One of the biggest obstacles to increasing advertiser investment into online display advertising is that there still is little definitive proof that online branding campaigns have a positive impact on offline sales.  Everyone agrees that online display campaigns can drive sales online and all of the major digital ad campaign management companies have introduced analytical tools, such as Eyeblaster’s Channel Connect for Search, that enable advertisers to measure the impact of the branding elements of their online campaigns on online sales.  The problem is that most consumer purchases still occur offline.  This means that, if online display advertising is ever going to be taken seriously as a branding medium, there needs to be definitive proof that display ads can drive offline sales.

The good news is there are a number of strategies for measuring the impact of online branding campaigns on offline sales.  None of the methods are easy or cheap.  However this research is necessary to increase advertisers’ confidence in online display advertising.  Ultimately, the industry will reach a tipping point after which the value of online advertising for branding will be taken as a given just as is the case for TV.  Until then, the following are the best ways we can prove the value of online display advertising.

Database matching

Description: This technique involves comparing online impressions to offline purchases using a representative panel consisting of consumers whose online behavior and offline purchases are both tracked.

Pros: This is the most rigorous methodology since all the data is electronically recorded.

Cons: It’s expensive.

Vendors: comScore, Dynamic Logic, Nielsen, Platform-A, Yahoo!

Longitudinal surveys

Description: This method involves conducting surveys over time among a representative panel to gauge their purchasing patterns as well as the online ads they’ve seen to determine whether there is an association between the two.

Pros: This is the least expensive methodology.

Cons: It’s less rigorous since it depends on the panelists’ abilities to accurately recall which brands they’ve purchased and, more importantly, which ads they’ve seen.

Vendors: Dynamic Logic, InsightExpress

Geographic testing

Description: This strategy separates different DMAs into control and test groups which are exposed or not exposed to online ads accordingly.  Offline sales in the different DMAs are then compared to determine the impact of exposure or non-exposure to the online ads.

Pros: This is a fairly rigorous methodology.

Cons: It sacrifices reach for the campaign in question, and sales if the ads prove effective,  since certain DMAs are not exposed to online advertising.

Vendors: This is usually done in-house by the advertiser.

Marketing mix modeling

Description: This scheme uses sophisticated statistical analyses such as multivariate regressions on ad serving and offline sales data to estimate the impact of online advertising on offline sales.

Pros: This is a proven method for measuring the impact of TV ads.

Cons: It tends to focus heavily on short-term sales impact and may be less suited for online.

Vendors: MMA, Marketing Evolution, MarketShare Partners

If you’re aware of any advertisers who have completed studies like these, it’s a good idea to share the results and/or the insights.  comScore has done a great job of this recently, publicizing results of some of its studies and even publishing some of its latest research in the last issue of the Journal of Advertising Research*.  Pooling our efforts as an industry is necessary if we’re going to establish once and for all the value of digital as a marketing medium. 

*Fulgoni, G.M., & Mörn, M. P. (2009).  “Whither the Click? How Online Advertising Works.”  Journal of Advertising Research, 49(2), 134-142.

Sean Gelles, Product Planning Manager