• by Jack Loechner , Staff Writer @mp_research, May 5, 2015

According to a new eMarketer report, “Cross-Device Search Marketing: As Search Goes Multidevice, Ad Targeting and Measurement Struggle to Keep Pace.”  U.S. internet users are conducting more and more searches that start and finish on different devices. This multidevice behavior makes paid search targeting and accurate performance measurement more difficult, but there are some methods and best practices that marketers can adopt to improve their results.

Analyzing cross-device search user behavior is difficult. For many marketers, improved methods for targeting multidevice search users will be key. This will involve using secondary search signals such as the type of device, along with user location and time of day, to further inform bidding strategies.

Paid search marketing has long been a discipline focused on maximizing the value of keywords. But the evolution of search toward cross-device leads many to conclude that keywords may no longer be enough.

John Cosley, director of product marketing for the search advertising business group at Microsoft, says “… if you are a marketer… building a marketing plan against an audience… you have a specific person you’re trying to reach… not a keyword… ”

And Jeremy Hull, director of bought media at iProspect, says “… focusing just on keywords is like reading an email … (without) contextual clues… just the text… looking at (the) information that the user passes when they search… is equivalent of being able to have a face-to-face conversation… where you can read their body language…”

To help solve this problem, marketers are increasingly using secondary data from CRM programs along with search cues like time of day, user location and device used to further target strategies for search. A number of sources suggest marketers typically use this information to help decide on bid adjustments.

For instance, a search user in a specific ZIP code might be worth more to an apparel company even if he or she doesn’t search for jackets.

Brian Lee, market research analyst at Marin Software, notes that “… it’s more context-based… (for instance) if the prospect lives in an area that gets a lot of rain… it’s related to their location and the audience… more than it’s related to their current search topic….”

A February 2015 survey by Duke University’s Fuqua School of Business, investigating U.S. marketers’ use of customer behavior data for targeting purposes, implies that the practice is increasing. Although slightly less than half of U.S. marketers in the survey currently used data for targeting, more than 90% said that their use of such data was increasing, concludes the report.

For additional information from eMarketer, please visit here.