A History of Marketing / Episode 48
This week, I’m joined by Scott McDonald, who spent three decades in the research trenches of America’s biggest magazine publishers before becoming president of the Advertising Research Foundation (ARF), an organization now celebrating its 90th year of trying to separate marketing science from marketing spin.
Scott led consumer research departments during the Golden Age of Magazines. His insights helped launch Martha Stewart Living, tripled The New Yorker’s subscription price, and he saw the internet disrupt the business model he’d spent years optimizing.
Along the way, he picked up insights that still resonate. Including:
The Strength of Weak Ties: How a core sociological concept explains networking and provides a framework for go-to-market efforts.
The Power of Print: Why Steve Jobs insisted that every new Mac launch campaign include an ad in Time Magazine.
Cultivating Authentic Brands: Behind-the-scenes stories of using qualitative focus groups when launching Martha Stewart Living.
Scientific Marketing via the ARF: Including the empirical rule that cutting your share of voice during a recession will reliably cost you market share.
Listen to the podcast: Spotify / Apple Podcasts
Now here is my conversation with Scott McDonald.
Special Thanks:
Thank you to Xiaoying Feng, a Marketing Ph.D. Candidate at Syracuse, for reviewing and editing transcripts for accuracy and clarity.
And thank you to Bill Moult, whom you may remember from episode 23 of this podcast, for introducing me to Scott.
Connecting Sociology with Marketing Research
Andrew Mitrak: You got your PhD in sociology from Harvard University, and then you got into a career in media and advertising. Sociology is such a fascinating topic. I always enjoyed my sociology classes in college. At a broad level, how did sociology influence your career?
Scott McDonald: Well, my interest in sociology went back to undergraduate days really, where I was mostly in the historical comparative wing of sociology and interested in social movements and things like that. And then when I graduated, I graduated from University of California, Berkeley and was totally broke by the time I got out of school. I needed a job. I went to the job board and found a job that involved program evaluation, just kind of project work, evaluating educational programs for the California Department of Education. And it ended up being quite fascinating because it was the first time I’d actually thought about how you would address structured applied problems using the skills of social science. So I cut my teeth on that, doing projects for the Department of Education, for Bay Area Rapid Transit, for all these sort of public entities. And that drove my desire to go to graduate school in sociology to learn the quant side, which I had not really studied as an undergraduate.
So that’s really the main throughline to the work that I’ve had in advertising and media because I approached it very much through a background in studying statistical modeling, pattern recognition. I was particularly interested in graduate school in demography. And so demography sits at the border between sociology and economics. There are other borders in anthropology and psychology and other things like that. But I was mostly interested in the border between sociology and economics. And that carried through, I’d say, through my entire business side career. But also I had really fallen in love with doing applied work as opposed to sitting around theorizing at a university. So I was much more receptive to those job offers.
And one came to me when I was just rehearsing for doing job talks, going around to campuses and presenting myself as a soon-to-be graduate of a PhD program. And quite randomly, a good example of the sociological theory of the strength of weak ties, that a job at Time (magazine) came up where they were looking for an academic social scientist to try to crack a problem that they found intractable. Because a guy at Sports Illustrated in the Time Inc. portfolio had gone to high school in Chicago with the wife of my thesis advisor. The weak tie led to the referral. I went to New York and hit it off and decided to move to New York and work for Time Magazine instead of joining the faculty at the University of Arizona as a starting tenure track professor.
Andrew Mitrak: Can you define more so the strength of weak ties? Like what is that idea? I haven’t actually come across it.
Scott McDonald: It was popularized as the six degrees of separation concept. That it isn’t so much who you know immediately, but it’s who people that you know know. That’s one degree of separation or two. So most jobs actually come to people through those kinds of referrals. Not exactly the person that I know, but someone else that I might be able to help them actually discover an interesting job. The exception usually in sociology is recent immigrants. Why do you have Haitian taxi drivers or Indian newsstand owners or something like that? Because their networks are small and they’re very specific to immigrant communities. But once you kind of move out of that, and of course universities themselves are super important as drivers of social networks, and they allow people to expand their networks a whole lot. There’s a whole field of economics now that has to do with the life chances that come to someone just as a function of whether they grow up in a well-networked place like say Austin or a poorly networked place like Waco. Geographically they’re not that distant, but they have very different social networks and different opportunity structures. So sociology, you know, again this is like demography, pattern recognition. When you think of the way that you would discover some of these theories and test them, they’re similar to analyzing the influence of say a magazine compared to a social media influencer. You can graph that stuff.
Andrew Mitrak: It sounds like a concept that’s really applicable to marketing in a lot of ways. And we tend to as marketers think of it as just social networking or your second-degree LinkedIn connections or your alumni network, or how you might build an audience through reaching out to influencers and connectors. But it seems actually useful to look at concepts from sociology that have probably studied this in a more rigorous way and come up with things like the strength of weak ties to frame some of your go-to-market efforts.
Scott McDonald: I’ve always thought of sociology as being very, very flexible partly because it overlaps with all these adjoining fields. And it’s always scrambling to try, it doesn’t have one unifying theory as economics does. It’s got a bunch of theories. So—
Andrew Mitrak: Sounds kind of like marketing.
Scott McDonald: It is, exactly. Exactly.
The Golden Era of Magazine Publishing
Andrew Mitrak: So you got to Time.
Scott McDonald: My first big post-graduating job.
Andrew Mitrak: And this was in the early 80s or so?
Scott McDonald: Yeah, 1982.
Andrew Mitrak: So what was the portfolio of Time magazines? Obviously everybody knows Time Magazine, and you mentioned Sports Illustrated...
Scott McDonald: Yeah, so the big moneymakers were the weekly magazines. It was Time, Sports Illustrated, and People (magazine). And they all made boodles of money. It was sort of the heyday of the magazine publishing industry. There were also a bunch of monthly magazines as well. And of course, Time Inc. owned a bunch of other things. Book of the Month Club, a publishing imprint. I forget exactly which ones they had, but they had a lot of things. And importantly, HBO. And so there was already kind of a media empire. They owned some cable systems and stuff like that. And then a couple years after I joined, they merged with Warner Communications, which brought them a movie studio, a music company, and a bunch of other of those assets, the Turner Broadcasting System, and CNN, and all that. So it became more and more of a media conglomerate while I was there. A very interesting place to work.
Andrew Mitrak: Yeah, we’ll talk about your time there, your work there, and how it evolved while you were there. But before we get into that, I thought this might be an opportunity to talk about magazines more broadly. You kind of called this the Golden Era of magazines. And they were such a huge part of American media and culture in the 20th century. And we haven’t really discussed magazines at all on this podcast aside from occasionally we reference an iconic ad that would have appeared in a magazine. And iconic ads are so critical to the medium of magazines. Do you have any thoughts on the rise of magazines in the 20th century and how it impacted the way brands marketed themselves?
Scott McDonald: Well, a lot of magazines are aspirational. And people kind of put themselves into that. Many are vertical. Time was an example of a fairly broad magazine, and it competed with other leading news sources. But it was much more in-depth than say what you would get from broadcast television news or something like that. Much more the middle-brow intellectual version of news. It wouldn’t be The Wall Street Journal necessarily, but something that was very, you know, they broke stories and competed in news. So a high-brow, well-heeled audience at a reasonable amount of scale that provided, say financial companies, any company that was trying to influence opinion would be a reasonable target. So like Microsoft when it launched, Apple Inc. when it launched. As a matter of fact, Steve Jobs always insisted that any new campaign had to include Time Magazine. So he was from a generation that viewed this as a super important, influential medium.
And magazines actually were that. They were criticized sometimes as being gatekeepers. Editors had a lot of power in setting agendas or anointing. I worked for Condé Nast. Vogue (magazine) is famous for anointing a new designer. Someone that Anna Wintour likes gets featured in Vogue and they’ve made it. It’s like they’re on the blotter. That’s less true now because you have competing sources of influence, but the appeal to advertisers in part was always that. And when you do consumer research, you would see that very often the readers of those magazines believed that the ads were really part of the value of the magazine. So a September Vogue was evaluated partly by how thick it was. Well, the thickness wasn’t editorial copy. It was a lot of ads for September Vogue, and consumers would actually think that Anna hand-selected the ads.
How Brands Measured ROI on Magazine Advertising
Andrew Mitrak: Can you take us behind the scenes of who are the players when it comes to marketers at a brand? Let’s say Apple, Steve Jobs wants his ads for a new Mac launch in Time Magazine. There’s Time, there’s the publisher, there’s advertising agencies, there’s Apple and the in-house company. What is sort of the relationship between how an ad actually gets into a magazine?
Scott McDonald: Okay. So the publishers, and of course since I worked for Time Inc. and then WarnerMedia and Condé Nast across 30 years, my view is a bit, the lens that I apply is from the publisher side more than anything else. Publishers very much wanted to have a direct relationship with the brand, with clients. And a lot of the communications were direct there. So at Condé Nast, I would go present directly to L’Oréal, for example, one of the bigger cosmetics advertisers for the house. And this was somewhat in conflict with the agencies. Agencies were supposed to be planning media across the board, but they often were really confined more to managing the television side and later on the digital buying. So the publishers preferred that because sometimes they didn’t compete with more mass media like TV on reach, but they were more influential. Very similar to what we look at research now, podcasts don’t usually have the same amount of reach as some other media, but they’re much more influential. They’re persuasive to the people who listen to them. And so they have a traction that is in some ways very reminiscent to me of what you would emphasize in conversations with publishers about the value, why they needed to be in Vanity Fair (magazine) or whatever.
Andrew Mitrak: How were the brands measuring their Return on investment on their magazine advertising? As we’ve looked at this era of marketing metrics and analysis, a lot of it tends to be around TV, and it feels like there was a lot more scanner panel data and things like that that were almost tied to television sets and stuff. But I haven’t actually heard it brought up on how it applied to magazines and such.
Scott McDonald: It was harder to justify magazines in terms of bottom-of-funnel metrics because they don’t work that fast. They are much more about building Brand equity and upper funnel. So the big studies of that era needed to take a pretty long timeframe. They needed to be in field for a year or more to actually be able to demonstrate the value, and the value often was a brand equity value. It wasn’t pushing product. Newspapers worked fast. You know, that form of print media, you’d have the inserts before the weekend. It was mostly promoting sales, so eroding your profit margin in the same way that other in-store promotions would, and ultimately undermining brand equity. The point of good magazine advertising was to build brand equity and pricing power.
So like a classic campaign that ran for over 20 years, the Absolut Vodka ad, was to me a great example of what’s different about print advertising compared to television or digital in most cases because it’s not interruptive. It works by invitation rather than shouting. It’s like, you want to put yourself in the picture? Oh, I want to be on that beach. I want to take that vacation. Or by being clever and witty, there’s a puzzle to solve. What have they done with that damn vodka bottle now? And, I mean, vodka is vodka, you know. But to be able to charge a couple of extra bucks because it’s Absolut is hugely valuable to that marketer. And so the game is a long-term game there. It’s not, and thus much harder to measure. And I think to the disadvantage of many advertisers that rely upon that kind of pricing power, it’s harder to sustain those forms of marketing these days because there is such a pull toward transactional bottom-of-the-funnel short-term metrics because they’re easier to measure. And they tend to be misattributed sometimes to shorter-acting forms of media that might have been, why did I search for that brand? But the search engine will get more credit than the advertising that made me type that brand’s name in the first place when I decided I wanted to buy something.
Driving Brand Equity and Subscription Growth
Andrew Mitrak: I want to come back to where we were in the story. You joined Time in the early 80s, and you continued to work at Time Warner and Condé Nast, and always in consumer research and insights leadership roles. And so what was your role in doing market research for major magazine publishers? Was it more looking at their own metrics, or was it looking at metrics for the advertisers, or what was your job there?
Scott McDonald: I set up the first consumer research department at Time Inc. And so the focus was almost exclusively on the demand side, on stimulating demand for magazines, working with the consumer marketing function and with the editors. And so a lot of work in magazine development, starting titles like Martha Stewart Living, Real Simple. Those were some of the ones that I worked on at Time Inc. And then there was a lot of magazine development work at Condé Nast as well, along with cover testing and developing forecasting models. You know, you have a couple different ideas for what you might run on the cover of Vanity Fair, which one will sell more. And so that was a key part there.
Condé Nast also had The New Yorker probably, a super influential magazine, one I still read all the time, very loyal to it. But the job there involved reducing its dependence on ad revenue and building up the consumer side of that business. So it really involved gradually getting people used to paying $150 a year for it instead of $50. And that was strategically vital to a magazine like The New Yorker, which isn’t a behemoth in terms of reach. And so it requires kind of a different mix in the business model. But yeah, at Condé Nast I had responsibility for the advertising side, but they hired me primarily because of my reputation doing work on the editorial and consumer side.
Andrew Mitrak: I make a lot of The Simpsons references on this podcast because I grew up watching The Simpsons. And I remember one of the first ways I ever heard of The New Yorker was a Simpsons joke where Marge is going through her mail and one of the envelopes was a rejection letter from The New Yorker subscription department. And I was basically a little kid, I was like, I didn’t even know what The New Yorker is.
And I looked it up like, oh yeah, seems like it’s this magazine for rich smart people. And it’s funny to think of how a magazine sort of segments itself. The New Yorker is different than Time, but there are some overlaps, right? That Time is on every newsstand, it has broader reach, it seems like it’s more ubiquitous, and The New Yorker wants to be big and everybody wants to know the name, but not everybody necessarily reads it or pays for it or subscribes to it. And I guess can you speak to the different approaches you had for how growing market share and maintaining market share for a very large widely circulated publication versus increasing the brand equity and justifying price increases and higher subscription costs for a more niche publication like The New Yorker?
Different Approaches to Managing Print Media Brands
Scott McDonald: Well, to some extent, I mean, some of it really is respecting the editors that you’re working with and trying to find a way to help them with the particular problems that they face. So a demand problem for Time (magazine) really involves something like newsstand. The New Yorker didn’t depend upon newsstand sales; it was a subscription magazine. So it’s partly just kind of understanding the differences in those businesses. And Time was probably in more need in some ways of the kind of research help that I could make because it did depend on newsstand sales. And that’s something where the forecasting tools can be of greater use and a testing program, particularly if you’re out every week, you get a lot of data points that you can then reconcile to how it actually sold and refine your forecasts.
So, but then a whole lot of times there’s a lot of news that happens. It’s not debatable what will be on the cover. It’s like what was the big story of the week. So your point of influence is more a slow news week where there’s what we would lovingly call a thumb sucker article. Just something that’s a bigger, in-depth piece that’s been cooking for a while and they’re looking for the right opportunity to run it. And for those they would really want to know some, it’s risk management. Like, how much will this appeal to people?
Andrew Mitrak: Did the business interests of increasing reach of say Time Magazine for instance influence editorial decisions as like who would be on the cover? Because I could imagine that there might be certain figures that you put that person on and it’s more likely to buy news, more people will buy it, right? Or you might have data like, oh when we put handsome people on the cover, we get more than... Did that ever...
Scott McDonald: It always is the editor’s choice. I’m just giving information. So there never was pressure from the corporation to, you know, just do what Scott says. It wouldn’t work well. It wouldn’t be good for the working relationship with the editors. It’s their remit. And so the principle of church and state was pretty much intact all the way along, and that would be, that wasn’t something that it was useful to challenge.
But there’s a lot of financial, a lot at stake. Or at least there was during that golden era. I mean the advent of this thing, completely changed the game because attention moved entirely to the phones. It hasn’t really left there yet. And people were no longer killing time at a checkout stand kind of browsing a magazine rack to figure out something to amuse themselves for the three minutes, the 2.7 minutes that they were in line waiting to be checked out at the grocery store. Yeah, so the forecasting became less valuable as newsstand just as a category declined.
Surprising Insights From Magazine Cover Testing
Andrew Mitrak: Are there any general insights or truisms that you’d be able to share about what are the markers of like, say who’s on the cover of a magazine and like this type tend to lead to a larger spike? Like what’s the type of insight you would share with an editor that they would choose to use or ignore or...
Scott McDonald: I’d say the things that are sort of durable truths, they didn’t need me for. I mean, put a Kennedy on the cover of People (magazine) and you’re going to sell. You can still, you can run JFK’s assassination 40 years later and it’s still going to sell. So I mean they don’t need me. They know that. But of course, if People magazine does this all the time, it’s not a good thing. They’ve got to find new things. So and Princess Diana, same thing. So there are cover subjects that for People or Vanity Fair (magazine) are pretty timeless.
The I’d say the better examples would be the ones that were surprises. Where they had some other strong options, but there would be a surprise that came out that they wouldn’t have automatically assumed. And so a test that would highlight that would encourage an editor to take a chance on something. And this would be true even for just an unusual shot that doesn’t look like the usual cover of Vogue (magazine). So a model or an actress in an unusual yoga pose or something, would be, or pregnant. Just something that is startling and feels a bit like a risk. And then you give the editor some idea of what is the level of risk and the probability of success for something that is out of box. And so I think it was used more for encouraging innovation and risk-taking than moving always back to something that was kind of a hardy perennial or too predictable.
Andrew Mitrak: So there were the truisms that were obvious, the JFKs and Princess Diana’s of the media. And then the things that were non-obvious that were unique insights that you were providing, were those sort of more temporal where you do that trick and then it sort of fades? I’m kind of almost likening it to people who analyze what’s trending on TikTok and social media trends and sort of seeing what are the types of stories that are this week. But you can’t, you kind of gotta hop on it now and it’s, this isn’t necessarily useful advice five years from now. Was it kind of like that type of thing?
Scott McDonald: One example I can think of from Vanity Fair was Heath Ledger. So it was like a year after he died. And they put him into the mix on a cover test. It wasn’t my idea, it was the editor’s idea, but it was against some other things that going into the test you plausibly would say, well these other ones have a pretty good chance as well. There’s no particular reason to think that people are still that interested in Heath Ledger. But they were, and it was quite evident. Now doing it a year later, it probably wouldn’t be the same. So these are kind of timestamped and the value of them is in being able to do that probe at the moment and fit it to a model where you’ve got other data on other covers and you’ve studied the competition and you know what their newsstand sales were. And so you can get that data back from the distributor. So you’re able to build a more sophisticated model because you’ve accumulated more data. And it was all great until the whole newsstand business collapsed in response to this more transformative launch of a smartphone and major change in consumer behavior. It’s part of what interests us right now on AI of course, and trying to get an early bead on this next transformative change.
Building Martha Stewart’s Brand with Consumer Research
Andrew Mitrak: I’m going to ask you about the smartphone and the internet and AI. I have one more magazine question before I do though. Because you mentioned Martha Stewart Living. And I think Martha Stewart might be one of the greatest marketers of all time. And I actually haven’t discussed her that much on this show yet. And just that there’s a magazine title with somebody’s name, Martha Stewart Living, there’s not that many of those. It’s not that, and to build a whole, and it seems like a unique thing at the moment to build a whole magazine around her brand. And do you have any stories of the creation of Martha Stewart Living or what was that about?
Scott McDonald: It was fun. The most fun part of it really was doing qualitative work, we did Focus groups with Martha in the back room. And it was of a genre of qualitative research where we decided that we really wanted to study the fans. So like this had worked very well for Warner Bros., so my confreres out at the Warner Bros. Studio, had this property Superman, that had been kind of damaged by this campy TV series in the 60s. And it wasn’t, they wanted to bring out a Superman movie that really worked, and they did it by studying the hardcore fans of the comic, of the original thing. So that was the approach with, and they managed to succeed in reviving the franchise for the movie.
That was our approach with Martha Stewart, we really tried to identify the people that just loved her. And that we studied what was authentic about Martha. So my favorite exercise from it was asking Martha to just from her, come from her house in Westport and bring us some stuff that’s in her house, that we mixed in with other things that were expensive, or utilitarian. I was like gardening gloves, or a little trowel, or just stuff that from her house, random stuff, compared to other stuff. And we threw it all on the table and asked people to pick out which things were Martha’s and why they thought that. And they could do it. They could do it. They understood her taste, some of which might be Shabby chic, but it was her taste and they were spot on. And it kind of helped the editors because here was a situation where Martha hadn’t made a magazine before, so she’s contracting with Time Inc. to boot up this magazine. And she’s got some professionals in magazine design and editors and things like that that she’s working with, but it’s a new venture. And that really helped to refine understanding of what the secret sauce was and this sort of passion for Martha. And I think it was a good example of, again trying to provide some information, but respectfully. I’m not a magazine editor, and you just set up the occasion as an opportunity to understand and refine the description of that brand and what’s the flavor of that magazine.
Andrew Mitrak: Right. Yeah, it seems like part of the core insight is really doubling down on the core fan base because if you’re making a magazine, you could sort of take it in different directions. And you can expand, if you have a lot of pages to fill, you could sort of dilute it and add a lot more stuff in. But instead, be like no, let’s really focus on what does this core group care about and try to get it to be the essence of Martha.
Scott McDonald: You know, and it’s interesting too because as we discussed before, the ads in a magazine are a pretty important signal of who’s in the room, who’s allowed into this club. So if you’ve got tasteless ads in a Martha Stewart Living or in any Condé Nast publication, you’ve got a problem. And it’s an editorial problem. I remember once at Condé Nast, the corporate sales department did a big deal with McDonald’s. And they ran McDonald’s in like all of the Condé Nast publications. And we got consumer complaints. “This doesn’t belong in my magazine. How dare you.” So there is an interesting balance that takes place that just has to do with the signaling about what’s appropriate for this particular environment.
The Early Days of the Internet
Andrew Mitrak: When did you first realize that the internet was going to be a big deal?
Scott McDonald: Right off the bat. The World Wide Web itself, which became I know was invented in 1989, but the first real operational browsers and effective implementation of the Web was in ‘95. And it immediately created a sensation, even though we were dealing with 300 baud modems and screeching sounds and all this stuff. Just the reality of having that amount of sort of global access to all these documents, was very bewildering. And for about four or five years, there was just a whole lot of experimenting taking place across all media.
Time Warner at that point had already been investing in Broadband and trying to pilot Video on demand. So they basically switched video systems, and it was they were too early. The technology was too expensive still, but I got to sort of play around with that. But there was recognition that something big was afoot, and people just didn’t know exactly what to do about that. And that was, that was a pretty fun ride.
Andrew Mitrak: Yeah, I imagine. It’s quite a ride. And so as a publisher, as the internet comes along, you know it’ll be a big deal, how does that impact your role as a researcher?
Scott McDonald: In some ways it led to just some interesting new things I had to figure out how to do. So again, because I came out of academia, I would constantly look back to see how certain methods, for example, for doing analysis and or forecasts, might apply in this situation. So my job at Time Warner kind of morphed into trying to understand the internet and the effect it would have on businesses. And so part of what I was doing was studying like what there’s a lot of complexity and chaos and difficulty finding things, and there were no good Search engines. So you’d start studying how people were actually using the available tools, AltaVista for example. And so it introduced me a lot to usability testing, user interface diagnostics, because internally people were designing things like more complex remote controls for TV for cable systems. And for proliferating channels of content. You’d start studying the dynamics of search and what led to satisfaction with a search result or not.
Time Inc. was experimenting with a satellite model that said, okay, we’re going to provide simplification, kind of like what AOL was at the beginning, where it was simplified into some aggregate content areas and you relied upon AOL or Time Inc. to filter all this stuff and make it simpler to find things because you’d aggregated content into kind of a hub. And part of my frustration was I wasn’t able to effectively convince the management of Time Warner that that was a mistake, and that that wasn’t actually going to win. That people wanted, they liked the freedom of all of those of being able to pull in documents from everywhere, and they didn’t really place enough value in that filtering design and structure. So Google would win. And as soon as Google showed up, Google didn’t even have a business model yet, but it was clear from day one when you’re studying that space that this is a significantly better search result. And you could see immediately that this is where Time Warner should be focusing its attention and not Pathfinder or something like that that was. Or, and it was the AOL deal, when that was announced, the merger with AOL, that was when I decided I was going to leave Time Warner. Because it seemed to me to be completely contrary to what I’d been learning.
Andrew Mitrak: Seems to have been a prescient choice.
Scott McDonald: Yeah, personally it was fine. I had a lot of options that became very valuable in that transaction and I could exercise them and walk away a happy camper. But it seemed like a very bad business proposition.
From the Walled Garden to The Open Web
Andrew Mitrak: Yeah, for sure. And it seems like the Time Warner AOL merger and sort of their Walled garden approach as opposed to sort of embracing the open Internet it seems like it also kind of ties back to their own business interests in being gatekeepers. And that if there weren’t gatekeepers that has sort of knock-on effects that might be bad for the publishing industry that sort of played out over the next couple decades.
Scott McDonald: It’s the Innovator’s dilemma.
Andrew Mitrak: Yeah, exactly. Did publishers start to see the writing on the wall there or when did that, when did because I’m sure there was a moment where the internet’s like, hey this is a huge opportunity, this is more, you know, free distribution, we don’t have to pay for paper, things like that. But then there seems to be like, oh but what if anybody can blog and what if people stop going through the gatekeepers? Like when did that turn or did you see that turn?
Scott McDonald: That was more in the 2000s. So it was really when I was at Condé Nast, and Condé Nast was wrestling with the same issues. In some ways it had a pretty big portfolio of brands, but it ended up pruning those to the most distinctive brands that could be defended and that could operate as digital properties on a global scale. So they kind of shifted scale and integrated their international, like they used to license Vogue (magazine) in a bunch of different countries, and they kind of consolidated and it became a global brand more. And would be sold, the advertising would be sold on a different basis thus. So there were different forms of adaptation, they all needed to figure out how to do what they were doing on a lower cost basis because the impressions became more commodified in that market. Particularly once Programmatic advertising took place.
And you know the, I mean the big change, the biggest change in my view was that advertising was severed from editorial content. Ads came from Ad serving. Advertisers bought an audience, they didn’t buy a placement inside a medium. And so the whole model and the kind of special relationship that I described where I’d be going over to L’Oréal and talking about our view of their customers, trying to share insights about customers that are gleaned from studying them in the context of Condé Nast magazines, was irrelevant because everything was much more commodified through that digital model of advertising insertions. The same issues are with us now with AI and you have different companies trying to decide do I license, do I make a deal with OpenAI right now or do I try and sue them, you know like The New York Times is doing, and require a different payment model for access to my content. And these are still commercial and legal questions that are not yet resolved but they feel familiar because they’re just a different iteration of the same business issues that developed in response to the Web.
Applying Lessons from the Internet to the AI Era
Andrew Mitrak: Yeah absolutely. Are there any other lessons that you’re drawing or thinking about from having navigated the internet’s disruption to the publishing industry and as we’re now entering or in the midst of this AI era, what that means for advertisers and marketers? Like are there any lessons that you’re thinking about that apply?
Scott McDonald: Yes and no. I mean I think the in some ways this feels somewhat different. And I don’t know, you know the question of whether AI dramatically changes the consumer the labor market, and the ability of people to earn incomes that supports the advertising system is a fair question. Even though the history of all these tech innovations is that they generate enough new jobs to replace the ones that have been rendered obsolete. But I don’t know at this point whether whether I believe that this time around. So that’s a fairly big unknown that would be different in terms of the consequences of the innovation.
If I was still working at a magazine publisher and or a publisher in general, it could be a TV channel that calls itself a publisher now, or any content engine, then I’d still be wary of how I monetize that content when it becomes Disintermediation. My advice still would be pay a lot of attention to trust and pay a lot of attention to the shifts in consumer behavior because advertisers always follow the consumer behavior. And consumers don’t always do what we as publishers want them to do. So you’ve got to be realistic about that and keep your eye on the consumer. That’s certainly a lesson I think from my Time Warner days where I don’t think they did that sufficiently. So.
Andrew Mitrak: I don’t know if this is a lesson, but something to draw from the golden era of publishing is editorial taste, that as a marketer that uses AI products, the AI products don’t always have good taste, right? Or they kind of have sort of a median internet quality taste and like, you know, obviously they’re very powerful and all that but like there is an element of if everything kind of looks the same, and you can’t differentiate your AI output from my AI output, somebody’s editorial taste on refining and coaching and directing it kind of becomes more important. And I wonder if there’s sort of people embracing their inner editor and developing taste to sort of know what’s good and not...
Scott McDonald: You know, this remains to be seen but it’s my observation that as AI improves, which it continues to do with breathtaking speed, it depends partly on you as the user to tell it what you want. It wants to please. So if you, so like in the context of say marketing applications or insight extraction, if you just ask a simple question, you’re going to get a pretty simple answer. If you actually feed it say peer-reviewed academic articles that you want a theoretical framework to be incorporated into the answer, you’ve raised the bar a lot. If you tell it that you want it to pretend that it’s a McKinsey & Company consultant, it’ll do it. It knows what you mean, and it will change the answers in response to your inputs. So I don’t see any reason why you couldn’t do that with regard to some matters of taste. If you could train your chatbot to be like those focus group respondents in the Martha Stewart Living example. And it seems in principle that you should be able to cultivate that.
The Advertising Research Foundation (ARF)
Andrew Mitrak: So I want to ask you about the Advertising Research Foundation (ARF). You’ve been president of the ARF for about ten years or so. What is the ARF for people who have are not familiar with it already and how has it evolved over the years?
Scott McDonald: Okay. So the ARF is the Advertising Research Foundation. It is celebrating its 90th birthday right now. It was founded in February of 1936. As at the behest of the two founding members, the Association of National Advertisers, the ANA, and the American Association of Advertising Agencies, otherwise known as the 4A’s. And it was set up from the beginning as an independent foundation dedicated to furthering through research the scientific practice of marketing and advertising.
So from the beginning days it wrestled with the kind of public facing questions of how advertising works. What’s the best way of measuring the audience of a Life (magazine), you know? Of not just the circulation but all the readers per copy and the people who look at it in barbershops and whatever, you know. What’s the best way of measuring the audience for a radio program? We know how many radio sets are in American households, but how many people actually heard a particular show? And then in terms of advertising, what makes some ads successful and others not? What’s the optimum frequency? How long does it take to burn in or to burn out? Those questions have been with us from the beginning, and they’re still with us today, it’s just a much more complex and fragmented media landscape.
And so to some extent you need to update that all the time. And that’s still the kind of role of the ARF. It’s the power according to its bylaws, the power over the organization is distributed among marketers, ad agencies, media companies, and service providers, which would include all the measurement companies and everybody from Nielsen Holdings to little Neuroscience consultancies or brand consultancies or attention measurement companies, any of those things. And so ARF is kind of the Switzerland in the middle of that ecosystem that conducts research on basic questions of how advertising and marketing work, trying to stay as close as possible to the values of scientific inquiry. And that means, that doesn’t mean anything goes. And you’re in an environment where people make a lot of claims. All these campaigns do really well. You go to a lot of conferences and they’re all just like success story after success story. And you know not everything works, you know? And so trying to separate wheat from chaff and kind of build a body of knowledge about how to think about these things is the mission. To try to improve practice through the application of scientific methods. So.
Andrew Mitrak: How do you, how do you deal with that at a conference or just in marketing in general? Because I think every marketer wants to say that they’re scientific. They want to say that they’re data-driven, but also every they want to say that their campaigns are working, right? They want to say yes and our campaign was great and there’s sort of a grading their own homework type thing. And there are ways where you can cherry pick your numbers, like “oh, our reach was great,” even if your conversion was bad. Or “conversion was great,” even though you paid too much. And I guess how do you sort out navigate that?
Scott McDonald: Yeah, it’s difficult. Partly because, you know, the association itself, it’s a membership organization. So you don’t really want to offend your members. But on the other hand, at some level you might have to because not everything can be equally true. So that’s why the north star remains. And you try and set up... I mean a classic ARF study, we just did it around different aspects of attention measurement. This is a growing field. And you have different approaches, some of which rely upon academic understandings of cognition and memory and things like that. And others that really kind of just follow the development of tools that might plausibly be used as proxies for attention. So eye fixations, because we have Eye tracking and good cameras on our digital devices, on our phones, on our laptops. You’ve got information that’s used for ad verification purposes that would indicate that yeah, there’s a human there. There’s a hand on the mouse. You know? So that’s a proxy for some level of attention that is a signal not very expensive to collect because you’re already doing all this ad verification work, but how closely can we establish that that relates to any sort of formal definition of what we mean by attention? And by that are we talking about, you know, just eye fixations and Saccades? Are we talking about evidence of memory and recall around an ad?
So there’s a lot of tests around that. And the ARF exists kind of to help sort out the quality of those. We have an academic journal. We connect to people who have, you know, where they’re peer reviews. There’s competition to get on the stage for our events. So people have to compete before a jury to even get a slot. And so it’s, it’s sort of through that process, which is similar to how it works in the other sciences. I mean, the best examples, if someone really wants to make a strong claim for their research, then they would, we’ll do an audit for them. We’ll run through and see whether we can replicate their numbers. We’ll see whether they did cherry pick. We will, and then we’ll take their data and host it on our website and make it available for anyone in the world who wants to have a go at it, to anonymize the data and, which is the same like if California Institute of Technology wants to make a big claim in the physical sciences, they got to make their data available to the team at Massachusetts Institute of Technology to build legitimacy around it. It’s a very similar concept. So that’s the space that we operate in. It’s geeky but it has some value in this ecosystem.
Andrew Mitrak: Yeah. If I was to draw an analogy back to earlier in this conversation of you and your publishing days recognizing “hey there’s truisms that JFK assassination and Princess Diana, that always sells magazines at newsstands.” But like the real insights are sort of the non-obvious things that are more unique or maybe more time-bound. Could you draw parallels and find like what are sort of the truisms that the ARF has helped establish or that you’ve sort of recognized over the years in your role there, versus some of the more unique, non-obvious things that research is uncovering?
Scott McDonald: There are a lot of them, I would say. We codified some of them in our, so the ARF acquired the Marketing Science Institute, which is a more academically oriented entity. We did this a couple of years ago. And MSI has published something they call the Empirical Generalization series, which only will, so it will formulate like “X causes Y.” And here are the estimates of effects, within this range and these categories, you know that might be covariates. But it’s reduced down to things that we think there is compelling enough evidence. And their filter on it is wherever there’s been a meta-analysis in like the top three or four marketing journals. So very high level of peer review scrutiny. And only where there have been 60 or more studies confirming this generalization that would allow you to talk about say the if you’ve got like a budget to spend and you need to spend some of it on advertising and some of it on price promotion, for example, in-store promotion, like what are the trade-offs and how do we think about that?
So but I think for the ARF itself, probably the thing we’ve studied the most over the years, is anytime there’s a recession or a big disruption in the economy, the pandemic, September 11 attacks, any of these things that suddenly just have a big dramatic effect on markets and consumer behavior, there’s a tendency to cut marketing spend. Short-term marketing spend gets cut. So what’s the effect of that? Since we’ve studied it like from the Great Depression, World War II, the Korean War, any of these things that have these kinds of shock effects. And you’ve got a pretty good record of it. And the answer to it, I call it an empirical generalization, is that when you cut your share of voice, so you withdraw from the advertising market and don’t spend, so you’re not really getting a share of voice within a category, you lose share. And you lose it fairly quickly, and it takes about five years to recover, if you can recover. We have had whole brands that just kind of go away because they lost their position within a category.
That’s connected to another generalization and truism that I think is there and is likely to remain there for a long time, that being the dominant brand in a category, which usually involves at least 20% share of market, although in some cases it’s a lot more, leader in the category. That leads to all kinds of benefits. Any advertising that’s done for the brand leader in a category has stronger coefficients of impact, both short term and long term. And to the dismay of the second and third or fourth participants in a category, their advertising is probably going to actually benefit the category leader. It’s an unfair world, but people just mistake it. And a lot of, it’s another sort of truism that I think remains, a lot of creative ads that are so creative that they don’t tell you who the brand is, people love the ad and they assign credit for it to the wrong brand. Because that truism was ignored. It might have won an award somewhere in an ad creative competition, but it didn’t really work for the brand because they didn’t integrate the brand, make it clear enough to the consumer what brand was being advertised. So there are a lot of regularities and it’s hard to not be like a broken record sometimes when you’re responsible for the catalog of those things. But there are mistakes that we shouldn’t be making over and over again. And I think MSI in its most recent iteration of the Empirical Generalization series had like 175 things that rose to the level of, okay these are generalizations. There’s like enough evidence, there is consensus around it. And that’s kind of how in my view science works. It still doesn’t mean that those won’t change and evolve over time as other situations develop, but you build it on the back of a lot of evidence that’s been objectively evaluated and critically evaluated. So.
Andrew Mitrak: Yeah. That’s great. It’s great that your foundation is able to advocate for this research, make it available and share it. So let’s learn from science, let’s learn from history and not repeat the same mistakes over and over again. So Scott, I really enjoyed this conversation. For listeners who have enjoyed it as well, where would you point them to online so they could find out more about your work and more about the ARF?
Scott McDonald: thearf.org and msi.org.
Andrew Mitrak: Scott, thanks so much. It’s been a real pleasure.
Scott McDonald: Thanks Andrew.
