• by David Danziger , October 31, 2016

What do you picture when you hear the term “walled garden”? The notion conjures up imagery that is soothing and pleasant. It paints the picture of a nice, flowery place inside protective barriers, safe from weeds, rodents or other threatening outside forces.

In the digital landscape, this garden isn’t as pleasant as its name may indicate.

As a refresher, a walled garden in advertising is a closed ecosystem where the company running the ecosystem tightly controls access (or doesn’t allow access) to outside applications, content, targeting or measurement. Not quite the picture you had in mind, right?

In the finance world, there’s a similar, closed structure that’s referred to -- fittingly -- as a dark pool: You can’t see what lurks below the surface, and you jump in at your peril, because it’s anything but transparent.

Walled gardens and dark pools serve certain limited purposes. In instances of a low-liquidity marketplace, they can give buyers access to sellers, and sellers access to buyers without whipsawing price swings. Walled gardens and dark pools can provide confidence, theoretically anyway, that the administrator of the garden/pool has the wherewithal to match buyers and sellers and to be fair to each. However, the advertising marketplace -- and particularly, the programmatic landscape that has evolved -- doesn’t have a liquidity problem. It’s a robust ecosystem with buyers, sellers, and independently accredited third parties for measurement.

When Facebook acknowledged last month that it had inaccurately measured viewability, meaning that its viewing numbers may have been off by as much as 60%-80%, it brought all the dark-pool murkiness of a closed system to the surface. If we put the best face on it, it was an unfortunate and unintended error. It eventually was discovered, advertisers were notified and now it’s fixed. Still, one can’t help but wonder how much quicker would the error have been discovered if clients of Facebook were allowed to use their own preferred measurement provider to validate the numbers that were being reported by Facebook? In an open ecosystem, that happens all the time.

This is the deeper, inherent issue of a closed ecosystem: It always raises the question of motives. If the scenario described above, Facebook’s accidental measurement error in this case, is the best face, then it doesn’t take much creativity to imagine the sinister scenario.

When a company makes its money by selling ads and that company also only allows its own in-house measurement validation, it drives up prices and the perceived value of that inventory. This creates some really problematic incentives. The same “incentives problem” occurs for companies claiming to represent the interests of buy-side and sell-side clients in the programmatic landscape. It’s pretty hard to do both effectively and simultaneously.

Luckily, the solution is simple, at least in theory: transparency and clarity. Whether it’s called a walled garden or a dark pool, the net effect is the same. There’s still a lack of visibility of what lies within the walls or beneath the surface. An open ecosystem offers a much clearer view of how dollars are spent, what’s working and what’s not, and what a company’s incentives are. The lessons to be taken from this are the same thing we learned as children: If the pool water isn’t clear, don’t dive in.