marketing was branding. Before digital marketing you could say there was performance marketing. Back then there were coupons that would come in the mail. So let's kind of level set back then all there was was tv, radio, print, direct mail, coupons, and performance marketing was probably coupons and direct mail, but everything else was branding. That's what things were back in the day. So when you think about all of the big brands that you know of today the Pepsi, the Cokes, the BMWs those brands have all been built based upon branding. So branding, we know for a fact, branding works and over the last 25 years or so there's been a lot of research to look at. What is the impact of this quote unquote performance marketing versus branding.

Speaker 3:

But we first have to understand how did performance marketing take off? Did all of a sudden the internet came around and marketers said let's move money over to performance? It really wasn't like that. What happened is that the Internet started up, there was advertising, that kind of started, and then there was the dot-com crash around 2000, 2001. And back then the newspaper companies and the owners of radio stations they were laughing and they were saying can you believe that those internet folks thought that they were going to put radio out of business and they were going to put magazines out of business? It's kind of funny now, but it's just important to understand that.

Speaker 3:

But then what happened is advertising started to take off. It started to grow slowly but surely to take off. It started to grow slowly but surely and around 2005, 2006, the boards of directors would start asking CEOs what is our internet strategy? And then that led CMOs to say let's try this out. So they would do a test and the reason they got excited about it is because when you would buy TV or buy radio, at the end of the month you would get an invoice that would just tell you what shows you were on and estimated impressions, and those impressions would become part of a mathematical formula that would get done once a year, something called marketing and mixed modeling. That would tell you how things work.

Speaker 3:

But you'd have to wait months or a year later, whereas with this new internet advertising, at the end of the month I got an invoice, but I also got this report that showed me how many people clicked on things, and click was much closer to the sale than this impression number.

Speaker 3:

So marketers quickly realized that if I spend dollars closer to our sale, to where I actually can get credit for things, that actually makes me look good. And so what happened is money started going from testing budgets to where it started moving over to performance, and everyone got very, very excited. And what ended up occurring by like the late, like 2010, 2011, a lot of branding dollars had been moved to performance, which is at the lower part of the funnel. And when you're advertising is at the lower part of the funnel, you're actually reaching a fewer number of people. So if you can picture, like the traditional sales funnel with X number of people in there, from just a hundred percent branding, now you're spending only half your dollars on branding and the other half down at the bottom. You're reaching fewer people, which means less number of sales overall, which means your advertising has to work twice as hard, three times as hard, four times as hard.

Speaker 3:

And that's so. That's how people kind of moved along, and then it got to the point where companies started turning off branding altogether because they couldn't measure it, because everything became click-based around 2015, 2016. And now we have this whole generation of marketers Everything is about click. And now we have marketers today who believe that the way marketing functions is we invest dollars to buy clicks and clicks lead to sales. But the reality is, the way marketing actually works is we invest dollars to buy eyeballs to get people's attention, and that attention is defined as impressions.

Speaker 3:

Every report we pull from Meta, google, wherever, there's always a column called impressions. Most people ignore it because I don't care about it, but impressions the job of impressions is that attention which builds awareness. The job of impressions is that attention which builds awareness. When awareness is built up, eventually people will walk into our store, and if our store is online, that leads to clicks and then clicks lead to sales. So imagine, everything you're doing, from targeting to measuring, is all about just at that lower point, which is where clicks are Like. Every marketer knows my cost per click, but really what you should be more worried about is the relationship between impressions to your sales, and that puts us back to how things were before there was digital.

Speaker 2:

Okay. So there's two, two kinds of themes that I want to pull from that that I think are, are, are really helpful, and I I love that history lesson and I I want to share that online and hopefully people will listen to it. I think it's great, um, and I think that people that are coming into the space that are new don't understand how it used to be Right and and a lot of the conversation around digital marketing or you know whatever. Like I remember when AdWords was pennies, okay, like it was pennies to do that and just the, the. The goal was to get someone to buy something online, like be able to want to put in their credit cards right, marketers coming into the space. Now you know this. This is just like, um, you know it's a, it's a, it's like a. Not even an afterthought, not even a prethought. Like, they don't even think about it at all and they're just looking at the clicks to sales, right, like there's no um view of, like the, the brand reputation of uh, of of the storytelling of what you know where that company came from, what, what's its history? Um, are people buying some nostalgia? Like that's not even into the conversation.

Speaker 2:

Like in in one area, when I talk to a lot of new marketers or I'm doing coaching or training, like they have no concept of branding at all. Okay, it's just direct. You know, and hey, that is the direct response world, right, like, hey, I'm going to spend this, I'm going to get this, I'm just going to sell that and. And a lot of those people that came into the space, maybe other on the agency side internally maybe had their hand at affiliate marketing, right, and it was like I can only spend this, I can get that. So it's a very like narrow view. But when you start working with bigger brands and bigger companies like narrow view. But when you start working with bigger brands and bigger companies, branding comes into it. And also there is some integrated marketing, there is some traditional marketing and I and one of the things that you said that resonates with me is I've come in a lot of times with really large brands where there's a multi cross-functional team or multiple agencies that are working together and we're bringing the digital aspect of it and to your point, they want digital work like 10 times as hard, right, and you know you're going after like a 3% or a very small percent of people that are looking to buy at the bottom of the funnel and and and typically like this is starting to change.

Speaker 2:

But radio TV get millions of dollars budget. Now TVs come way down right, and we're doing a lot more streaming over the top stuff like that. But terrestrial radio, um, for certain brands is a great driver to communicate a story and those dollars are huge. And so a lot of where, uh, we've been brought in is to test how to make that conversion, how to kind of how to digitally transform their advertising so that we can track stuff better. But, man, I run up against um on these cross-functional teams. Uh, traditional marketers that maybe right, they're, they're great at what they do, they've done it for 15 years. The analytics and marketing attribution, all the stuff we're going to get into is they?

Speaker 2:

Not everyone, but I've run in a lot of them do not understand this stuff because it's like you buy a flight. That's how they look at it and that's how some of these bigger brands look at it with digital advertising. It's not management, it's like we're going to spend this amount, we're going to run it. We're, we may pivot, we may not. Because, like, once you buy a billboard, you buy an ad in a book or a magazine or whatever. Like you just got to like hope that you got the right message, like there's not a lot of testing, You're just relying on experience. Well, like I'm a data person, I look at like how are people responding to it? Can we AB test? Like I love running emails where I use the whole alphabet on um, you know the, the, the headline let's, and then typically or not typically, but always there's like three or two headlines that work like 80%, 90% better than the rest of them and you have a clear winner right. So when you're AAB testing, you can get that with data, you can iterate, you can improve and you can make those marketing dollars work harder.

Speaker 2:

I've just seen that, like the smaller companies um are are more focused on the metric side of things. The bigger companies might still be really focused on the brand side of things, but but there's a lot of companies that are lost in this kind of traditional advertising or try to view digital as traditional and not measuring and not doing anything. And I think that that's where um, uh, your product, your company shines because you really, if you're spending a big amount of dollars and you have different variables that you're trying to test, um, if you don't have that data all linked together, it's almost like okay, well, we're going to add this variable to it and we're going to see what the output is or the impact is. But when you're running multiple campaigns on multiple platforms, tying everything together and talking about that marketing attribution gets really challenging and really difficult. I would love to hear kind of your viewpoint on, maybe, what you're seeing in the space, because, also, I've seen, too what I've typically seen.

Speaker 2:

When you start spending on meta or Google or anything like that, a lot of clients are like next day, like do we get any sales? Like where are the results? Like whatever it is, like it depends on the client you're talking to, but I've typically found nothing's going to happen for like two weeks. Like you got to get enough eyeballs, you got to get enough interest, they got to see the brand out there a number of times and and then, based on the customer journey, which I think that's a good conversation that we could jump into if we have time. It's not linear, it's not you go here, you go here, like you know, it's almost more like a, like you have a center point and then you go here, you come back, you go back to that point, you go this direction, you come back and then you got to get kind of in that buying phase. But they're, they're, they're looking at multiple platforms, they're having to see the ad multiple times and you know, doing a weighted attribution is probably the best but, like, adwords gets a lot of the credit for, I think, brand work sometimes because, um, advertisers, uh, marketers, like to run a lot of ads on the name that, to your point, as we were talking before, is going to get the click anyway.

Speaker 2:

And I've even seen that, like in the affiliate space, like with a lot of e-commerce brands, a lot of marketers are running ads on the name and that's a whole debate in itself of like, would you get that click anyway? Do you got to protect the brand? Do you got to drive what that experience is to a landing page, whatever? But? But a lot of marketers are measuring and giving good data to the client, like, and if you don't know what you're looking at, it's tough, and they're just saying, oh, look, how great AdWords is doing for you, but those are clicks. That I think you kind of said this but didn't say it like a little under the surface. Were you going to get that business anyway, right and so how you use these tools.

Speaker 3:

Yeah, and also to kind of expand on that. I mean, here's a. Here's a real life scenario. Someone's running campaigns on meta they're measuring using GA4, which is Google's analytics and a consumer sees an ad in their Facebook feed. They just kind of scroll by it, but they become aware A couple hours later they maybe hover over an ad on Instagram and they say, hey, that looks really cool, I think I want to get this. But people get distracted. And then the next day they say, oh yeah, let me see if I can find that thing. And then the next day they say, oh yeah, let me see if I can find that thing. And then they go and they Google it. They click on the ad there and the sale comes in and credits Google AdWords under the name which we would call brand search and in their analysis at the end of the month or the end of the week, they would see that Google AdWords is doing great.

Speaker 3:

Let's shut down meta or let's slow down meta, but the reality is in that instance is that meta is the one that you could say originated. The sale Broke the awareness initially. Yeah, exactly, and this gets even more complicated as you start to think about today's omni-channel world. So let's expand that a bit. Someone sees something on Meta, they Google it, they click on Google, they go to the website and then they say, huh, I wonder if I could get this on Amazon. And then they take their phone. So now they're in a completely different thing. They open up the app and then they order from Amazon and they click on the ad that the team is buying from Amazon.

Speaker 3:

So in that scenario, you've got an agency that's handling meta. You've got an agency that's handling, let's say, google. We've got a whole separate team, like the Amazon Marketplace team, that is handling Amazon Completely different budgets, sometimes completely different departments as well, because the way most large companies are set up things like Amazon is set up as a whole separate team, completely separate from the website team. And so now we're talking about things like Halo how ads on Amazon can drive sales to your website. Ads on Meta can drive sales on your website and back to Amazon. It gets very confusing for folks and very difficult to manage. But going back to your brand really, really quickly.

Speaker 2:

Showroom sales to add another layer to it, right. So so you drive them into the store. They're looking at the product, right. And then they want to price shop, right, like where they're looking at the product and they're like do I buy in the store? Right, because you got, you got people setting up displays. You got people driving stuff in store. You got, um, you know, mark, my, my, my sister works for Nestle, um, uh, ferrara now, but but she, she's doing that every day with candy, right, and water, and um, you know, there there's so many different factors on what drives that purchase at the end of the day, to, to, to your point. And so the bigger you get, the more money you spend, the more complex it gets, the more you need to understand the data. So right.

Speaker 3:

And what most people don't know is that they think you know, I'm a small entrepreneur, I've got a small business. My God, if I had money like a Nestle or some of these even larger companies that are so sophisticated. They have all these special tools and the reality is is that the larger the company, the more regions of the world that they're in, they're more likely on last click and very unsophisticated metrics, even though they know it's wrong. It's consistently wrong, so it's consistent, and in the corporate world, consistency matters more than being correct. Now, this is the first thing about metrics and analytics is that none of these things are correct. There's nobody out there that has devices that are hooking into people's brains to find out specifically what ads actually drove them down that route.

Speaker 3:

But what you want to do as a marketer is you want to say, okay, you just have to accept the fact that what you're doing today is probably wrong. Okay, just accept it. I don't care how sophisticated you are, and your goal as a marketer is that your decisions each month, let's say you want them to be less wrong Like this month. I want my decisions to be less wrong than they were the month before. It's a very simple goal, because if you say I want to be more correct or I want to have the right answer, there is no right answer in this stuff.

Speaker 3:

So that's, first and foremost, is that everyone is using some sort of mathematical model. But whenever it comes to these mathematical models, one of the things that you should ask yourself is where is the proof, where's the validation? Because, think about it, math you can always. You know in high school, remember you're doing math. You always check. They always ask show me, show me your work. And then you check to see if the number is correct. So the cool thing with models that deal with marketing is that and this goes back to the time of before digital we would actually validate things, and the way we would do it is you give the data a bunch of models. So, think about, like AI, you train the model.

Speaker 2:

And it's all math right. It's all math right. That's how the search engines work. It's all math.

Speaker 3:

And then we get to a point where we're like the model's cranking along and it's doing good, and then you say, okay, you're predicting what the sales will be, what the orders will be based upon, how many ads are out there in the combination of ads. So what I'm going to do is I'm not actually going to give you the sales number I want you to actually now, I'm just going to give you the ads that are out there and I want you to predict it. So then you look at the model and you see how well does the model, what is the number for today? Today's sales were a million. Did it go over a million? Did it go under a million and how close was it?

Speaker 3:

And then you look at that over a period of time, like a month or even a year's worth of data, and that's what gives you confidence to say, hey, I can trust this. Now a lot of people will say, well, I'm using GA4. How do I validate a GA4? Well, you don't, and you really can't. And that's the issue is that if you're using an approach in terms of making decisions on where you're going to spend your next dollar and there's no way to predict and to show how well your model is predicting the future, then you got to realize you're not using, you're not up to par in terms of what should be and what are best practices these days.

Speaker 2:

Well, you know. So, jeff, a couple of things that that at least these are like high level metrics that I look at when I'm looking at what the team's doing, right, so we have layers of account managers and experts, but I'm trying to kind of oversee the strategy, what's going on, and so I'm taking snapshots right Of like what's happening and try to look at it really quickly and measure different things. Like I'll look at search console, depending on what we have set up for the client, right, but we'll have. I'll look at Search Console versus GA4. And you know, I think Search Console is a better metric because it's anybody that's driven to that page and one of the things that you'll see and why Core Web Vitals and like load time are so important, is you drive advertising to a page and it doesn't load, or they jump and leave before the pixel fires on the GA4. And then you got a delta here of like the differential, I guess, of people that went to the site and left immediately before the pixel fired. So either your website's slow and we've identified that for clients I had to put together, for a really large client was spending a lot of money. I was like, look how much money you're wasting by having a slow website, right, like we have to fix this. You're losing so much so we couldn't even measure the advertising because, like the website so that's why it's almost pass fail of core web vitals. Like everything's got to load so quickly because, especially if you're pulling somebody off one platform to another, there's a level of trust there you need stuff to load basically instantly.

Speaker 2:

But then also, when I look at the advertising, so if I'm looking at advertising, one of the metrics I'm eyeballing and looking for is going, what is that rate of organic, maybe, or paid, or the combination of, because Google doesn't give you everything. But if I'm doing paid or if I'm doing display, like how quickly are people leaving? Like how long are those sessions? And I'll see some campaigns if we first start, which I'll have to give it a little time, but it's like under 10 seconds, right? So I'm like, is this the right audience that we're reaching? Because we're trying to, just like anything else, this is an answer to their question or solving a problem for them? Is this the right information to be showing them? And if they're going there and then five seconds or 10 seconds or three seconds are leaving, how can you say that advertising is effective.

Speaker 2:

So, looking at the quote, unquote, dwell time, which I know that Google has a combination of things that add up to 12, but they say they don't measure dwell. That's really important. Are people finding the information they're looking for on the page and are they consuming it at that level? Now, if it's 10 seconds and you're getting all these signups or purchases or whatever, maybe it is working great, but if it's not, and you're, you're in that testing phase. That's one of the things that I'm looking for immediately.

Speaker 2:

Okay, we've driven somebody here. Do they stick around and for how long? And then, how is that weighted against all the other channels to, to kind of see, of course, display less um or or typically is less, but yeah, it shouldn't necessarily be if, if you're getting the message out there and they're ready to buy and you're giving them something, uh, above the fold top of the page, here you go. Um, like, display is very, very effective of reaching a lot of people and building a brand story, and it's again how you're using the tools. But you know I I have some great rates on on display advertising where people say, oh well, that that's just not um. You know that that's just not that that's a normal rate. I'm like well, is it working effectively? Right, and I think that that goes in to the analytics of the more you can see and how you can measure it and what you're using it for. Um, you can start getting pretty sophisticated on how you're running your ad campaigns.

Speaker 3:

Well, time on site is incredibly important and I'm going to give away a little secret now which I should have saved for later.

Speaker 3:

But I'm going to give it away now because I'm going to add a new metric to what you're already doing, because what you're doing is you're looking at time on site and relating it to how that person showed up, what was the entrance that they made to the site? But what we know is that, in order for people to get to the point where they're ready to engage, there's other things that came before that which are not captured in GA4. And, of course, we would want all clients to invest in advanced analytics and do all this stuff, but the reality is most are not. I know that that's a fact, but one little hack that you can do is that, as you're looking at GA4, and you're looking at it day by day and you're looking at time on site based upon the UTM source or whatever, you can also go back in time for that client for the last year and download from every place where they advertised how many impressions there were for each ad for every single day. And so now, when you start to think about that relationship between time on site and the click source instead of looking at the source, because my gut tells me that for the majority of your clients, when you're looking at the time on site and you find that time on site is really extended and long, and they actually take action, that they come in through organic or brand search, meaning that they just show up. And so GA4 tells you that and this is true of most clients that the people who actually buy from you and do the actions that you design the website to do are actually coming in through a source that you don't know, meaning it's some other source. And so the secret to figure that out is if you gather all of the impression data for the last 12 months by day and then you graph it, create a graph so you can see it day over day, and then you create a graph of all the clicks and then create a graph of the time on site and create a graph of your sales or leads. Whatever you're looking at Now, start to look at when those things happen.

Speaker 3:

When all of a sudden, time on site goes up, clicks go up. Now go back to your impression graph and go back not just a couple of days, but a couple of weeks. You'll see an increase, maybe a day or two there. Then you want to dig into that day and figure it out what happened on that day, what ran during those days, during that timeframe, because that's the actual source of what's driving that organic traffic, that increased the time on site and increased the sales. So that's the big kind of secret and takeaway.

Speaker 3:

And that goes back to the understanding of ads, because you're buying ads and doing stuff to build awareness and that awareness building takes time. In fact, every ad that you buy, every ad you produce, even user generated content, it has its own unique time to conversion and I'll explain that a little bit more. Let's assume that one of your clients decides they're going to blow $8 million Next year it'll be $10 million on a Super Bowl ad Awesome. Well, chances are, if they're not ready for it, their website will crash. But let's say they're all set.

Speaker 3:

They're ready to go with Cloudflare, they're good to go. That day their web traffic, as soon as the ad runs, is going to go through the roof, and then, as soon as the ad is over, you would expect it to go back down to its same level. But the truth is that's not what happens. It stays up, it goes down, but it stays up at a high level for that entire day and, in fact, the next day it doesn't go down. It goes down a little bit, but it still is at a huge level and it'll stay at that huge level for at least seven to 10 days and then eventually, it may wake its way back down.

Speaker 3:

That phenomena is something that's called advertising ad stock, or it's also known as carryover. There's a great Wikipedia article about it. Now, that's an extreme example of a Superbowl ad, or it's also known as carryover. There's a great Wikipedia article about it. Now, that's an extreme example of a Superbowl ad. And it's because, my God, 12 million people, people see the ad. Maybe they don't go to the site right away, maybe they go to it a couple of days later, who knows?

Speaker 3:

But every ad that's out there, every message that you put out there, has a degree of this ad stock, of this carryover, that that, and what I mean by that is that there's an immediate effect when the ad happens, and then there's there's a decrease and a decline over time, and that's everything from a a pay-per-click ad, to a meta ad, to an Instagram ad, to a CTV ad every one of them has it and it's by figuring out that ad stock curve and there's a curve to it, then you can determine what we call that time to conversion, that timeframe, and then that's why kind of the hack is. Well, let me go and grab all the impressions for all the ads we've done, and let's start there. Let's look at that curve over time and then correlate that to clicks, and then correlate that to time on site, which is an awesome metric, and then to sales, and then you'll start to be able to unpack some of the secrets of what's working and what's not working.

Speaker 2:

See, I my mind immediately went to event-based marketing or conference marketing of like, set up the tracking for B2B companies to be able to look at that. And a lot of what we've been doing for clients recently for conferences is pre and post marketing as well as marketing while at the conference. Because you know the complaint we get a lot is you know, I spent a bunch on this conference. My salespeople went, they came back, they they have some names, they maybe followed up on it and then what is the time to that conversion? Like, what else can we do? A lot of these sales managers have big budgets and, um, you know that that since post COVID, uh, I feel like like everybody got a douse of like how powerful digital is over COVID and now everybody's going back into the office and things are kind of going back to the way they were. But now they're like okay, like it forwarded the progression curve of like a digital transformation a lot faster. So so I absolutely see that and also one of the things I do see in a lot of clients data cause I get to. That's one of the cool things about um working at an agency. I get to see so many different campaigns and, um, our methodology works across multiple industries. Like we, we build marketing programs that produce leads and produce new customers consistently for clients and across different industries. I see spikes in data, okay, and you know I'm usually moving a little bit faster, but I do say to the account managers it's like, why did this spike? Okay, um, what I've really seen is Reddit, reddit organic.

Speaker 2:

If you get, like you know, I guess, user generated content or people talking about your brand or you're using, like you know, something to monitor mentions or something like that, and some event or some action happens and then people start to talk about it. I mean, I've seen spikes in traffic. Like I'm talking, big sites triple in traffic for a day or two and then, and then certainly it typically comes down pretty close to where it was, but I have seen, depending on the campaign or the things that we're doing, that new level is higher than it was, and so that advertising stock is a very interesting concept that I would love to dig into more. But, yeah, the time to conversion is absolutely there. But, yeah, there's some action that created enough awareness of all the other ads that are going on that they're familiar with your brand based on what, what, what your, you know um, brand promises of what you're trying to solve, um, or the pain point that they have that connects you to it. And then you know really, um.

Speaker 2:

So Google has this methodology of uh, seven, 11, four like seven hours of consuming content, uh, senior brand, 11 times on four different platforms. Like they like seven hours of consuming content, seeing your brand 11 times on four different platforms. They've kind of boiled it down to that metric and I can see when that big action happens, or whether it's a PR event or maybe influencer post something or whatever the brand repetition. They get comfortable with the brand and then if they're searching for it, and then you show up in organic or SEO, uh, for this term, and then they search for this long tail term and you show up there again and then they also see the ad there. There's an affinity towards your brand, um, because they're familiar with it, uh, and, and then they have a higher chance to buy from you If you're solving the problem, if you meet those requirements.

Speaker 2:

And of course, you have to create trust and even in math that's a lot of my spring break reading is going to be some math models of basically how things are plotted on a graph or a matrix and where those edges are. And really each search in organic there's kind of a perfect score that you're measured against and then you're competing against everybody organically on who answers that question the best for that keyword combination. And there's kind of weighted entropy and there's a bunch of math associated with this and the ads are the same way. And so to understand SEO is to understand how the ads work, because they're working off of some of the same platforms and and they're looking at the clicks and there there's a lot. There's a lot to understanding the math and the analytics to give you the visibility to make the better decisions.

Speaker 2:

And I know I went on a little bit of a tangent there, but I thought it was worth sharing.

Speaker 3:

It was a good tangent, because what you're demonstrating, matt, is that marketing, which used to be primarily a creative pursuit, where people would get into marketing because they wanted to come up with really cool ideas, is now in a world of math and overly complicated. And everything you were talking about there, concerning SEO, has drastically shifted just in the last 12 months from two things. One is that Amazon now controls 35% of the search, and anybody who is searching in the US for any product not a service, but a product bypasses Google and pretty much goes direct to Amazon. So that's the first thing is that the SEO for products is completely different than it used to be, because now you have to be optimized and have a page in Amazon or you're going to miss out on business. That's number one. And then number two is now we have Gemini, which is showing up right above everyone, and HubSpot used to be the king of SEO for all their terms and it's been published and shown how their organic traffic just dropped down to. I mean, it's still huge, but it looks like. The graph looks horrible because all of a sudden, people are finding that when they Google a question, that question is already being answered right there and we haven't even discussed the percentage of the population that is now going directly to CatGPT or other AI platforms like DeepSeek or Perplexity or Claude or Grok. I mean, there's so many different places.

Speaker 3:

So the world of search and again, to pan the camera back a bit, this world is completely complicated and, like I've been saying since I first got into the digital space in the late 90s, and what I always have said to clients is that if for six months you've been doing something and it's been killing it, you are on borrowed time. And what I mean by that is that all of these techniques and all of these secrets and all of the ad buying that you are doing, just like you mentioned about Reddit all of these techniques and all of these secrets and all of the ad buying that you are doing, just like you mentioned about Reddit all of these companies are publicly traded and they face the pressures of quarterly earnings, and what I mean by that is that they tweak algorithms, they change their platforms in order to benefit their shareholders in order to benefit their shareholders, not the advertisers and not the marketers. So you have to always be on your toes. In today's marketing, it's less about creativity and more about trying to keep up with what's coming next.

Speaker 2:

So I want to get in and we can go a little bit long. I love our conversation if you have time because I think we're now starting to get into and we've teed up, like why analytics is so important, right, because it's not one channel anymore. That's a silver bullet, and that's why I wanted to bring the conversation also back to the brand conversation, cause your, your brand, floats across all channels. That brand consistency and, and you know, if you're trying to do something on LinkedIn, for example, okay, well, there's, there's a LinkedIn algorithm that you have to learn, and then organic and then also for paid, and how do those two work together? And then you're adding another channel of meta. Now, right, and I mean, we're doing a lot of stuff in WhatsApp. Okay, like, the WhatsApp follow-up is phenomenal right Now.

Speaker 2:

It's a moving target because shorts were doing really great and then those changed, right. And then live a couple years ago it's constantly moving, and I agree with you, right. And then live like a couple of years ago, like it's constantly moving and I agree with you, it's on borrowed time and it's speeding up. And then that's the big thing that we're focused on right now is ranking in the large language models like chat GBT. So the shift has been shifted like to Bing. Right, like cause, they're pulling from Bing If people are using chat GBT and then grok. Well, twitter, right, like. So now you have so many different things playing together and Amazon is just becoming like the juggernaut right Like there it is. They're owning everything, right, and, and you have to warm up those accounts and there there's certain metrics that that that you have to hit, and they're building communities inside Amazon. Right, like cause. They want to just be the one app for for everything. And so there is then the, the, the, the zero position, and the, the, the Gemini and AI overviews and ranking there. And that's not pulled from the first church result, like you know. So you're getting a schema. So, you're right, everything has gotten more complicated. So for the traditional marketers out there that may be listening to this podcast trying to figure out digital I'm sorry, the world's gotten hard, okay, the world's gotten a lot harder and for the new marketers that are out there, you need to go back to some of the basics of branding, right, so the pendulum swinging the other way, absolutely.

Speaker 2:

Now getting into the conversation that it took 40 minutes to have that I know we both want to have is well, analytics sits in the middle of it and you have to understand what's going on and how all these things interact. And now there's so many different variables. But you have the power of ai now to help you figure out the connections. I mean, if you look at what doge is doing with the government, like I think he's got the whole government mapped out of where the spending is going, but he couldn't have figured all that out like I mean, this is going to be a popcorn kind of watch what happens over time. And I'm not trying to point any political things, but the access to the data and the power of the data that he has of like the connections you can find, can, of course, be applied to marketing first, and that's how all these big companies have applied it.

Speaker 2:

And now we're talking about monetizing data, monetizing information, attention, like that is the real currency, right Is is your brand on people's mind? Like how does it make them feel? What does it get them to do? What is the call to action that you have? Like, so it becomes overly complex, but it's also quite beautiful to, to a certain degree, of like understanding the art of how all this interacts. But you got to have eyes, you got to be able to see what's happening, and the more you're able to see, the better decisions you're able to make. So I would love to transition the conversation to Provolitics and what it offers, and what are some of the challenges that you're solving when we talk about analytics.

Speaker 3:

Yeah, I think the biggest problems that we solve is marketers today don't have a single source of truth that they can rely upon. And the reason that that's important, especially for larger marketers and I'm talking to chief marketing officers, vps of marketing is that there's always that very important budget question. As a marketer, I want more money, because the more money I have, the more bigger team I can have, the bigger win I can have. And that's what you want and what's happening right now in organizations. So imagine a typical company. You've got a bunch of in-house marketers that are managing, let's say, five different agencies and you're using, let's say, ga4. And each agency is working on a different channel. They each have those platform metrics, like what Google says, what Meta says. So that's five different sources of truth plus GA4. So we're at six, and then each one of them is also going to do their own analysis. So that's another five, that's 11.

Speaker 2:

Okay, and they're all siloed and not talking to each other and you don't know what impacts are happening where right?

Speaker 3:

And then on top of that, we've got an Amazon team that has a completely different set of metrics. I have a wall. They don't talk, they don't talk like Amazon.

Speaker 2:

There's like a wall there. They do not talk to the rest of the marketing, yeah.

Speaker 3:

And then same with same with Walmart too. So all these teams have their own budgets. And here you are, you're ahead of all of these marketers. None of them are talking. So now we're up to about 15 different sources of truth.

Speaker 3:

And you go in and you talk to your CMO or to the CFO finance and they're looking at their internal dashboards and it's bringing in GA4 data and the problem is is that you want to go and get a bigger budget for connected television because your connected television agency told you it worked really well and you saw something. You saw more sales come in through organic. And your CFO is looking in GA4 and looking for the tab that says connected television and there is no connected television tab in GA4. There's no podcast tab, there's no TV tab, there's no TV tab, there's no radio tab. There's none of that stuff. That's inside of GA4.

Speaker 3:

So what we do is is that we come in with a platform that takes in data from all of these different platforms, takes in data from the client's data warehouse, and we built an AI driven model that literally kind of looks at the world through the lens of how it was before digital and the way it was before digital is that everything was based on attention and eyeballs. How many eyeballs were you trying to reach Impressions, and so the best thing about impressions is that it allows us to look at everything under the kind of the same lens. So think about it. Podcast advertising is huge. It's taking off. A lot of ads in podcast advertising today, a lot of ads in connected television, a lot of ads with what we call digital out of home, those digital billboards you see when you drive and yet you know really quickly.

Speaker 2:

Do you remember when they sold cable, as you got to pay for cable so you don't have to have ads, right? Oh yeah, that was back in the day.

Speaker 3:

Yeah, I was just talking to someone this morning at the gym about this about now. I can't download and watch the whole show all the way through. It only comes out once a week, I have to pay for it and there's ads and I'm like, okay, how did this?

Speaker 2:

improve Well, even on the phones now the game. So I want to give my kids the like games on my iPhone. The game on my phone that come with it now have ads in the games, like I mean we're it is. It is wild the access. It's bad on one end if you're the consumer and you can access people anywhere at any time, like doing anything, so the world opens up of what you can do, right, like. So it's always a double-edged sword. Sorry, I interrupted you. Go back to it.

Speaker 3:

So what we do is is that we enable the marketer to bring all this data together in our platform, then we use a model that goes through and figures out what's working and what's not working, and it's based on incremental impact. So, just like we talked about at the beginning of our conversation, if an ad if, if, if an ad wouldn't have actually driven additional sales, it's not going to get credit. So it's all based upon incrementality. And then we bring it all together and unify all of this stuff and then actually tell them what the very granular recommendations are where they should be spending and how they should take their budget and reallocate it, not just for those five agencies, but we also incorporate the Amazon and the Walmart.

Speaker 3:

So we look at the whole brand holistically, because what's happening today is that you run an ad and you tell people to buy on your website and that ad builds awareness and they say, oh, let me check out Amazon, like we talked about earlier. So that's the key is that, even though you're telling people to buy in one place, they're buying someplace else, and that makes measuring more and more difficult, and so what our platform does for these much, very large advertisers is tell them how to actually move their money around in order to get more return for the same advertising dollars, so it enables them to be more efficient and increase their sales, and so that's really what we solve. The biggest problem is creating that single source of truth within an organization.

Speaker 2:

All right. So I want to drill down. I know we're a little over time here, but I wanted to drill down and just get a little geeky on me on the AI that you built and the model of, like, how it's analyzing the data right, because AI is so powerful. Okay. So, like, just what are some of the thought logic that's happening with your AI? Because, like, hey, there's going to be people out there that are saying I do AI analytics, right, like, that's going to be the new hot thing, like everybody's like everybody's calling me like AI agency. It was funny. It was like. This was like nine months ago we landed a client and, um, uh, they called and said the only reason we're talking to you, the only reason we're talking to you, is because we were searching on perplexity and you came up number one as the SEO agency on perplexity.

Speaker 2:

Okay, and this is eight months ago, okay, and and and my first question was like what is perplexity? Like that, like we were, we were doing, we were doing yeah, I mean we were doing quality SEO, right, like that's what we were doing. We were doing quality SEO. I wasn't kind of like trying to trick the algorithm to optimize for that, like I was just doing. You know quality SEO and it recognized that Right, and so you know.

Speaker 3:

I'll give you. I'll give you a couple of the secrets of what we do and what the AI does.

Speaker 2:

Okay, so what it's? I'll give you a couple of the secrets of what we do and what the AI does.

Speaker 3:

Yeah, see how it works, right. What it's looking at? Yeah, so what it looks at is it looks at that causal relationship between all of these paid, earned and owned campaigns. For some of our clients, we're taking in all of their PR data, all of their organic Facebook posts, everything. And what it's looking at is we talked about this earlier it's looking at that ad stock, meaning those campaigns, how long they have an effect for. But then it also looks for this concept of what we call saturation. What I mean by that is that everything you do where you're spending dollars, there's something called a marginal response curve, and what that means is is that, hey, I can keep spending dollars on meta, they'll keep taking my money, and I can keep spending more, I can increase my budget. So you hit a plateau or a threshold, right?

Speaker 3:

Yeah, you're going to reach a threshold, whereas you keep spending. So let's say, for every dollar I'm spending, I'm getting $10 in revenue back. So I double my spend and now I'm at a nine, I'm getting $9 back. That's great. I doubled again, now I'm down to a five, and what that means is that eventually that curve is going to start to go down. And so, as you're looking across a very large campaign, you want to start to think about well, once I get from a 10 to a nine to a five, is there something else that, right around the same time, is actually going to give me a 6.5? Maybe I should move some dollars over there, the level of complications that are here.

Speaker 3:

Then, on top of that, the third thing is this concept of what we call media synergies.

Speaker 3:

What I mean by that is that an ad has a certain response that's associated with it, both its immediate response and that long-term ad stock response.

Speaker 3:

But when you combine one ad with another, when someone is exposed to, let's say, a connected television ad and maybe a display ad or a certain retargeting ad, it's not one plus one equals two, it's one plus one equals seven. So in order to figure out those actual synergies, the math that's involved is actually even more complicated, and you have to run what we call simulations, where we take every single day and we remove a very grand. It's like we take that day and we say to the platform we're going to remove this one ad this day. You had a million dollars in sales. Now we're going to take this ad out, and I want you to run a simulation and figure out what our sales would have been for that day, because that allows us to actually calculate that synergistic amount. So it goes through this whole process and then it layers in, interestingly enough, two techniques, both of which have been around one for a very long time. The first technique is called Bayesian. Bayesian is what's driving most of the AI today and all of the machines.

Speaker 2:

The Markovian, basically the advanced Markovian right.

Speaker 3:

Right, and that was developed in the 1700s, a long time ago before there were computers. The other technique we use is a technique out of the University of Chicago that was developed in the early 60s, called seemingly unrelated regressions. Traditionally, when statisticians and math experts would solve these types of marketing equations, they would separate them based upon the channel. So they would solve the math for TV, separate it from search, separate it from digital and then manually combine them together. But that creates a scenario that doesn't simulate what the real world is. The real world is all. This stuff is all happening at the same time.

Speaker 3:

Seemingly unrelated regressions allows us to solve that all at the same time. Once we solve it all, then we want to go back and see how did we do? We want to check our math, and the way we do that is we use a validation technique called K-fold, which is common in learning engines and machine learning, where you want to quickly iterate and what that allows you to do is to hold back a certain amount of, in our case, days, and what we do is we hold back 20% of the days. We train it with a month's worth of data, we hold back 20% of the days and then we run it five times, so we actually do what's called K fold five, and so we end up with a validated result where we hold back a hundred percent of the data, and so we can actually say, hey, this is how well the model performed for the entire last year's worth of data, and that gives our clients confidence to know hey, I can start running some tests based upon the recommendations of this model, which is really cool.

Speaker 2:

No, I love that. I love that. All right, jeff Provalytics, jeff Greenfield, go check them out. Where's the best place to find you? Where are you actively posting, talking about some of this geeky math stuff, which I love? Like I, I've really enjoyed our conversation, um, I, I think we could do a part two, uh, and and go even deeper on some of this stuff, because I think, uh, there's there's a lot of uh things that we're talking about uh from a different angle.

Speaker 2:

That definitely connect, uh, but yeah, how? How's what's the best place for people to find you or find out more what you do? Check your out.

Speaker 3:

Well, they can go to Provalyticscom, but the place where I'm the most active is on LinkedIn, so find me on LinkedIn, check me out there. I pretty much post something just about every single day.

Speaker 2:

To hit the algorithm right. Yeah, of course, one of the biggest tips that I got that you know, 15 minutes a day. Set your timer 15 minutes a day because you've got to feed the algorithm. Like, and you know, youtube says they don't punish creators for not producing content, but they absolutely do Right, yeah, right, and so that's where these compilations and things for not producing content, but they absolutely do right. Of course they do, yeah.

Speaker 2:

Right, and so that's where these compilations and things people are running stuff live like cause they, you gotta hit, you gotta hit the metrics. So everybody, go definitely check out Jeff's work. Go check out Provolitics. I'll put the link to his LinkedIn and his company in the show notes. Jeff, thank you so much for coming on Everyone that's listening. If you enjoyed this conversation, please share it like it. It really helps us. We are trying to move over to YouTube. We are going to be pushing out some more shorts.

Speaker 2:

We may be changing the name of the podcast. I know that we've had this podcast name for 13 years, so that's really scary to me, but I will let you know more about that. If you are looking to grow your business, if all this stuff is making your head spin and you need a trusted source, a marketing agency, to reach out, to reach out to EWR Digital for more revenue in your business, A lot of people don't connect this podcast with our agency. We've been in business since 1999, beginning of the internet, and we've seen it happen. My mom was one of the first employees of Microsoft. I got some great stories. Going to have some more great guests. Thank you for supporting us. Until the next time. My name is Matt Bert. Great guests, Thank you for supporting us. Until the next time, my name is Matt Bertram. This is the Unknown Secrets. Bye-bye for now.

People on this episode

Head Shot

Matt Bertram

Host