Genome uses big data to continuously update target audiences
Editor's Note: "Genome Decoded" is our new Ad Blog series that tells the story of Genome from Yahoo!,our new data-driven audience-buying solution, from many angles, including interviews, videos, client stories, infographics and more.
"Know your audience" is a marketing commandment that should be chiseled in stone. But when marketers think about audiences, they should also remember these wise words---people change.
Every marketer likely has some core attributes in mind when defining their target audience — think age range, gender, income level. But those attributes, which may remain constant, don't necessarily tell the whole story. They miss unexpected audience attributes that can also drive conversions, but are more fluid---think trends or seasons---and require deeper and more detailed analysis to find.
"Audiences are dynamic," says Pedro Ponce de Leon, director, Solutions Engineering, Yahoo!. "To grow your campaign, extend your reach and get better performance, you need to look for data that changes over time and discover the not-so-obvious audience data points that will perform for you."
Unexpected opportunities
Here's a simple example: Say you're setting up an online campaign to drive subscriptions for a men's magazine. In the past, most of your conversions came from college-educated males 18-34, so you buy an audience that matches that profile. Unfortunately, that profile doesn't show that during the holiday season, mobs of college-educated women 18-34 buy subscriptions as gifts, so they're left off your list.
The moral is that marketers should dig through more sources of data to find unexpected opportunities and beef up their targeting---and do it continuously. As Ponce de Leon says, "It's important to constantly optimize a target audience because the data facts that define it are ever-changing."
Big data, better audiences
Defining an optimum target audience by continuously analyzing huge amounts of data, quickly and comprehensively---that's what Genome from Yahoo! was built for. It starts by ingesting the biggest plate of big data in the industry, piled high from:
- A constant flow of new information from our partnerships with 25 leading data providers
- Yahoo!'s mountain of proprietary data, including registration, search and behavioral data. (Yahoo! has user data from 76% of the U.S. population and over 300 million daily searches around the world.)
- Data from our advertisers, including conversion, site-visit information and even entire house files
"We have the most data relationships and the most data," says Ponce de Leon, "and the more data you have, the more confidence you can have in your audience models."
Always analyzing
The Genome team leverages its massive data base---and the number-crunching prowess of OSM, a proprietary technology stack that manages diverse data sources---to provide the kind of continuous optimization that produces better audiences.
Ponce de Leon offers a basic example. Say an advertiser's campaign goal is to drive more purchases as efficiently as possible. First, the Genome team will tag every purchase with a pixel that identifies each buyer. Then it will leverage Genome's comprehensive data base to analyze that new purchase data and paint updated, detailed pictures of the latest set of converted buyers, which marketers can use to keep their campaigns current. "With Genome campaigns, we perform data analysis and optimization throughout the lifecycle, so advertisers can be sure that the right data is being used," he says.
All this, and save money too
Finally, Genome's big-data brainpower can help advertisers make wiser decisions when buying media and audience segments. Its analytics reports can show the amount of audience lift associated with each data fact of a target audience, so advertisers can weigh whether purchasing a particular data segment will bring enough positive impact to be worth the price.
Now that's big data working for you. For more information on Genome, contact your account representative or call 866-803-7994.
