Ad Tech’s Impossible Mission
Why ad tech won’t be the internet’s savior in its battle for television dollars.
When it comes to the promise of the internet conquering advertising and solving Wanamaker’s dilemma of “Half the money I spend on advertising is wasted, the problem is I don’t know which half,” the internet’s main weapon in its arsenal has always been advertising technology, known in the industry as simply “ad tech.” The improved technology of advertising was to be the secret weapon that would bring additional efficiency and solve Wanamaker’s dilemma. To this day, financiers and other bullish proponents of the internet claim that all of the ad money spent in our economy will inexorably migrate over.
While the internet has recently begun to expand past its advertising roots, advertising dollars still comprise a massive share of all internet-related revenue — to the tune of over $200 billion worldwide. Moreover, advertising shapes the very products of the internet. As we’ve mentioned before, ad spending has exceeded VC investment in the internet every year since 2001.
Ad tech showed early promise on the internet, vastly improving upon older media and methods for direct advertising: that is, sales- and promotion-based advertising. Google was instrumental in the migration of billions of dollars, as their keyword-based advertising siphoned money from our yellow pages, classified ads and newspapers. One would be forgiven if one looked at the money migrating to digital in the first decade of this century and thought that the internet was unstoppable.
As we’ve seen, however, the bulk of the other $200 billion spent on advertising is a different type of animal: brand advertising (Just do it!). This type of advertising is less receptive to the internet’s charms. The appeal is emotional, and thus cannot be easily tracked. And the bulk of this money is spent by a relatively small number of large brands, whose audiences are effectively every human being on the planet, rendering the internet’s improved targeting ability of limited value. This nuance is lost on many internet economy pundits, and most predictions about the internet’s seemingly unstoppable appetite foresee the internet continuing on as before, cannibalizing more and more ad money from other mediums. The truth is not that simple.
The question before us today, then, is whether the internet and the ad tech of today is up to the task of capturing the other half of advertising money it can’t already claim. I believe that the prognosis is not promising.
The holy quest to solve Wanamaker’s dilemma
To be fair, ad tech has made incredible advances since Google. We’ve made significant strides in end-point attribution: the ability to associate a purchase with an ad, even if the consumer saw the ad days earlier. We’ve vastly improved ad tech’s location awareness. We’ve accumulated mountains of data — perhaps immorally — on every consumer. We’ve improved social tools, harnessing our friends and family as ad partners.
But we’re nowhere near the holy grail of advertising, which is twofold:
- Attribution: To definitively ascertain which ad convinced someone to buy a product.
- Allocation: To efficiently allocate our dollars most efficiently against that information, reducing or perhaps eliminating waste completely.
If we can nail these two things, we’d eventually solve Wanamaker’s dilemma.
When one thinks about these two concurrent goals deeply, a challenge for ad tech quickly makes itself evident. Even if ad tech succeeds at the attribution goal, there is no guarantee that the allocation goal will yield to the internet. Think about it. It’s possible — though, I believe, unlikely — that our ad tech data capabilities will eventually get to the point where we can perfectly tell which ad convinced someone to buy a product. But ask yourself: even if we get there, what is the likelihood that the ad will have been seen on the internet?
Virtually every study confirms what you probably already suspect: that television is more effective than the internet for emotionally inspiring someone to purchase something they weren’t already thinking about purchasing.
One out of two ain’t enough
There is, then, a scenario where ad tech dominates in the attribution goal, but not the allocation goal. Thus the media money — the real money — would still be spent elsewhere. Of course we can’t predict, now, exactly how ad tech will accomplish this seemingly impossible attribution goal — it’s a holy grail still far off — but if the attribution goal is accomplished, it won’t necessarily predicate the spending of ad money on the internet. It will involve a data spend. A SaaS spend. A fee, like other data providers charge. In this scenario, one or many companies — Google or Facebook, perhaps, or perhaps something with more location chops like Foursquare or Placed (recently acquired by Snapchat) that can tell when you walk into a store — will bill companies a fee for analytics, and once the brand possesses said data they will spend their money elsewhere.
To conquer the other $200 billion of advertising money, ad tech must not just be able to tell us where to spend our ad dollars most efficiently, but also, that the most efficient ad must also run on the internet. To conquer only one part of Wanamaker’s dilemma is to consign the internet to fee-based dollars, ceding the much larger pool of ad media dollars.
In search of something new
Now, you may be thinking to yourself “well, maybe that’s not the case. Maybe the data is the ad spend. Maybe there’s some cool new future ad tech ad unit that is required to capture the data. The ad unit is the attribution solution.” Well, yes, maybe. I mean, I can’t think of how, but ok. Let’s talk through that possibility. This would happen one of two ways:
- By fiat: one company or an oligopoly of companies develops the data technology and requires significant ad spend with them for a brand to access that data. This could well spur antitrust concerns, but even putting that aside, unless this was the perfect, be-all unit, it’s quite likely that brands would try the service, see that the ads that accompany the data are not the best performing ads, and determine that the waste of buying elsewhere is worth it.
- By holy perfection: someone actually does invent an ad unit that not only captures the best data on purchase inspiration, but also is the best available ad for purchase information. In this case, simple economics dictates that the provider of this unit would price it as high as possible to retain as much of the profit as possible while still keeping their unit as the best performing. Say, a 5–10 percent premium of efficiency over the number two performing ad medium. The end result for the brand, then, would be negligible, and many brands might resent the whole affair enough to not bother. Additionally, competing mediums such as television would adjust their rates downward to match the efficiency gap, and we’d have a deflationary price war.
What is the Internet and TV merge?
TV is migrating to the internet, of course. And with that migration, it will be more possible to apply ad tech to television. This “grand unification” could make it so that, if TV becomes the internet, that ad tech’s vision could be fulfilled: it could provide for attribution (through ad tech) and allocation (through TV-on-the-internet). In this case, however, if the attribution tells us that digital ads were the ones working great, then the internet didn’t need TV. If the attribution tells us that, say, billboards worked best, the internet didn’t need TV. And if the attribution tells us that TV shows worked the best, then those ads would be on TV-on-the-Internet but they’d still be “TV Ads” in format, so did the “internet” really win? We should also dig into why internet boosters claim all ad money is coming to the internet. They don’t claim this out of some rah-rah love of IP addresses and distributed networking. They claim it because it gives ad tech companies and platforms higher multiples in their valuations.
So in this scenario, if the internet and TV merged, the new, mythologically perfect attribution capability of the internet would tell us not just that internet-delivered TV ads worked best, but it would help us tell us where they work best: which shows performed best, and which content was priced right.
And if that’s the case, then all of the internet is once again media companies competing on the merits of their professionally-produced content, with a nifty data attribution capability that makes it even harder for the second tier to compete. So why is it this deserves an inflated multiple over HBO?
Slowing progress towards an impossible goal
It should be noted that these are the best case scenarios for ad tech. Far more likely is that when we have our ad tech data singularity (well, far more likely is that the internet never even nails the attribution goal, but that’s a topic for another time), we learn what most studies have told us and we already suspect: that television is probably the best medium for advertising we’ll ever have, and that shoppers aren’t generally convinced by one single ad, but rather the atmospheric, ambient presence of many ads seen or absorbed throughout a campaign. It’s also possible that the new, perfect ad tech data attribution capability will make television more efficient: helping advertisers spend their money on the more appropriately priced networks and shows and spending less on shows that don’t convert as well. But in that scenario, the money still stays on television.
For ad tech to succeed in conquering the other half of our ad dollars — the big brand dollars, still predominantly spent on television — it must perfectly define attribution and be the perfect place to allocate ad dollars. And while ad tech has made significant strides in attribution, I struggle to think of a single instance where modern ad tech has even incrementally improved allocation. The last interesting new ad units ad tech proffered to us were native ads and native ads: that is, in-stream ads on Facebook and Twitter (Fred Wilson’s original concept native ads) and the new glorified content marketing that now passes as native advertising. I seriously doubt that even ad tech’s biggest proponents claim that either of these units stand a chance at performing better than television.
Without a perfect singularity in both attribution and allocation, then, the chances for ad tech to conquer the $200 billion still in television seem slim, indeed.