Blog post

Why Ad Tech Is More Complicated Than Wall Street

By Martin Kihn | December 04, 2014 | 0 Comments


Welcome back, marketers! Last time we spoke, I was dropping ominous science about the shadow world of high-frequency trading on Wall Street – a liminal space populated by Russian geniuses and assorted shoegazers who chase minute information advantages and total obscurity to do nothing more than bully regulatory and tech loopholes into shifting billions in decidedly grayish profits from ratatat buy-sell orders on a disconnected heap of exchanges and confusion in a massive casino game concatenated to drive stock prices up . . . and it thrives on the ignorant indifference of the S.E.C., and you and me.

But that’s Wall Street, according to Michael Lewis’ Flash Boys. Surely, here in the world of advertising – you know, sight-sound-and-motion, puppies and yuppies, glorious sunsets through the high-tinted luxury windshield advertising – things are not quite so bad.

Sure, there’s real-time bidding for impressions that looks a lot like real-time bidding for stocks. There’s anonymous people (and a frightening number of bots) doing Wall Street-esque stuff like buying and selling within fractions of a second at a profit (arbitrage) at scale. But – hey – it’s advertising. It’s can’t be that serious, right?


Well, before I go and tell you that’s wrong, let’s follow a Bloomberg News reporter into a lair in NYC’s Chinatown. In a recent story, the reporter found a shifty-seeming anonymous ten-person boiler room-style shop that claimed to arbitrage 500 million impressions a day on digital ad exchanges. Profit: “tens” of thousands each day (underestimated for tax purposes, no doubt).

A kingpin in this shop implied that he was buying and selling impressions on ad exchanges that were (temporarily) mispriced because . . . well, because ad agencies and other media buyers, including hapless advertisers themselves, were “too slow.” This is exactly the argument high-frequency traders make on Wall Street – except the slow pokes there are the big mutual funds and hapless folks at home with no better access to the markets than e*Trade.

Of course, ad exchanges and insiders have no incentive to let media agencies or in-house ad buyers using demand-side platforms have a level playing field, or even to know what “level” looks like. What’s intriguing is that in the story leaders at ad exchanges were quoted downplaying any so-called information imbalances. And one leading exchange’s president no doubt regrets admitting that his platform “doesn’t block the traders [arbitraging inside the same exchange] because it’s impossible to control their activity on other exchanges.”

Which is a bit like saying, “I’m not going to expel this shoplifter because he will just go shoplift somewhere else.”

No, my friends, I’m afraid there is plenty of reason to think that ad tech – and particularly the world of real-time bidding, which comprises over 25% of all display impressions sold in the U.S. – that it’s just as shifty, conflicted, bewildered and disingenuously crooked as Wall Street. In fact, it’s probably worse.

How can I say this?

Humbly, I offer up these . . .



1. It’s more complicated

Digital marketing technology today consists of at least 2,000 significant vendors – and those are just the ones we track in our secret database. Many of them touch in some way on the subcategory we can call advertising technology, including demand- and supply-side platforms, ad exchanges and networks, ad servers, media agencies and trading desks, data and analytics vendors, and so on.

But ad tech is not just commercially more convoluted than Wall Street. It is informationally more complex. Michael Driscoll, the brilliant CEO of Metamarkets and a former computational biologist, puts it this way: The world of ad tech processes about 400 billion transaction-like events per day, while the New York Stock Exchange processes a puny six billion trades; and programmatic ad trades use up to 100 data fields each, while a stock trade only has ten (you know, bid price, ticker symbol, etc.). So there.

2. It’s getting bigger much faster

There’s a perception that Wall Street is trading real money while advertising is in the kids’ room using Monopoly cash. That it’s a whole different scale of dinero. Certainly – for now – Wall Street exchanges see more money and brokers peel off more and so on than players in digital advertising. Wall Street is bigger than ad tech. It also had a 97-year head start.

But it’s worth pausing a moment to ponder: Just how much bigger is Wall Street?

As we’ve said, the part of ad tech most analogous to those high-frequency traders we were jealously punking up top is, obviously, real-time bidding (known as RTB). Right now, this portion of the digital ad universe represents some $10 billion per year. How much of this booty goes to the middle-people, the traders and exchanges who sit between the advertiser (the buyer) and the publisher (the seller)? It varies and is not really clear. The “ad tech tax” is probably at least 40%, but let’s be very careful here and assume a mere 20% margin to the traders themselves.

So we’re at $10 billion x 20% = $2 billion profits to the traders for RTB advertising.

Okay. A recent estimate of the total commissions paid to Wall Street banks last year for stock market trades was about $10 billion. So RTB trading is only one-fifth the size of stock trading? Maybe. But the ultimate prize is much, much bigger. Total global media spending is approximately $450 billion. If any significant portion of that ends up on real-time bid exchanges – as seems all but inevitable – then you’re not doing too much fancy math to project that in a very few years ad tech brokers will be making more funny money than Wall Street brokers.

Yes, I said more. That’s the prize here, my friends. That’s why all the data scientists and enterprise data warehouse barons and salesforce automation women and electric gangsters are rushing – rushing – into the arms of advertising. It’s not because they thrive on understanding customers.

3. It’s less organized

Ad tech is harder to define than Wall Street. What is the product? What is the value? Many years of academic work has gone into constructing a theory of stock valuation. Impressions are also valued by the market, somehow, but ultimately advertisers try to bid for them based on those 100 data fields we mentioned above, and on their own estimates of what those fields say about the value of the person to their business . . . plus some estimate of how likely that impression is to make an impact on that person.

That’s a lot of guesstimation. There’s also the sad truth that actual prices paid are obscured by layers of middle-people – for example, your ad agency’s trading desk may not tell you what they paid for those impressions, but they’ll certainly tell you what they’re charging you for them. Often, you as an advertiser won’t even know where they ran.

And for all its faults, Wall Street has a fairly clear set of hard-nosed regulations, codes of conduct, conflict of interest policies, a federal regulatory overseer, and the public’s full-throated skepticism to keep it from drifting too far. Ad tech doesn’t have much at all. Regulations and standards are sketchy and poorly enforced. Whole formats (like video ads) and key issues (like viewability or brand safety) barely have standards at all. And as yet, there’s no real “currency” for online advertising akin to television’s GRP.

4. It’s (even) more criminal

Capitalists in any industry are like publicists to the stars: they deny their project has a raging problem until it’s so obvious they have to break down and admit, yes, there may be an issue with, um, dehydration. This year, the problem of fraud in display advertising was apparently to obvious even insiders had to admit there might be something – something minor, to be sure, and easily solved – but something going on.

Most recently, one beleaguered former ad tech high-flier, whose stock fell on widely-reported questions about impression quality, has come out strongly claiming its fraud level is very, very low. Others followed suit. It all reminds me of a comment made by good-hacker Michael Tiffany of White Ops, an expert in display ad impression fraud. Tiffany said comparisons to credit card fraud miss the obvious point – namely, that credit card fraud affects only about 0.07% of transactions. Meanwhile, fraudulent impressions on ad networks and exchanges could be as high as 30%.

In other words, here again, ad tech is worse than high finance.

The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.

Comments are closed