Digital has ignored the theory of signaling — to the detriment of both advertisers and society. Where a customer is, is crucial.
In a column of January of this year, journalist Walt Mossberg relayed a storythat probably sounds all too familiar to many publishers:
About a week after our launch, I was seated at a dinner next to a major advertising executive. He complimented me on our new site’s quality and on that of a predecessor site we had created and run, AllThingsD.com. I asked him if that meant he’d be placing ads on our fledgling site. He said yes, he’d do that for a little while. And then, after the cookies he placed on Recode helped him to track our desirable audience around the web, his agency would begin removing the ads and placing them on cheaper sites our readers also happened to visit. In other words, our quality journalism was, to him, nothing more than a lead generator for target-rich readers, and would ultimately benefit sites that might care less about quality.
Digital has made targeting the holy grail of advertising. We have discussed in previous columns why this is a flawed approach for big brands. For a large brand, targeting is largely useless: every consumer is a potential customer for your brand. Furthermore, the additional cost of targeting technology would offset any potential gains. And third, even if ad tech successfully nails the challenge of attribution (figuring out which ads worked), there’s no reason to assume that that information will definitively tell us that the ads that worked are on the internet. We may well learn what we already suspect: that television ads work better.
Revisiting signaling
But today we are here to talk about something that is closer to the publisher’s immediate concern: one of the most unique, surprising, and established theories of advertising economics. One that has, by and large, been completely ignored when it comes to digital. And one that speaks to the heard of why the advertising executive with whom Mossberg was speaking was — like many digital ad buyers — profoundly misguided. We are talking about signaling.
In 1974 economist Phillip Nelson published a landmark paper entitled “Advertising as Information” with a somewhat McLuhanian pretense: that the act of advertising is information in and of itself. This concept leads us into fascinating areas of thought around advertising and economics, and gave rise to the concept known as signaling. Simply put, the very fact that a product is heavily advertised, regardless of its message, is evidence to the consumer that the quality of the product is high. Nelson got to this point by delving into the informative view of advertising and asking himself “how exactly does an ad convey information? And does an ad convey information even when it is not explicitly, verbally stating a fact?”
He posits that it does: that the very act of advertising is useful, and consumers can derive information out of its mere existence. “The consumer believes that the more a brand advertises, the more likely it is to be a better buy [emphasis added]. In consequence, the more advertisements of a brand the consumer encounters, the more likely he is to try the brand.”
In some ways, this is not a unique notion. Economist Emily Fogg Meade noted as early as 1919 that “Many people are impressed with the mere fact of advertising.” But Nelson broke down that sense of being impressed into component parts and developed a robust theory around the tendency for people to be impressed with advertising. “The consumer can learn that the brand advertises,” he said. “I contend that this is the useful information that the consumer absorbs from the endorsements of announcers, actors, and others who are paid for their encomiums.”
Signaling is kind of a crazy theory at first blush. But the more you think about it, the more it makes sense: the best products sell the best. They can charge a decent price and make a good profit. Which means they have more money for advertising. Which means they can spend more on advertising. So, as a consumer, you can be forgiven that assuming that a product that spends a good amount of money on advertising is probably a pretty good product.
The theory of advertising signaling was of huge importance in the mid twentieth century because it was one of the first formal explanations of why advertising could work even if it wasn’t informational. Economists love what is called “the informative view of advertising” — that advertising works because it provides additional information to the consumer. They love it because it fits neatly into their view of consumers being rational and everything in economics having a rational reason. Yet they weren’t so blind as to realize that a lot of advertising wasn’t — strictly speaking — informational. An ad that says “Just do it,” or “Think Different” isn’t really conveying a whole lot of information. Then signaling came along and said “Hey! yes it is! it’s conveying that they product can afford to advertise and so is probably pretty good! no matter what it says!” Keynesian acolytes Edward Hastings Chamberlain and Joan Robinson had told us, back in the 1930’s that advertisers had two different types of ad spend — direct and brand — but they didn’t really go into why brand advertising worked.
Signaling addressed that. And everyone was happy.
Like most economic theories of yore, however, advertising signaling was just that: a theory. And the world is full of economic theories that don’t stand up to empirical investigation . Signaling, a theory borne of the 1970’s, didn’t have to wait too long before the rise of econometrics — the study of large amounts of data to analyze economic theories. And before too long, a robust body of work came along proving that, as crazy as it sounds, yes, consumers can intuit quality of a product based on how much money that product spends on advertising.
We turn, then, to the work of Amna Kirmani and her various associates, which I find incredibly interesting. In a series of papers, Kirmani et al set out to discover not only if higher advertising spends indicates higher product quality, but whether the consumer can perceive this higher advertising spend and make use of the information. She finds that in “the absence of other information about a new brand, people may use the amount of advertising as a signal of quality. They may infer that the firm is exerting effort in promoting a product to which it is committed or that the price of the product is likely to be high.” I find this to be very interesting: consumers, without knowing much about the advertising industry in general, have the ability to sense a large ad campaign, and understand that a brand is willing to spend on a large campaign when it believes in the product. Not only that, consumers can intuit this spend relative to its competitors. That is, they don’t think in dollars and sense, but rather “Oh, Nike is spending more on this shoe than they usually do, and more than Adidas. They must believe in it.”
The consumer, it turns out, is incredibly smart about knowing whether a brand is skimping or not. Consumers are smart enough to intuit ad spends, and better still, they intuit them in comparison to competitors. They don’t need a published budget. They can tell if you are spending more. They also can infer higher quality from certain expensive advertising tactics, including the use of celebrity endorsements, blockbuster special effects, and really expensive media placement such as the super bowl. These findings are echoed by Pamela Homer, who in 1995 confirmed Kirmani’s findings and found that the ads don’t even need to convey quality-related information to be effective in imparting quality levels to the consumer.
(As an aside, Kirmani et al caution that there’s a limit to this madness — what I call the “Zune effect.” It is possible for advertisers to spend so much money that consumers can tell that they are overdoing it, and trying to substitute ad spend for quality. Brands can, therefore (and not surprisingly) take on an air of desperation with high media spends. “The desperation undermine occurs when the amount of expenditure seems excessive or more than reasonably warranted to convey product benefits.”)
While Kirmani et al were focused on consumers intuiting ad spend levels, a parallel set of research was taking place confirming that yes, in fact, consumers can intuit quality from that ad spend.
In 1983 Robert Archibald et al set out to find out whether advertising and quality were related when the consumer had access to otherwise useful information regarding a product: In this case they use the published ratings of shoes in Runners World magazine. They find that “ quality and advertising are much more likely to be positively related in the presence of quality ratings.” Archibald et al point out that the existence of ratings “has a major impact on the advertising behavior of firms, which makes misleading advertising much less likely.” Now, the existence of ratings can seem like a big caveat, but in this day and age of Amazon reviews for virtually every product, the caveat becomes less and less meaningful. They emphasize, too, that once a category has published ratings, “this effect is so strong in our sample of data on running shoes that, after the ratings’ publication, advertising levels are not only good indicators of high quality products, but also good indicators of good buys.”
Finally, an interesting finding of Archibald et al is that you, as a consumer, don’t even need to review the ratings once they are published. “Our results show that individuals who consult only [emphasis added] advertising levels to gather product information…will get much more accurate information after the publication of ratings.” The implications of these findings in a post-Amazon world are sort of staggering.
The confirmation of signaling theory has been repeated, several times, with several different methodologies. In 1988 Gerard Tellis and Claes Fornellconfirmed it using using an analysis of the PMS database. In 2004 Rolf Fareand his colleagues do the same by using extensive analysis of internal corporate data in the beer industry.
This is a lot of references. But that is the point: this is not some half-baked “theory” espoused to prove some pre conceived notion. It has been empirically proven, time and time again: consumers can smell a cheapskate advertising campaign, and their impression of your product suffers for it.
Signaling and Digital
Returning to Mossberg’s anecdote, then, you can see where this is going. That advertising executive was, in short, wrong. Buy buying a cheaper publication in lieu of Mossberg’s AllThingsD, he or she was signaling to customers that their brand was cheap and, thus, probably inferior.
And if a consumer can tell a brand’s spending level across the whole world of ABC, NBC, billboards, The New York Times, People, train stations and the Super Bowl, they sure as hell can tell that advertising on some podunk perhaps-robotically-generated blog or app isn’t as expensive as advertising on AllThingsD. And they can — and do — judge your product for it.
Digital advertising has relentlessly followed a single theory of advertising — targeting. It is a theory that is nowhere near as well-proven as signaling. Targeting is a far more empirically unsound theory: indeed, most work on the subject has found it isn’t particularly effective for large brands (I refer you specifically, again, to Byron Sharp, Andrew Ehrenberg and John Philip Jones — the people who have done or are doing the most empirical research on this topic.)
Signaling, by contrast, though weird sounding at first, is on far more solid empirical grounds. Yet our entire digital ecosystem is predicated upon throwing it out the window: we assume the venue — the publisher– is irrelevant. “Target your customers where they are.” “Hit your target for less money.” Hit your target — over and over that’s what we hear, as if it’s the most important thing in advertising. But it’s not, and it’s not even proven it matters. Meanwhile, we — personified by Mossberg’s ad exec conversational partner — completely ignore that where we hit our customer matters.
So instead of the best places to advertise getting money for it, we instead stiff them, and give money to a crappy place instead. We justify it by saying we’re giving the crappy pub less money than the good pub, so we’re saving money.
This is analogous to saying I’m giving my money to a slot machine instead of putting it into an interest bearing savings account because the slot machine is cheaper.
Except in this case, the publications – quality information and news — suffer as a result.
Not only are we screwing over good media, we’re not even getting anything for doing so. That sounds like a pretty bad deal all around.
