The field of marketing has experienced various changes over the past decade. Digital marketing, social media presence and sponsored content are becoming increasingly important, but traditional channels retain their attractiveness to many organizations. Machine learning and better data availability will turn the game around for agile organizations with the courage to implement advanced analytics solutions.
Return on Marketing Investment
One of the most crucial aspects of marketing is effectivity. Therefore, associated financial outcomes should be evaluated and analyzed when marketing spending is decided. It is feasible to calculate return on marketing investment (ROMI) for each channel and marketing campaign.
Marketing resource allocation can be considered in the light of return evaluation. ROMI modelling also allows earnings forecasting in which marketing spending is used as an input. Simulations increase demand predictability and thus improve decision-making. The greatest financial benefits are typically realized in large B2C organizations.
Data Requirements
ROMI modelling requires revenue and margin data. Furthermore, organizations have to document marketing spending by channel, product and time stamp to utilize the approach. The quality of the results is largely dependent on data completeness, length and frequency. The picture below illustrates the sales of a product sensitive to national holidays with relatively sizable seasonal fluctuation.
Earnings Modelling
Earnings generation has to be modeled in order to obtain an estimate of ROMI for each marketing channel. The determinants of earnings vary according to product characteristics and operational environments. The resulting time series model can be deconstructed into smaller parts to obtain a quantitative estimate of each element with an effect on revenues.
It is often feasible to include a trend component when earnings modelling is carried out. The baseline of the model is built on observation history and endogenous behavior of the dependent sales series. Identification of relevant external factors is also essential to modelling success. Structural anomalies can be filtered out with a suitable statistical method often based on residual standard deviation.
Earnings generated by each marketing channel are displayed along with the rest of the model components. ROMI can be calculated based on these cash flows. The statistical model which inspired this blog post explained 94 % of the underlying variation in sales.
