Four years ago, the independent media agency Horizon Media took stock of what it was doing for clients and realized that there was a huge over-reliance on third-party data and third-party technology resources.
“We’ve been using a lot of DMPs and CDPs on behalf of our clients and noticed how you can use data and there are many gaps and limitations in this technology in terms of data granularity,” said Paula Conard, Horizon Media’s Chief Personalization Officer.
For example, he said, “You really couldn’t use digital features and viewers outside the channel.”
The source of the third-party data is also unclear, he said, and it is often difficult to know when the data was modeled or when it was last refreshed.
“Honestly, it’s kind of expensive and doesn’t really work,” Conard said.
Horizon followed his review with the launch of an indigenous data platform called Blue, which uses PII data licensed directly from Transunion as its identity level (or “backbone” in agency language). From that database, Blue. Determining consumer profile data is collected from over 283 million people in the United States, including names, phone numbers, physical addresses and emails.
Clients can modify these profiles with their own familiar consumer characteristics, such as employment status, population, home ownership and income, and looklike or trend modeling.
“DMPs and DSPs felt like black boxes,” Conard said. “We want to position them to help clients do more with their data and give them control over what they take home.”
Conrad spoke to AdExchange.
AdExchanger: What does a Chief Personalization Officer do? It’s not a title I’ve come across before.
PAULA CONNARD: There are many different ways to apply data resources, and media is only a part of what we do. It’s all about creating a personalized experience across the screen, whether it’s through advertising, on a website or even in an offline situation like an event. That’s what I do.
Clients are usually so silent in the media, CRM, email, creative and their website. But we have data technology that can combine those things to orchestrate a real customer journey.
Why did Horizon decide to build a data platform instead of a buy one?
As an independent company, we still have a P&L controlled by our CEO, Bill Koenigsberg, and I work directly with him on our investment plans. Because of this we are able to change, develop and build quickly. We have the flexibility that our competitors, many large holding companies do not have.
How blue. Separate from other agency data platforms, Of Which There A lot?
OK, we don’t use third party cookies. The change that we have created will come. We license our data directly. We say go straight to the farm. It comes to us and goes directly to AWS at the most grainy level We then prepare it for exploration and place it in the snowflake.
I have a team of marketing scientists and we create custom models for our clients. Once the data is in our UI, we can do cluster analysis, we can use advanced statistics techniques, we can build Boolean logic audience and the data is refreshed every day. If you use prepackaged third party data and templated methods, this is not something you can do.
What kind of custom models have you created for clients?
In recent times, many of our clients, such as Hispanic listeners, have become interested in understanding underdeveloped audiences. We scored our complete database using machine learning techniques to determine which IDs have a tendency to be Hispanic, and then we came up with a retail client loyalty data and looked at the file across our backbone to see what the lifetime value of those visitors actually was.
What kind of questions do you get from clients about dealing with signal loss?
The first thing they want to understand is the impact of third party cookie changes and how it will affect their data sources. They want to understand the impact of things like remarketing and attribution, and the changes they need to make to the marketing or technology stacks they use.
It’s a lot of learning because it’s not like everyone lives and breathes like this all day.
What do you think of the work being done in Chrome Privacy Sandbox and are you involved in any API testing or what are you planning to do?
We’re a little bit involved in this, but there seems to be a lot of talk and a lot of work. It’s like a car before a horse. We always have to check everything, but at the moment it is a lot to wait and see.
Like FLOC, there are always going to be a lot of starts and stops. This is important work, but it takes a lot of effort and time, so it seems slow and repetitive. This is another big thing to do, and to deal with it you have to create almost another team.
This interview has been edited and summarized.