The problem with ad tech | Adweek The problem with ad tech | Adweek
Advertisement

We Reach Everyone Online. No, We Do!

The stream of meaningless ad statistics keeps piling up, cheapening the medium

Yesterday, AppNexus announced that it had pulled in $75 million in new funding. Pretty impressive, particularly in a climate where you hear things about startups like Tumblr having trouble raising money.

Even more staggering of a stat was this line from AppNexus’ press release: Over the past year the company has “processed a peak of one million requests per second, and quintupled delivery to over 13 billion ads served to websites each day.”

I don’t know about you, but that sure sounds like a lot. According to CEO Brian O’Kelley, that’s only the number of impressions AppNexus serves. It "sees" 40 million. On Twitter yesterday, comScore’s svp corporate development Kirby Winfield guesstimated that AppNexus is involved in about 10 percent of all banners served every day online. I would have thought it was even more. Man, the Internet is big!

Still, AppNexus must be well on its way to building out the ad tech platform that will rival Google’s, right? They’ve got the exchange product. O’Kelley’s hinting at some sort of supply side platform, some other stuff they do. With numbers like those, AppNexus must be unrivaled, right?

Wait, what about Turn? According to the demand side platform’s press materials, every day the “Turn platform makes over 30 billion advertising decisions, analyzes over 1.5 trillion customer attributes.” So Turn must be handling the back end ad delivery or decisioning or serving or targeting or something for most of the Web then, right?

Hold on! According to comScore, in December the demand side platform The Rubicon Project reached 214.2 million monthly unique users, or 96.7 percent of the U.S. Internet audience. Not only is that more than Google (for the sixth straight month!) 96.7 percent is almost 100 percent, i.e. the whole Internet. AppNexus isn’t even close! It seems to me that you must be able to buy and target everyone online through Rubicon (do they sell ads that way? Or just facilitate? Or provide the platform for? Or something else?).

Stop the presses!

Rubicon rival Pubmatic handles 50 to 75 billion gross impressions served and 500 billion to 1 trillion bids monthly, according to its website. I think trillion might even trump quintuple in its impressiveness.

The list of big-talking tech firms goes on: “Through the eXelate marketplace, we are able to capture rich targeting data on over 350 million global unique users via a rigorous set of data collection, filtering, segmentation and normalization procedures.” 350 million people is a ton. They must be spooking Google as well. Surely BlueKai is, as its “Audience x Media provides marketers with scale across 300 million unique users.” Hey, even Yahoo's Right Media's RMX exchange is said to handle billions of impressions a day and reach 700 million uniques. Marissa, you don't need to buy anything!

Yes, I’m mixing DSPs, DMPs, SSPs, exchanges and all sorts of other ad tech company flavors. But you see my point. Big numbers are always going to be important to advertising. Brands love to be in high-rated shows, they want campaigns that deliver national reach, reach 120 percent of the possible in-market car buyers at any given moment.

But at a certain point, these eye-popping data boasts become more akin to rap brags. Saying “I’ve got more rhymes than the Bible’s got Psalms” sounds impressive, until you read on Wikipedia that there are only 150 Psalms. And that’s not even close to the whole Bible.

Of course, there are plenty of companies in this space doing all sorts of amazing things when it comes to using algorithms for sophisticated ad targeting. AppNexus is certainly one of them. But after a while these numbers do two things. They constantly serve to remind people that there are way too many online ads being served, which is why far too few of them to be worth much of anything. And it serves to make one's eyes glaze over—since the reach, impression volume and targeting numbers cease to mean anything. In effect, they get cheapened.

Ad tech companies must figure out a better way of communicating what it is they really do and why it's so groundbreaking and world-changing other than "we handle a lot of impressions really really fast." They have enough image problems. They don't want to be lumped together with House of Pain.

Advertisement

Advertisement