It didn't pass without note this week that media agencies belonging to WPP's GroupM family started adding a curious clause to their digital insertion orders. According to MediaPost, the clause, which appears as an addendum to the IAB/AAAA Standard Terms and Conditions, reads as follows:
"Notwithstanding the foregoing or any other provision herein to the contrary, it is expressly agreed that all data generated or collected by Media Company in performing under this Agreement shall be deemed 'Confidential Information' of Agency/Advertiser."
Not surprisingly, this is causing something of a stir.
Ad networks, technology companies, and publishers believe that valuable marketing information is being gleaned from ad buys. And it is. They also believe that there's a value exchange taking place for which they ought to be compensated. And they're right about the value exchange part.
It's the expectation of compensation that I'm a little fuzzy on. But we'll get back to that in a minute. Let's talk about the GroupM clause for a second.
The fact that this aberration even exists shows me that GroupM’s lawyers, if not the people who run their media business, have a rather skewed sense of what constitutes "data."
By way of analogy, let's say GroupM was a patron in a crowded bar and a website publisher was the owner and barkeep. GroupM strolls in and offers to pay the barkeep to tell him what drink a customer just ordered. The barkeep tells him that Bubbles down at the end of the bar just ordered a Cosmopolitan. GroupM thanks the barkeep, pays him, and then asks him to keep the fact that Bubbles ordered a Cosmo confidential. Can the barkeep be reasonably expected to keep that fact confidential?
Of course not.
There are other people in the bar who might have seen Bubbles order her drink. There may be other people around who heard GroupM inquire about who was ordering which drinks and overheard the subsequent conversation. Not to mention that the other bartenders need to know that Bubbles ordered a Cosmo because they'll need to know what to give her when she orders a refill.
Web pages and the data gleaned from them behave in the same way. When an ad is served into a page, there are an enormous number of parties who are partially or wholly privy to that transaction. They include ad servers on both the publisher and agency side, analytics companies, outside software vendors, content syndicators, and a plethora of other parties.
Any good confidentiality agreement has a few carve-outs in the definition of "confidential information." One common one ensures that information generally known to the public can't be considered confidential. That carve-out is usually there to prevent one party from using confidentiality as a weapon by binding the other party from using a piece of information that everybody else not bound by the agreement can use. The IAB/AAAA Terms and Conditions contain this carve-out, as it should.
I'm not a lawyer, and no one should construe anything in this column as legal advice, but it seems to me that with the large number of parties collecting data whenever a particular web page is served, such information might meet the definition of "generally known to the public." Certainly, the expectation that it won't be leaked is ludicrous under the arrangements we currently have in web advertising, where multiple parties are setting cookies on a user or otherwise gathering information with every data transaction. The GroupM clause doesn't address this in any way that I can see. So I'm left with the notion that this might just be an unrealistic expectation on GroupM's part.
But let's get back to the big picture…
While it's true that publishers can discern certain facts about web users based on things like surfing behavior, I'm not certain that it's the publisher's exclusive realm to own those facts. This may come as a shock to some publishers, who have become accustomed to the notion that data collected on their site is inherently theirs.
Look at it this way:
If I invite my buddy Gus to a backyard barbecue and he shows up wearing a loud Hawaiian shirt, I know about it and I can tell other people that Gus wore a colorful shirt to the party. Despite the fact that I'm the host, though, I have no exclusive ownership of the fact that Gus wore a loud shirt. Other people may have seen him at the party, or on his way to it. Others might know about Gus' affinity for loud Hawaiian shirts and assume he wore one to my party. There are a lot of ways to get to that information, so the notion that it belongs exclusively to me is suspect.
In this way, publishers and ad networks may be selling data that they're perfectly in the right to sell, but that they don't exclusively own in the traditional sense. This makes the idea that you can treat a digital audience data transaction confidentially a little silly.
Regardless, I think we're in for a big battle over who owns observed audience data. Even though it will be tough, by definition, to relegate audience data exclusively to an agency, a publisher, an ad network, or whoever -- I don't think that will stop agencies from trying.
Tom Hespos is the president of Underscore Marketing and blogs at Hespos.com.
