Rand Fishkin, Co-Founder & CEO of SparkToro, joins our host, Camela Thompson, Go-To-Market Thought Leader and B2B Insights Expert, in this episode of the Revenue Marketing Report. Rand shares why he doesn’t feel the end of third-party cookies will meaningfully impact marketers and got into a debate about the utility of attribution with Camela.
We are here to talk about the cookiepocalypse. What are your thoughts?
“So I generally believe most analysts and prognosticators that Google is finally going to drop third-party cookies from Chrome. There are some people who mistakenly think this means all kinds of cookies. This isn’t true. First-party cookies aren’t affected. So for instance, if you go to your bank’s website and you log in and your bank saves that machine’s details that you’ve been there before and you have authenticated whatever the cookie that the bank drops on your machine in your browser, those continue to exist. Third-party cookies are essentially cookies that primarily ad networks serve you through. They’re used to track your behavior across the internet on multiple different websites and build a profile of you based on what you’ve visited and then send that data back to ad servers so that advertisers can buy targeted advertising toward people who visited different websites in the automobile field or whatever it is.
“Those are the problems. Those are the ones that are going away. And they are probably going to be replaced by an inferred Machine Learning-based system that aggregates behavior that Google, Facebook, Apple, or third-party ad network of your choice like Tabula. Those kinds of people know about you and are cohorts, what are the sort of cohort behavior of people who look like us on the web, might get grouped into a category and then Google will sell access to that category to advertisers, which isn’t nearly as targeted as the kinds of advertising that people do today. This is what’s generally called the cookie apocalypse for folks who aren’t familiar. I have many thoughts about this. My biggest thought about this is I don’t think it’s that big a deal from a technical and practical perspective because I think that a ton of big tech advertising is actually Google, Facebook, Apple and Amazon being really good at predicting where someone who was eventually going to buy was already going to go and showing ads to them and then you paying for the privilege of having the analytics to see it.”
I think that’s what people need to understand why Google has delayed so many times is until they have a real way to cope with this, it’s going to impact the effectiveness of their advertising and their advertising income.
“I don’t think so. It’s not going to affect the effectiveness. I think in most cases, in a few cases it’s effective, but there are edge cases where it will impact the effectiveness. But I think in the vast majority of cases, Google was and is taking credit for sales already would have happened anyway.”
Fair. We have seen the transition from Google Analytics to GA-4 and all the unknowns that are, it’s just yes, that probably didn’t make sense to those of you who aren’t in Google Analytics all the time, but it is painful.
“Well, and they removed some of the options for attribution that were in Universal that aren’t in GA-4. People have been emailing me saying it sure looks like they have switched to GA-4. It looks like a Google pay-per-click is responsible for all my conversions. I don’t remember that being as true in the previous version of GA. Did something happen? Yes, something happened. Guess what? Google’s advertising and analytics people sat down together and said hey, do you want to make a bigger bonus this year. Alright and a little handshake.”
That sounds very shrewd, but you’re probably right, probably spot on and it’s a good reminder for people to track UTM parameters and those things so they can verify it in different ways and not just rely on Google Analytics.
“Perhaps part of me wants to encourage people to try and find technical solutions and workarounds and a bigger part of me wants to say if you’ve chutzpah to do it, you should just throw out the idea that attribution can be measured on the internet.
Oh, we are getting into it early. This is going to be fun. So I have a tough time with that since I come from B2B space with really long sales cycles and I’ve been one of those marketing leaders who have been hammered for every dollar you put into each channel, how many come out? It’s hard to change the executive mindset. I don’t disagree with you, but it is really hard to change their minds. Therefore, maybe we can get into that a little bit.
“Yeah. I think it is like trying to convince average Americans that crime has gone down.”
I mean statistically, but if you heard it enough times.
“Right. So statistically, crime in the last 40 years, every five years is so much better than the last brief spike during COVID-19. Sure. Now it’s down again and though almost no Americans believe that crime has ever dropped, the truth evades the mentality. And many people when you bring in the fact crime is way down, say but my cousin, his car got broken into or I was robbed at this gas station in California. The anecdotes aren’t statistically sound aggregates to convince executives. So executives I think have to be intelligent and thoughtful people or maybe we shouldn’t work for them. We should go find new people to work for and having a conversation in the abstract is a good way to go. Therefore, not saying, hey, you shouldn’t track your metrics. You should throw out all that stuff.
“But instead of starting with the broader conversation of hey, do you think it is possible that what Google’s really doing with all the ads that appear say, oh, buyers who are likely to buy from your site, we know that they’re going to visit ten or twelve sites and pages across the internet in their journey. Let’s make sure your ads show there. So that we get credit for a lot of the sales that would have happened anyway. And executives will say, oh yeah, I am sure they do shady stuff like that. In fact, check out this Department of Justice case. They unearthed this email. This doesn’t matter where you are in politics. This started under the Trump administration and continued under the Biden administration and the DOJ has all these documents showing collusion and sketchy stuff happening inside Google to do exactly this. And you can also prove it to yourself by looking at all these case studies.
“Go to Chase Manhattan Bank had a great one that was featured in the New York Times a few years ago where they cut their advertising from 95% of websites and saw the same exact number of signups, transactions, and conversions. Uber did the same thing and saved around $100 million. AirBnB cut their advertising spend by more than 90% in 2021. And Brian Chesky did an earning report that was quite famous in the marketing world about this. eBay had a good case study around this. So there are tons and tons of these big companies who have cut their advertising spend by massive dollar amounts and seen the same or better performance. The only explanation that could possibly be true is a ton of ads that Google, Facebook, Amazon, and Apple claim are creating new demand, are actually being served to existing demand.”
And I have to wonder if there are some other things going on. I mean those are huge brands with tons of recognition and this is an awareness play. Therefore, I wonder if we have hit market saturation or if it is difficult to track inefficiencies. Yes, for all the reasons you just listed. But stepping back a minute, I liked your analogy because I feel, and this is going to pay into my next statement, that many leaders are still intuition-led even though they think they’re data-driven. We’re still trying to scale B2B companies the way that used to work ten or fifteen years ago and the buyer demands have completely changed, buyer committees have gotten bigger. There are all of these things that are pointing toward marketing and customer success becoming more important in the sales cycle. So what are your thoughts?
“I fully believe that in terms of most B2B purchases, product reputation and brand matter most, marketing distribution and reaching awareness building are second. And then everything else which would include your pay-per-click ad in Google appearing above your organic listing. Your puppy dog, third-party cookies following me around the internet advertising or social media ads. That kind of stuff is very good at taking credit for sales that would have happened anyway. And, if I were coaching, and I often do that with lots of executives in the B2B world, I would tell them to do as much as they can to get rid of the idea that attribution is something that’s possible on the internet in 2023 and instead move to a world of measurement.
“And if you don’t believe me, that is fine. Some of your competitors will and they will probably eat your lunch. And the way to prove this is to use a quarter where you’re feeling comfortable about it, bring your advertising spend on a platform or two down to zero. Repurpose those dollars into something else like brand campaigns, content, whatever your product is, new hires, new agency, whatever. Put those dollars somewhere else and then see what happens if we cut all of our Twitter spend. It has been a hundred days and it sure looks like we’re getting 80%, 90%, or 105% of the same results year over year based on the average growth rate compared to the last quarter last year that we got from Twitter where we are seeing whatever it is. And then ask yourself, if we spent 0$ and we still got, let’s say 80%, then the marginal contribution of that channel was 20%. Even though the channel took credit for this many conversions. Now you have real data. Therefore, if you want to be data-driven, this is the way to go about it, the problem that arises from suggesting that to executives, especially CMOs is that they are so scared of a 1% drop in potential new sales or growth rate that they’re unwilling to trade that for a massive increase in profitability or cost of acquisition.
“And if you’re VC-backed, I get it. VCs want 95% of the companies that they invest in to die and then the other 5% have explosive growth and get to a billion dollars or whatever, become a unicorn. But the die trying bunch, those are fine too. So they’re going to encourage most of their companies to way overspend to hopefully become the monopoly in their space, which I think sucks for a million reasons. However, it works for their model. So I don’t think you’re going to convince those people to change their minds. But anyone who is not venture-backed, you’ve got a huge competitive advantage in cost of customer acquisition in the profitability of your potential business, your ability to invest in alternative channels that might be longer term to show ROI, but much higher ROI when they mature. So that’s how I encourage and coach folks to think about it.”
And then with attribution, I’m a fan of measuring what you can and using that to help you to make decisions, which is much of what you just described. I think where attribution becomes really problematic is when it manifests as this pressure to prove the amount of pipeline and bookings you’re doing and to get credit for things. And that’s just a disaster waiting to happen.
“Whatever you come up with is wrong. I guarantee 100%. I would put all the money I have in a bank account on the fact that everybody’s attribution model is wrong and incorrect and there is no perfect model. If you get it right, it is because you won the lottery in the context of statistics.”
Yeah. I would say that attribution being wrong is okay if you know what you’re using it for and you’re not trying to battle with other executives for credit. So if you’re simply using it as a directional, it’s never going to be right, as you said, the way to figure out what’s working quarter over quarter, fine. Nevertheless, this whole battle in the boardroom, it is something we’re never going to win since I think so much of it is actually intuitive-based, how people used to grow businesses and wanting to continue that trek. It’s confirmation-based or the opposite of that. They’re going to look for whatever they can.
“It is very frustrating. Marketers spent a good century, the 20th century, investing in marketing, advertising, branding, and content in lots of different ways that never showed attribution, but did show incremental lift. So Coca-Cola would run two different billboard campaigns, one in Cincinnati and the other in Cleveland and they would look at whether the Cleveland billboard outperformed in terms of lift of the same source sales over Cincinnati ones. If it did, that would be the nationwide or Ohio-wide campaign for the next year. That’s great, right? That is a phenomenal way to test incrementality and lift-based measurement, pretty smart, particularly in isolation cases like that. What they never tried to do is say, Joe Schmo walked into that convenience store and bought this many cases of Coca-Cola because of this billboard.
“The fact that we tried to do that in the first fifteen to twenty years of digital advertising on the internet. It is kind of dumb. It was never a good idea. I think even though I am someone who for years encouraged that sort of behavior. I was an attribution advocate. I showed people how to build models. I filmed videos on how to do it. I spoke on stage. I tried to show people; here’s what SEO is responsible for. Here’s what this keyword produced in terms of conversions and this piece of content. How many conversions did it lead to? I did all those kinds of things and I now believe that I was foolish, mistaken, and misled. As was the case with many marketers of my generation. I certainly wasn’t alone. We did that since the big tech and big advertising companies had such a huge incentive to try and convince every executive to buy into that.
“And what’s funny is if you remember, let’s say we were having this conversation fifteen years ago. It’s 2008, I bet it would be the other direction. It would be us trying to say, gosh everybody’s still addicted to the 20th-century style of incremental lift-based measurement and that’s old school and we’re going leave you in the dust with our new attribution-based systems that can measure all these different things. And there may have been a moment in time when Google was still providing keyword data in analytics, when third-party cookies had no expiration limits, when there were virtually no privacy laws or regulations, when there was virtually no one who blocked cookies on their devices or blocked tracking of any kind, when there were no privacy-centric browsers, when no one had ad blockers. There was a window where you could get reasonably close, reasonable people could disagree about whether attribution was possible in a directional sense.
“Today, I don’t believe even in a directional sense, I think you’re probably lying to yourself if the directions are or if the numbers are so big and trustworthy and you’ve tested cutting off a channel and then bringing it back in. Okay, you might be able to prove to me in special cases that your increment, your attribution-based model is directionally useful and accurate. Some people probably have that. But for most folks, I would urge them to just throw out that way of thinking and go back to the 20th-century style measurement.”
Yeah. I have been one of the lucky few who have been able to tool a huge system to prove incremental change, but it took a ton of education on how to use it, what it is good for, what it wasn’t good for, and how to think about it. And that’s the hard part.
For more content on B2B marketing trends, listen to the full Revenue Marketing Report episode at the top of the article or anywhere you podcast.