By Mark Jeffery

A Microsoft and Kellogg School of Management collaboration

The business value of marketing is inherently challenging to capture. This article discusses best practices for return on investment (ROI) marketing and how to design marketing campaigns for measurement.
 
"Half the money I spend on advertising is wasted; the trouble is I don't know which half", said John Wanamaker (1838-1922). Marketing managers still struggle to quantify the return on investment (ROI) of marketing a century after the famous quote. Compounding the challenge, today consumers have a plethora of entertainment options with the rise of the internet, blogs, video games, podcasts, hundreds of cable channels plus on-demand programming, and digital video recorders (DVRs). This media fragmentation spells the demise of the 30-second TV ad, and dilutes the impact of traditional marketing. Mike Sands, chief operating officer of Orbitz and former chief marketing officer said, "I know half my marketing budget is wasted but today I know which half [the TV advertising]." For Orbitz, the value of online advertising is directly quantifiable.

Many senior executives today are frustrated that marketing does not have a better handle on value. As one marketing executive said, "Each week I go to a gun fight [the senior executive leadership meeting]. I'm tired of carrying only a knife."

Marketers sometimes like to argue that marketing is a creative endeavor that defies measurement. Yet high performing marketing organizations are differentiated today, not just by the creative teams, but by a culture of measurement. That is, they design all marketing campaigns to be measured. Following are three keys these organizations use to unlock marketing ROI:

(1) DEFINE THE RIGHT METRICS
The marketing behavioral impact model is an old idea—design marketing campaigns to take customers through the stages of awareness, evaluation, trial, and loyalty. Different metrics for measurement can be defined for each stage. Trial is the easiest: marketing in this category, such as coupons or discount promotions, is intended to promote short-term sales. Sales revenue is a natural metric. Since the time lag is often relatively short, and coupons can be tagged with an identifier, there is often a direct correlation for trial marketing campaigns and sales.

Loyalty is a little more challenging, since it requires a firm to know who their customers are. However, if the customers are known, repeated purchases can be quantified. The important metric for loyalty is churn—how many customers choose not to continue doing business with your firm in a year.

Firms with a direct channel to their customers have an advantage, such as Dell or Amazon, and retailers or banks. Firms with indirect channels have to be more creative in obtaining direct customer data. Microsoft for example sells the vast majority of software indirectly. To better understand their mid-market customers Microsoft established a Web portal with tools to help mid-market firms manage software licenses. Capturing these data enables Microsoft to gain significant insight into the software owned by individual mid-market firms. Through the portal they can quantify sales revenues at the firm level, customer churn, and take rates to marketing offers.

More difficult to quantify are evaluation campaigns. Such marketing might include white papers, or advertisements listing product benefits and features. Financial metrics are now not appropriate but one can still define metrics; product downloads or white paper requests are two examples.

Finally, awareness, or brand marketing, is well known to be the most difficult to quantify. Since there can be a long delay between customers viewing a branding advertisement and a purchase, the linkage to revenues becomes nebulous. One typically defines metrics like CPM (cost per thousand) or attendees at events. A key metric, however, is customer satisfaction since this is a leading indicator of future revenues.

The first step in quantifying ROI for marketing is to understand where a marketing campaign fits in the spectrum of awareness, evaluation, trial, and loyalty. Then define appropriate metrics. Beware though; defining too many metrics can lead to data paralysis. The best practice is to define approximately five to ten key metrics for each campaign.

(2) PUT THE RETURN INTO ROI
Wouldn't it be nice if there were a 'Killer App.' for marketing ROI? The good news is that there is—it's called Microsoft Excel. For trial and loyalty campaigns the return can often be quantified with hard dollars. The approach is conceptually straight forward: (1) Understand the base case, what will happen if you do not do the new campaign, (2) figure out the total costs of the campaign, (3) determine the upside revenue generated as a result of the campaign, and then (4) calculate the incremental cash flows, the net cash generated as a result of the campaign. The financial return on investment (ROI), net present value (NPV), and internal rate of return (IRR) can be calculated once the incremental cash flows are known.

Firms selling indirectly do not need to know who actually buys the product to do this calculation. Rather, they can measure the change in sales through channel partners as a result of the campaign.

Look out for a common pitfall of financial ROI analysis, however. The assumptions used to define the new revenue generated as a result of the campaign are critical to any model. The best practice is to use conservative assumptions, based where possible on market research. At the executive level, conservative assumption make the ROI analysis more believable.
 
A best practice approach to probing the importance of assumptions is to do sensitivity analysis. With the Table Function in Excel it is easy to vary parameters in a model and see the best and worst case. This range of possible outcomes gives insight into the risk of the project—if the best and worst cases are very different this suggests the campaign has a lot of risk.

Campaigns at the awareness and evaluation stages do not fit the financial ROI model, but one can still define a base case for important metrics such as customer satisfaction prior to a campaign. Management must then weigh the anticipated change in the metrics as a result of the campaign, and decide if this change is worth the cost

(3) DESIGN FOR MEASUREMENT
Fail fast and often. Rather than creating big bang campaigns, high performing marketing organizations are continually experimenting. They build flexibility into the campaigns and design small experiments to test concepts.
For example, Harrah's Entertainment, the world's largest gaming company, designed an experiment to test existing offers. The sample consisted of frequent slot players in Jackson, Mississippi. The control group received the regular $125 package: free room, two steak meals, and $30 of free chips at the casino. This resulted in the usual response in subsequent months. A test group was given a new offer: $60 of free casino chips. The gaming activity was higher than the control group for the test group in the subsequent months-this experiment demonstrated that less was indeed more.
Similar experiments can be designed to probe effectiveness of awareness and evaluation campaigns. Control and test focus groups can be used to look for changes of qualitative metrics such as customer satisfaction and intent to purchase.

WHAT'S NEXT? TAKE THE SCORE CARD CHALLENGE
Marketing measurement is a creative endeavor. A simple but effective exercise prior to each campaign launch is to have a campaign team create a score card and think through how the metrics will be measured. This exercise can takes just a few hours, but the results can have significant impact on campaign performance. Defining metrics up front and thinking through how they will be measured is the key to best practice ROI marketing. The team activity also helps build a marketing measurement culture.