I look at hundreds, if not thousands, of online ads each day, but the chance that any one of them will seduce me into making a purchase -- let alone clicking on an ad -- is pretty to close to zero. And I have to think I'm not alone in this.
But new research suggests that the mere presence of online ads has an influence on the products and services we buy, and that businesses that advertise online are probably getting a better bang for their advertising buck on the Web than in other media, like television, radio or print.
There have been plenty of attempts made at computing how advertising influences sales, but actually doing this is no easy task. That's largely because the relationship between ads and revenues is a two-way street. For example, many companies set their marketing budgets as a percentage of sales. That means when a company expects a good quarter or year, it's more likely to budget more ad dollars. So in the real world it's very difficult to tell if sales are being driven by marketing or the other way around.
To get around this problem, two researchers at Yahoo, Randall Lewis and David Reiley, set up a clever controlled experiment. Teaming up with a national retailer, Lewis and Reiley identified 1.6 million registered Yahoo users who were also in the retailer's customer database. (The study did not name the retailer. Large companies that participate in studies like this typically choose to remain anonymous for competitive reasons.) They split these users into two groups. The first, comprising some 924,000 users, was able to view three ad campaigns from the retailer while surfing on Yahoo sites between the fall of 2007 and the following winter. The second group was restricted from viewing any of the retailer's ads on Yahoo sites.
Surprisingly, Lewis and Reiley found that users who had seen the retailer's ads (which were not shown more frequently than other ads) wound up spending more than 10 times the ad campaigns' cost to the retailer. That sort of return on investment is way beyond estimates for other media.
Lewis and Reiley also unearthed a few other insights into how ads can influence users:
* Online ads don't necessarily generate online sales. The vast majority, 93 percent, of sales associated with the three campaigns took place offline rather than on the Web.
* While the marketing industry is very concerned with how many people click on ads, it turns out that 78 percent of generated sales came from users who did not click. (But, as would make sense, people who click on ads are much more likely to buy online.)
* Users who saw ads were likelier to spend more than they usually would at the retailer.
* The effects of the ads persisted for up to four weeks after the campaigns were over.
While it's always tricky to generalize, these results would seem to provide some optimism for the troubled news industry, which has seen much of its fortunes evaporate in the Internet era.
Lewis and Reiley's findings suggest that if evidence of the effectiveness of online ads grows, news sites should expect to see the return of some of the ad dollars that have disappeared in recent years.
