The Hunt for ROI Continues
BY MARCIA LAYTON TURNER
Blame it on John Wanamaker. The turn-of-the-century
department store icon, who famously wondered which half of his marketing
dollars were wasted, started a quest for return on investment that continues
to this day.
So far, the results are not pretty. A recent survey
found major corporations using no fewer than six different outcomes to
measure ROI, everything from incremental sales from marketing to changes in
In fact, the confusion over how to measure ROI is such
that the Association of National Advertisers recently joined with the
American Association of Advertising Agencies to create a new way of
measuring how consumers interact with marketing.
“In this age of accountability, new metrics are
necessary to adequately reflect the impact of nontraditional messages,” says
Barbara Bacci-Mirque, senior vice president of ANA.
MASSIVE CALCULATIONS LEAD TO PREDICTABLE RESULTS
So, what’s a beleaguered ad executive to do? Some are
turning to econometrics, a highly sophisticated statistical modeling method
that has the industry talking. Supporters say it can help companies
determine ROI, predict outcomes and optimize their marketing spend.
Critics say something
along the lines of, “yeah, right.”
Essentially, econometrics involves stuffing a database
with historical data (usually one or two years’ worth) on sales, marketing
spend, product enhancements, even outside factors such as weather or
seasonal effects. Unlike other statistical modeling programs where you pick
and choose what goes into the analysis, econometrics takes everything you
It sounds simple, but it requires a massive number of
calculations and a boatload of data.
“You go through tens, maybe hundreds of thousands of
different models with an automated statistical modeling process and you’ll
eventually come up with the right fit,” says Randy Stone, chief executive
officer of Marketing Management Analytics Inc. (MMA), a consulting firm
based in Wilton, Conn. “It’s like a pattern-matching test. You measure
everything that was happening during a period of time, including marketing
campaigns, competitors’ activities, the economy, even the weather.”
Econometrics finds relationships among all the data
being analyzed, and that allows it to make predictions.
Once you understand how two factors, say sales and
television ad spend, relate to one another, that information can be used to
make predictions: Increase your spending X percent on television advertising
and sales will jump Y percent.
If it all sounds a little too good to be believed, well
then, Barbara Lewis understands. “The first question we get asked is ‘How do
you know how many people came in from a television spot?’ and the answer is
‘We do.’ Through analysis of all the data across many weeks, we’re able to
tease out the effects of the television spot on the number of customers,”
says Lewis, president of MarQuant Analytics, a Beverly Hills, Calif., firm
that consults with companies to capture, measure and analyze data.
Lewis adds that once she and her colleagues run
through the numbers, though, the skeptics become converts. “They see the
process and they understand,” she says. “Then they start to get excited.”
She cites the example of a financial services client
who uses econometrics to decide what to spend on marketing activities, its
products, each region of the country and even on certain types of customers.
It’s that ability to provide statistical insight into
decisions that used to rely on gut instinct that has marketers intrigued.
Well known in academic circles for years, econometrics
has only recently become the darling of the analytics set, thanks mainly to
the massive amount of data that companies now track and the leaps in
technology that make it possible for a laptop computer to run calculations
that used to require a supercomputer.
A number of European companies have adopted the
process, but U.S. corporations have been slower to follow.
Interestingly, Lewis notes, when doing analysis for
the clients she works with — a gamut of blue-chip names — direct mail
consistently has a high return compared to other marketing channels. “That
might not be true for every company, but it is for the companies we work
with,” she says.
LOOKING BEYOND ROI
But determining pure ROI isn’t the only game in town.
Companies are also seeking budget optimization to get the maximum lifetime
value out of their customers. It’s part of a movement in the marketing world
to look at return on customer, or customer equity, she says.
“They’re not just looking at ROI,” says Lewis. “That’s
short term; they want long term as well.”
Sometimes, for example, the corporation’s long-term
interest is in gaining market share, she says, and it is willing to take a
loss on its marketing in order to gain customers.
She also notes that larger corporations, with huge
budgets and massive amounts of spending, tend to find the most value in
“They spend hundreds of millions on their marketing
budget and the authority is so big, they can’t get their arms around it,”
says Lewis. “They can’t figure out what’s working and what’s not working.”
Right now, she says, most discussions about
econometrics are happening with the analytics departments within
corporations. Charged with finding ways to optimize spending, those
departments are looking for a scientific method that can help them produce
Ultimately, though, those conversations need to
include representatives from the financial side of the business — the people
asking about return on investment — so they too can understand the way that
marketing efforts can pay off for a company.
“We’d like to be talking at the chief financial
officer level,” says Lewis.
FOLLOW THE NUMBERS, BUT TRUST YOUR GUT
Not everyone is sold on econometrics, however.
Because econometrics focuses on what your company is
doing right now, it may not be as helpful if you are launching a new
product or expanding into a new territory, cautions Mark Dominiak,
principal strategist at Insight Garden, a Whiting, Ind., marketing
Consultant Erwin Ephron takes it a step further.
“Econometrics ROI is
a backward-looking measure, as you can only tell what you have done,” says
Ephron. “You can’t do econometrics modeling unless you have extremely good
disaggregated data. It only tells you about what you did, not what you want
Use it as an aid to making marketing decisions, but
not as a replacement, says Ephron.
Marketing Management Analytics’ Stone agrees. Don’t
rely solely on the numbers, he says, trust your gut, too. “Combined, you
have a winning proposal. Don’t allow the model to be the dictator.”
Still, in a world of maybes, econometrics holds the
promise of being one model that can actually help corporations get a handle
on exactly what they get in return for the marketing dollars they spend.
“We are looking for the ideal budget: How should the
company allocate its funding,” says Lewis. “And econometrics can help us
find that answer.”