Advertising, entrepreneurism

That joke’s not funny anymore

“[Y]ou have to assume that humans are capable of looking at facts, finding root causes and formulating solutions. On my planet there’s not much evidence to support this assumption… If humans had the ability to look at facts and make good decisions, think about how different the world would be. There would be only six kinds of cars on the market and nobody would buy a car that was second best in its price range. There would be no such thing as jury selection since all jurors would reach the same conclusion after viewing the facts, and all elections would be decided unanimously. That’s not the world we live in. Our brains are wired backwards. We make decisions first–based on irrational forces and personal motives–then we do the analysis. The facts get whittled until they fit into the right holes.”

— Scott Adams*

[The first section of this post was previously published on AdExchanger, and there’s an excellent comment thread there, so you should go read it. This is the rambling version.]

Jokes all start with one of a few stock set-ups. “A man walked into a bar.” But the punchlines are all different. VC pitches are the opposite: the set-ups are all different, but the punchlines are all the same.

Stop me if you’ve heard this one before. “We enable the half of advertising that is not yet online: brand advertising!” Been hearing that one for more than ten years now. It permeates the hopes and dreams of every adtech entrepreneur and investor. And still no one has cracked the code.

My friend Tim Hanlon got together with Tim Chang a few weeks ago over on AdExchanger to offer some reasons why this might be. Worth your time, but here’s the tl;dr**:

  • No standards or consistent measures of “success” other than outdated or inadequate metrics like CPM and CTR;
  • Limited real-time intelligence;
  • Unsuitable display ad formats; and
  • Lack of creativity in formats.

The prize is huge. As Tim and Tim point out, two-thirds of advertising spending is brand advertising, but online only one quarter is. In fact, if brand advertising dollars moved online in the same proportion that sales advertising has, it would almost exactly close the famous gap between time spent online and ad dollars spent online. The $50 billion gap that Mary Meeker mentions is exactly equal to the missing brand spend.

So I understand the urgent desire to figure out online brand advertising. If we did, we’d more than double the online advertising market. Online pubs would rejoice, online marketing pros would have more excuses to go out drinking with prospective clients, my portfolio value would quintuple overnight. Good things all. And I appreciate the optimism that Tim and Tim have, their willingness to keep suggesting solutions. But I think it’s the triumph of hope over experience. Each of these things has been tried, and tried and tried. And still we believe that this time it’s different, that this year an online branding play will work. Online video maybe, or Facebook, or Pinterest. Every new company is touted as the one that will make branding work online.

But what if we try all these things, like we’ve tried everything before, and they don’t work? What if we eliminate all the possibilities and what remains is… nothing? I’m going to be branded a heretic for saying this, but what if online just doesn’t work for branding?

I mean, not to be defeatist, but we understand branding pretty well. Marketers have been creating brands nigh on one hundred years now, it’s not a black art. And the solutions I hear, even Tim and Tim’s, are not untried. More, they are not what makes brand advertising effective in other media. I don’t buy that these are the solutions. I think it’s distinctly possible that there are no solutions.

Maybe the medium itself is antithetical to the way brands are built. Like direct mail, maybe the very fact of delivering your message in a low-budget, specifically targeted way can not in any way build a brand. Brands attempt to exist autonomously, they are objects of desire, they want to distinguish what otherwise is indistinguishable. The psychological processes of branding are inimical to the idea that the brand has been chosen for you. Brands do not choose you; you choose brands. Brands are aloof, they aspire to be the Platonic ideal, their competitors just shadows.

Perhaps. I mean, I could be wrong. It could be that even though we sell ourselves to clients as brand-building geniuses, we don’t know what we’re doing; that we’re groping in the dark, throwing random darts and just haven’t hit the bulls-eye yet. We could be a bunch of monkeys at typewriters and Shakespeare will just roll on out one day. Could be. But it seems unlikely.

What if I’m right? What if online branding is a mug’s game? If it is, it won’t be too much longer before marketers get wise and just stop listening to online branding pitches. Maybe they already have. Maybe they never did listen to them. What’s the fallback plan? How do we go about getting the brand advertising dollars if online brand advertising doesn’t work? What can we do that will cause brand advertisers to move their branding dollars out of advertising altogether into some other online channel? How do we disrupt branding?


What is branding, really? Why does it work? For the marketer, branding is a way of wrapping all of a product’s attributes in a neat package and giving it a handle so they can refer to it easily. For the consumer, brands are a shortcut: for products where the potential benefit of making a choice is smaller than the cost of choosing, a brand is a fallback. Brands are Scott Adams’ whittled down facts, except that marketers have done the whittling for us so they can control the outcome to their advantage. The hole they fit into is your brain. And research shows people only have a few of these holes in their brain.

This can work in a couple of ways. A consumer confronted with a dozen pasta brands in the spaghetti aisle*** would have to expend some time and effort deciding which was best for them. The small potential benefit from finding a better quality pasta is less than the cost in time and effort to determine this, for most consumers. So even if De Cecco is a better pasta, it is a rational for someone familiar with Ronzoni to buy Ronzoni. The pastas are similar enough that someone who just wants a bowl of spaghetti should not expend any effort distinguishing them: just choose your brand and move on.

On the other end of the spectrum, some things are extremely costly to evaluate. Choosing a motorcycle, for instance. The variables that come into play include not just the motorcycle itself, but complementary goods like service quality and availability of third-party components (if the stock pipes are too tame, say.) There are also intangibles, like aesthetics, community, and how you will be perceived among your peers if you are riding a Honda instead of a Harley. A brand can ensure that, even when the objective technical qualities of the bike itself are the same or inferior, it has an advantage among certain consumers because the cost of objectively evaluating differences between bikes is, for most people, impossibly high.

This cost/benefit analysis masks another obvious aspect of branding: risk-mitigation. I have, driving down the highway with my kids, chosen McDonalds though this would be at other times not on the list of possible dinner spots. But I know exactly what I’m getting, how long it will take, and how much it will cost. Because there’s a fixed cost of investigating new options, even if for one-time use, the risk/reward curve is not linear.

The answer seems obvious. If branding is a needed compensation for something our brains are just not good at, a low quality way to reduce search costs, an easy alternative to remembering tons of facts, then the answer is to provide a better, more efficient way to sort alternatives. The internet, in its ability to instantly connect you to huge data sources and extremely fast algorithms no matter where you are and what random question you’re asking, seems to be the perfect answer.

This is what computers are good at. Lots and lots of data, changing prices, personal utility curves. Right now I might walk into the supermarket looking for a relatively healthy breakfast cereal that my kids will eat. Given the huge number of choices, I might settle on Frosted Cheerios (even though they’re probably about as healthy as a glazed donut) because Cheerios has pounded the idea that they are healthy into my brain. I can imagine, instead, walking into the supermarket, scanning the Lucky Charms with my smartphone and asking it to rank healthier alternatives for me. I can imagine a world where I tell an application what I like and dislike about my current pair of sneakers and it recommends a pair that would be better for me. I can imagine a world where I enter the specs and spec tradeoffs for any good imaginable–skis, cars, laundry detergent–and my computer finds me the best match.

But I’m not leading you down a garden path here. I don’t do Socratic dialogue. I do not know the answer to the question I’m asking. I do not think these ideas will work because nobody will pay for them.


Generally, and certainly with advertised goods, the seller is the one paying to find buyers and not vice-versa. This has resulted in all sorts of market distortions. Sellers are motivated to sell, and not necessarily only if the product is right for the buyer. That the seller is paying to find buyers–any buyers–instead of the buyer paying to find the perfect seller is a bit of an historic accident. The media was once solely a broadcast mechanism, a mass-reach vehicle. Before the internet, seller-financed advertising was cost efficient while buyer-financed search was cost prohibitive. Even though that is no longer necessarily true, advertising is stuck in a local maximum.

It would certainly be more systemically efficient today if consumers decided what they wanted and then went out and searched for their best match themselves. Then advertising would be less effective so there would be a lot less of it. If sellers did not need to advertise, they could lower the cost of the product and this lower cost would–I’m guessing–more than compensate buyers for the time needed to find the right product. The buyer would end up even on cost (lower product cost ~= higher search costs) but with a more appropriate product. That’s the ideal world, but it requires massive behavior change from both sides of the market at once. There’s no way to achieve that kind of coordination.

As long as advertising continues to be cost-effective, sellers will advertise. As long as they advertise, they will not lower prices. And as long as they don’t lower prices, buyers would have to pay twice if they decide to do the search themselves: once for the advertising and once for the search. This is a long way of saying that no one except the sellers themselves will pay for anything to do with informing consumers about products and services. We are stuck with what we have, efficient or no.

Want to quibble? We now have some seventeen years of evidence that, even if it’s a better way of doing things, consumers will not pay for search in any way other than by looking at ads. This is, if you think about it, probably the most bizarre thing about internet marketing. People pay for media that compares and analyzes products, but in a way that undermines the value of these comparisons. Yelp, Google (like Car & Driver magazine offline) critique the very industries that finance them. Their interests are in question. Marketers are attempting to influence people right as they are trying to make uninfluenced choices. Search for a product on Google and you are inundated with ads. Odder, many are clicked. People are searching for someone to convince them, they are going through the motions of choosing and then avoiding making choices. The media soothes the cognitive dissonance with a pleasing veneer of objectivity, but the objectivity is–has to be–a sham. Follow the money.

The non-profit Consumer Reports has overcome this criticism by refusing to accept advertising. But they may be the exception that proves the rule: despite almost certainly being worth the subscription price for anyone who buys even one thing in any category they cover, they only have some seven million monthly subscribers. Consumers will not pay to inform themselves. That’s why we’re stuck with marketing.


Some other, possibly spurious, correlations to note:

TV Print Online Display
Measurability Low Medium High
Involvement Absorbed Absorbing Engaged
Audience Mass Select Targeted
Branding Yes Sure Not so much
CPMs High OK Low

Many intelligent observers, when contemplating low CPMs or recalcitrant brand advertisers say we just need more measurement, more engagement, or more specific targeting. But these things seem to go the wrong way. On the other hand, there are some exceptions: the trade press is more targeted and has higher CPMs; search is more engaging and has higher (effective) CPMs; etc. So the point here is not that we’re doomed, but that the easy answers will not do–we’re probably analyzing success along the wrong dimensions.


Regardless, I believe in the power of the internet. I think that we can achieve a better product-consumer match by using personalization, community, data and machine learning. In fact, this seems almost too obvious to say. The internet has the power to create a much better branding mechanism: one that works better for brands and for consumers.

I also believe that brands can be valuable. They are proprietary marks, so can guarantee implicit promises and ensure repeat business. They allow trust, and trust is a necessary lubricant for commerce. If there were no newspaper brands, no one would read newspapers, because they would not be able to judge the quality of the news they were reading. If there were no retail brands, every purchase would be like walking into a generic electronics storefront in Times Square: buyer beware.

My objections are not to the internet or to branding. My objections are to the way we are approaching disrupting**** branding. I do not believe the success of online brand advertising is about waiting a bit longer, or measuring better, or creating more engagement. We’ve waited long enough, we measure better than any other medium, and we are as interactive a medium as they come. If you are espousing those ideas, then you have to also explain why you are right now when you would have been wrong all these long internet years. Things do change, but sudden change is either because of a compounding effect or a catalyst. It does not look like to me that online brand advertising is increasing in an exponential way. Nor do I recognize a catalyst*****.

Or, and I think this is a more promising path, we need to accept that branding may not change to accomodate us, we may have to change to accomodate branding. No more complaining that brand marketers just don’t get it. No more waiting on incremental change in measurement or attribution technology. Find a way to allow brands to hone and prove their promises, while giving them a much larger payoff for doing so. Don’t think about how to service Procter & Gamble or Coca-Cola–disruption starts off by creating new markets, not servicing old ones–think about how you could help a quality product or service build a brand from the ground up for far, far less than a TV branding campaign would cost. If you do that you will have the big brands’ attention, and an amazing business.

* Quoted in Bruce Tremper’s Staying Alive in Avalanche Terrain, a must read if you like the backcountry in Winter.
** Seriously? What are you doing here?
*** I was in a supermarket in a hispanic part of Pennsylvania recently where pasta was in an aisle labeled ‘Ethnic’ while the Goya black beans were in an aisle labelled ‘Beans.’ Where I live in Hoboken, the exact opposite is true. I wonder if there’s a place where the ethnic aisle stocks Oscar Meyer and Easy Cheese.
**** I wrote a typically long blog post on disruption and what it means in this context here.
***** A catalyst has to be a new technology or an entirely new way of utilizing it. The internet itself, say, or social media, or collective intelligence, or online video, or data-driven matching. These could all have been catalysts but, as it turned out, they were not.