May 09, 2016

Interpreting time in Marketing ROI

Companies invest in marketing ROI analytics to optimize KPIs

Companies typically invest in marketing ROI analytics to optimize investments, sales, margins, and other important KPIs. Even though the value of ROI is generally understood, there are different definitions of marketing ROI which carry different implications for business. Hence, marketers should understand which definition their analysts are using, given the implications.

Marketing ROI is the relationship between sales attributable to marketing and the spending on marketing. Sales attributed to marketing are commonly determined from marketing-mix modeling, but can be estimated from other analytics, such as test and control studies, primary research, and econometric analytics. Since marketing drives business overtime, the manner in which time is considered in the ROI is key to the implications.

Successful marketing drives KPIs during and following marketing execution
Successful marketing drives KPIs during and following execution of marketing. Investments thus generate returns over ensuing periods.

This form of ROI presents a number of problems. The most contemporary marketing investments lack periods thereafter to calculate the forward ROI. For example, May 2015 marketing may drive sales in May, June, July, and August 2015. Yet if the data ends in May 2015, the ROI will be underreported by excluding the impact of May marketing on June, July, and August sales. This contrasts with older time periods, which do not have underreporting.

Forward ROI metrics have business value, however. They show which marketing channels are expected to provide the best future returns, since budget planning is based upon expected future responses following marketing execution. They also enable decision makers to distinguish between the short- and long-term effects of marketing. Driving forward ROI and forward-looking sales are objectives, since marketing is designed to improve KPIs following its execution. Hence, the forward ROI is the metric of choice when trying to understand precisely what one can expect in the future for a given investment in a specific period of time.

Forward ROI is designed to quantify what can be expected in the future from a current or planned investment, and for determining the impact from any historical investment. The forward ROI should thus be in the tool belt of the marketing budget planner and marketing forecaster.

Understanding sales drivers during a specific span of time
To understand sales drivers during a specific span of time requires backward ROI analyses. Sales generated in a specific time-period are influenced by marketing from that period and from prior periods.

As with a forward ROI, this form presents a number of problems. The earliest time periods in the dataset will have mis-stated ROIs, since sales in these earlier periods could be due to marketing from even earlier periods. For example, if a dataset begins January 2014, any sales in January could be the result of 2013 marketing. Without earlier data, the backward ROIs for these early time periods will be misleading. This contrasts with recent time periods, which will have more accurate ROI reporting.

Backward ROI metrics have benefits, however. They show which marketing drove which past sales. Analysts can leverage this type of information when decomposing sales effects, since they are ultimately designed to attribute sales at each moment in time to all its sources.

Backward ROI is designed to quantify which historical activities drove sales for a specified point in time. Backward ROI should thus be in the tool belt of the brand and category analyst and those who manage insights.

Using the ratio of the total attributable sales to total marketing spend within a time period
Most common approaches to marketing ROI use neither forward nor backward ROI concepts. Instead, they use the ratio of the total attributable sales to total marketing spend within a time period.

There are implications of average ROI for strategic or tactical planning.

  • Recent marketing drives ROI beyond the data and the earliest sales were driven by marketing prior to the data. Both of these realities are excluded from the average-ROI calculation.
  • For example, the average ROI of 2015 excludes the impact of Q4 2015 marketing on 2016 sales, and excludes the impact of 2014 marketing on early 2015 sales.
  • Averaged over time, such as one year, the ROI is the less sensitive to changes in marketing. Therefore, it could take months of additional data to observe changes in the average annual ROI.

However, there are still benefits to average ROI. It allows comparisons of marketing channels, is less sensitive to noise and spurious activities, and is easy to compute. Given the value of average ROI to conveying marketing impact, this form of ROI should also be in the tool belt of marketing planners.

Ideally, analysts should report all three forms of ROI as they tell the holistic story of how marketing is expected to drive sales (forward ROI), how marketing has driven current and past sales (backward ROI), and which marketing tends to generally be most productive (average ROI). Since optimization is core to marketing, marketing decision-makers should leverage all ROI tools available.

About the author

David Hauser, PhD

David Hauser, PhD

Chief Science Officer, CPG and Specialty Analytics

Follow David Hauser, PhD on LinkedIn