Commentary

Search + Display + Attribution Modeling = Transparency

Josh-Shatkin-Margolis

Magnetic, which relies on search data to retarget display ads, has signed a partnership deal with C3 Metrics, a company that provides attribution modeling. The two plan to announce the deal within the next month, according to Magnetic CEO Josh Shatkin-Margolis.

It's the next step in Magnetic's attempt to build out optimization features on its display ad retargeting platform and self-service interface geared toward search engine marketers, Shatkin-Margolis says. The platform launched in December. About two dozen advertisers are testing the platform. "We're seeing, on average, about a 74% increase on click-through rates, compared with the advertiser's average, as well as a 325% return on investment," he says. "Overall, the big focus for us is to help advertisers improve ROI, which means improving the optimization part of our platform that goes on behind the scenes."

advertisement

advertisement

It's all about data and media optimization. Shatkin-Margolis believes search data provides the best source for retargeting display ads and improved optimization. The Magnetic platform analyzes ROI metrics such as CTR and cost per acquisition (CPA), and considers context vs. keywords, so when a consumer searches on "cell phone" the system automatically buys media on electronics-related Web sites. The optimization leads back to attribution modeling, which will allow advertisers to attribute the keyword or the search engine that led from one click to the next.

Search marketers looking to expand basic knowledge into retargeting display ad technology also might want to know that Magnetic recently expanded its data centers, pushing out a 100-node data processing cluster on the open-source platform Hadoop. I know it sounds a bit geeky, but you'll want to know because the technology lets Magnetic process optimization calculations, along with about 5 billion searches monthly. For each search the calculations examine several dimensions. That's where the data clusters come in handy. Shatkin-Margolis says it's the same type of technology Google and Yahoo built their back-end systems on.

Serving up an ad using search data and intent-based signals to a consumer down the road could end up costing the advertiser even less. So after putting the technology together correctly and finding the inventory across the Web that costs the least, a company can lower its CPA on display advertising. That's the premise, anyway.

For Magnetic, the focus has been on building out its optimization technology. That means launching additional match types that allow advertisers to control keywords, and targeting ads based off data sources and keywords. It's based on the "and-if" programming language. For instance, if a consumer searches on the keyword phrase "black Lexus" and "SC430 convertible" in the past week. While the process to improve the platform is never ending, Shatkin-Margolis and team appear to have made a positive start.

 

Next story loading loading..