Last week, demand-side platform, [x+1], announced Open Data Bridge which it says gives advertisers the ability to “simply integrate and use client and third party data for online targeting.” Read the release.
Ken Rona, VP of Data Strategy and Analytics at [x+1] discussed the new product as well as provided insights on the uses of data today.
AdExchanger.com: What are some of the key differentiators of Open Data Bridge in comparison to your demand-side platform competition? For example, many DSPs offer access to TARGUSinfo, Datalogix, eXelate, BlueKai, etc.
KR:
- A major component of the ODB is that it allows you to quickly and simply make any third party data (not just those that we are integrated with) or custom data source (that can be matched online) available for targeting. Clients and agencies can bring their own data relationships. We have developed technologies to make the loading and use of the data into our platform very simple. For many marketers, a valuable differentiator is their own customer data, and now, the value of that data can be unlocked by ODB.
- Unique to the platform is multivariate targeting. The platform allows you to mix, match and combine data sources.
- The platform also does a series of quality control steps to ensure that the data is being loaded correctly and has the proper amount of consistency. Different data elements have different levels of consistency. For example, we don’t expect that most people are consistently in the market for a vacation, but we would expect that most people’s income is fairly stable.
- All of the data can be made available to the POE engine for optimization, which automatically finds the relevant data from all the sources and builds custom models for each client and customer.
- The data can be used across our site optimization and media optimization solutions. This is a unique capability that allows us to optimize up and down the purchase funnel. So, we can create different customer experiences based on the data we used to do media optimization. This allows the marketer a “single view” of the customer.
Discuss [x+1]’s experience with data providers as it relates to the following: are you finding that data has a shelf life? Can you share any experience here? Perhaps shelf life for Travel shoppers, etc.?
KR: As you could guess, shelf life is less about the provider and more on the customer as represented by data. We just used in-market for home electronics to good effect, but we did not go back and show ads to people who converted to see if they would convert again (grin). So, in-market for some categories has a natural shelf life.based on if the product category is durable goods or if it is a considered purchase, etc. Much of the 3rd party data that we have brought into the first phase Bridge is based on demographics and lifestyle segments. Those are fairly stable.
Shelf life is less of a problem than the “level” of the data; The problem with some online data is that it is not keyed to users. It is keyed to a computer or a HH, or even worse a zip code. Even targeting on a Zip+4 can provide meaningless data, depending on the variable. For example, it does not make sense to target gender on the basis of a zip4. The ratio of men to women does not change much at the z4 level; It changes at the user level. So, when talking to the data providers, you need to ask very explicit questions about what the level the data is on and understand that the campaign results you are getting might be due to the fact that you are targeting based on the household, not the user.
Where is the opportunity with social targeting today? Is it all about look alikes?
KR: It’s an important challenge because marketers are looking for social to generate real success this year. The quick answer is: Not all. For me, the bigger opportunity is in audience expansion. Let me back up. There are two types of data that could be thought of as “social targeting.” The first is for audience expansion: like Media6 or 33Accross offer their media clients. We use a xGraph to place a tag on a seed user and then they provide me a list of user IDs who are friends with that seed user. For campaigns where there is a natural community, this approach works great. For example, if you want to reach stamp collectors or people that work at a specific company, using social media for audience expansion makes sense. These people have a reason to be connected that supports your campaign. By contrast, I have a hard time thinking about using social media to expand an audience for vacuum cleaners. People buy vacuums when they need one, not generally because their friends just bought one or are really into vacuums. Now that I think of it, I can think of a way to expand the audience for vacuum cleaners, but it would be a small audience. I’ll post my idea in the comments section in a couple of days to let your readers make some suggestions.
The second use of social targeting data is what Lotame and Rapleaf offer. They extract users’ interests (and some personal characteristics) and tie that data back to a user ID. This is targeting derived from social media. It’s also a valid way to target, but at this point, the audiences are a little thin. This is getting better all the time, though and I expect that in the future this will be very powerful in finding extremely targeted audiences. We are actively using both types of data.
By John Ebbert
