Some advertising industry insiders believe the data-buying model is broken. Most is bought on a CPM basis and then tied to the CPM of the media.
Ryan Rolf, who spearheads data sales at Lotame, is calling on the industry to change its pricing model.
"It would be easier to transact the purchase of data through media platforms" he said. "If every demand-side platform could transact, report and track, and let their advertising clients purchase the media with the data under different models, it would help everyone. Brands going in house should also have the option to transact the data directly with the provider."
Initially the industry developed a CPM buying model because the DSPs, and the media buyers at agencies were used to buying on CPMs. Before that Blue Kai, which Oracle now owns, began selling data through what Rolf called a “cost per stamp” model for data.
D&PI: What are some of the other data buying models advertisers might use?
ROLF: A performance data model. For example, Aspen Resorts may want to buy media and find those who would book a room. You can buy a lot of inventory such as display for $0.50 or $0.25. If we assume the cost of the media is $0.25, and add it to the cost of the data, it would double the performance to serve one impression to one user.
In some cases advertisers say this is just too much unless they get double the performance, because they could have bought two impressions and reached two different consumers.
D&PI: It sounds like you want to put data buying on a scale.
Rolf: Absolutely, The Trade Desk has done very well with their percentage of media cost model. They let advertisers choose to pay for the data on CPM or as a percent of the cost for the media. If the data is successful in driving the impressions, rather than tacking on a fixed CPM, they give a percent to the data partner.
[There's] more frequency if an advertiser needs the data for creative optimization or analytics in their own data management platform before deciding who to target. Some advertisers are comfortable with a flat-fee model.
D&PI: What are the data buying options?
Rolf: Marketers need options when buying data. There should be several pricing models, such as CPM, percent of media, or flat fee. If the advertiser has all three options, they could test all three by taking a third of the budget to try CPM, another third bought as a percent of media, and the last third as a flat fee. I think you would see a lot more value.
D&PI: Are these models being used today?
Rolf: Yes, but they require direct relationships with the data provider. Then friction occurs when the data provider and the advertiser show up to the media platform with their own data deal.
We’ve always been open to test new models, but since we don’t run the media we fall short. It’s not on us [Lotame] to support the data pricing model and integrate it into the media pricing model. For smooth adoption you really need more platforms to adopt these models.
Today, around media buying, the biggest questions have become the quality of the data. Advertisers now see how much they spend on data, especially if it’s a line item on their bill, and they are starting to question whether the data works and why they use it.
The data must be priced correctly to perform. I hear more times than not, "Oh, the data just doesn’t perform." The data could be great, but it’s used or bought the wrong way.
D&PI: Are you speaking to advertisers to let them know they have pricing options?
Rolf: Yes, we ask if they’re open to other data models or open to using the data in a different way such as to find trends and links.
D&PI: How do you make sure advertisers uses the correct data pricing model?
Rolf: It would be great if the demand-side platforms drove the change, because the DSPs see the full view of the client’s campaigns. Data companies also need to take the initiative and act as consultants and strategists.
Price is one piece of a larger trend set by advertisers. They are talking about ethics, transparency, and quality. Price is just the first step. If I’m not buying something correctly, I’m probably not using it correctly. The pricing model needs to be fixed before we can move on to tackle other challenges.