Thursday, April 16, 2015

3:09 AM

Download the pdf

Online Ads Drive Sales
Garrett Johnson

New research from Assistant Professor Garrett Johnson
and his coauthors shows the cause and effect of online
advertising and retail sales.
One of the largest and most statistically powerful studies of its kind provides proof that online ads improve retail sales.

A large-scale study by Simon assistant professor Garrett Johnson and two coauthors provides for the first time strong evidence that online retailer ads increase Internet and brick-and-mortar sales.

Especially effective is repeated ad exposure among the 80 percent of users who see up to 50 ads per person over two weeks, they found.

Johnson conducted the study as an intern at Yahoo! Research while earning his PhD at Northwestern University. He and coauthors Randall Lewis and David Reiley ran a controlled experiment examining two consecutive weeklong ad campaigns on Yahoo! that targeted three million customers of a nationwide retailer. Yahoo! Research and the retailer partnered to explore how online display advertising affects both online and in-store purchases. The coauthors share their results in a paper, “Location, Location, Location: Repetition and Proximity Increase Advertising Effectiveness,” now under journal review.

Advertisers have long questioned how much advertising a firm should purchase before diminishing returns set in. Equally puzzling is whether ads should target a company’s most loyal customers or woo those on the margins.

Effective advertising frequency is higher than many might expect, Johnson says. The returns to advertising hold strong even up to 50 ad exposures in two weeks. The coauthors also found ads had the biggest effect on the retailer’s most active customers and those who live near a store: in those groups, online and in-store purchases went up the most.

In the study, a full treatment group saw the retailer’s ads, and a control group was exposed to unrelated control ads. The study showed that retailer ads increase sales 3.6 percent in the full group relative to the control group—a statistically significant rise, Johnson notes.

The experiment is among the largest and most statistically powerful in the ad effectiveness literature. And because the research combined consumer-level ad exposure with valuable retail purchase data, it established important and unprecedented causal effects.

“Typically, it’s hard to get a sense of who sees the ads,” Johnson says, “and even harder to connect these people to their sales.”

As a result of the research, in 2011 Yahoo! launched Proximity Match, which helps marketers find consumers who live near their brick-and-mortar locations. The product has since brought in significant revenues for the company.

—Sally Parker


Post a Comment