By John Nardone
Modeling the return-on-investment of internet marketing doesn't have to be a challenge. After having spent nearly 10 years at Modem Media developing Internet marketing programs
for Fortune 500 clients, I'm often approached with questions about marketing mix modeling's ability to measure return-on-marketing investment (romi) for Internet advertising.
There is a
persistent belief among marketers that mix modeling somehow favors television, and that it underrepresents the impact of online, or is unable to measure online effectiveness at all. This belief comes
from the fact that the industry most advanced in the use of mix modeling is consumer packaged goods, which is one of the industries least developed at using the Internet as part of the marketing mix.
Our experience at Marketing Management Analytics in working across many categories is that the sales impact of the Internet is as measurable as any other marketing tactic and subject to all the
same modeling requirements:
>> There must be quality data, with enough data points from which to build the model. >> There must be enough spending behind a tactic so that its impact can be
isolated. >> The tactic must have a sales impact to isolate. Given these conditions, we are routinely successful in "reading" online effectiveness, and often find that Internet marketing romi is
comparable or superior to other tactics in the mix, including television. For instance, a financial services provider found that the Internet compares favorably to tv in its ability to cost
effectively contribute to new customer acquisitions. This client also found important synergies between tv and Internet, suggesting that the two tactics should be used in an integrated fashion for
maximum impact.
Companies that are frustrated with their ability to model the Internet in their mix are often in industries such as packaged goods, which are not typically "Internet
considered" purchases. In these industries, weak data and low spending are critical weaknesses. Here are some tips that can help overcome these challenges:
Spend enough to be meaningful. Mix
strategies are designed to tease out the sales impact of separate marketing vehicles, but can only identify elements that are meaningful. Depending on the amount and quality of your data, they can be
sensitive; they also have limits. If you only spend one half of a percent of your marketing on interactive, don't expect these strategies to register.
Stagger spending in short bursts. Be
strong when in. If your interactive budget is low, you may get a better read on its contribution if effort is pulsed in short, strong bursts.
Don't decompose digital programs. View your
integrated digital tactics as one "program." If you're trying to define the impact of Web site page views, ad exposures, click-throughs, and e-mails separately, you may be splitting some very fine
hairs. Instead, define a common metric and assign point values for each interaction. This approach might make interactive planners cringe and might not be the most efficient use of the
interactive budget. However, this trade-off may be necessary to help marketers get a "common basis read" on the dollar-for-dollar impact relative to the rest of the marketing mix. This approach is
already proving itself for one packaged goods client, despite a low Internet budget.
With even a few solid points of reference, marketers might be willing to allocate more dollars, making it
easier to measure the next time around. Whether ultimately effective or not, marketers will feel more confident in Web budget allocations when they are based on quantitative, apples-to-apples
evaluations.
John Nardone is executive vice president, product development and marketing, for Marketing Management Analytics. (john.nardone@mma.com)