Today, it’s more important than ever to use marketing models to quantify the effects of marketing efforts. Why?
First of all, because marketing professionals are under pressure to improve their accountability to regain their influence in the boardroom (Verhoef and Leeflang, 2009); and, secondly, because technological developments are creating a host of new marketing investment opportunities.
Over the last couple of decades, the amount of data available to marketers has exploded – culminating (so far) in the Big Data phenomenon. Today, data on individual consumers are available by the day, the hour, and even the second. Moreover, beyond more traditional data, such as purchase histories, we now have clickstream data, data on customer interactions with front-office employees via social media, and feedback provided by customers in the form of reviews, just to name a few of the possibilities.
Furthermore, in addition to highly detailed data and new and direct forms of customer feedback, media have become extremely flexible. Digital media, in particular, allow for quick adaptation. Websites can be changed in an instant and can even differ for different visitors. With the introduction of real-time banner bidding, online advertising decisions, such as whether to present an ad to a particular customer, are made in real-time.
Three challenges for better exploitation of your data
These developments – rich data, direct customer feedback, and flexible media – have given rise to novel challenges for marketers that, if well addressed, improve marketing accountability and create a competitive advantage for the firm based on better exploitation of the available data. Three of these challenges are to:
- Track the effectiveness of marketing efforts over time and spend when effectiveness is highest;
- Identify customer segments based on various data sources and approach each segment with the right message;
- Design and execute field experiments to establish clear causal effects and use the insights to improve allocation decisions.
1. Effectiveness of marketing efforts over time
Detailed data over time and direct customer feedback allow managers to track the effectiveness of marketing investments over time. In doing so, it is important to distinguish between short- and long-term, and even persistent, effects (Osinga, Leeflang, and Wieringa, 2010) and to allow for underlying trends in the data.
- Short-term effects are those effects that disappear as soon as the marketing efforts, which drive the effect, are stopped.
- Long-term effects last for a period of time after the marketing efforts have ceased.
- Persistent effects are those effects that remain even after the marketing efforts have ceased. Persistent effects can occur, for example, when a consumer tries a new brand because of a TV ad and then continues to buy the brand even when the TV ad is no longer broadcasted.
The effectiveness of marketing efforts may change over time for a variety of reasons, such as: the phase of the lifecycle the focal product is in, whether the company enjoys a positive buzz on social media or product review websites (Hanssens, Wang, and Zhang, 2014), the introduction of new products by competitors, the content of the campaign, etc.
Ideally, one should spend when marketing effectiveness is highest – i.e. when marketing efforts generate persistent or strong long-term effects. Advanced time-series techniques, such as Kalman filtering, are required for separating short-term, long-term, and persistent effects and to filter out underlying trends (e.g. Osinga, Leeflang, and Wieringa, 2010). Such tools, combined with rich, high-velocity data and flexible media, make it possible not only to track effectiveness over time but also to adjust spending, and even the effort’s message, in real-time.
2. Approach customer segments with the right message
Next to studying marketing effectiveness over time, it is important to approach each customer with the right message. Consumers may be in different phases of the purchase journey and may have different motivations to purchase. For example, one consumer may not even be aware of the existence of a particular product, whereas another may have a narrowly defined set of alternatives. Also, where one consumer may be enticed by a low price, another may look primarily at product reviews.
New sources of data make it possible to determine a consumer’s purchase journey stage and purchase motive with more precision and in real-time. Suppose, for example, that an energy supplier wishes to market solar panels to its existing customers. As soon as a customer logs in to the firm’s website, the energy supplier knows that customer’s current energy consumption, the relationship’s duration, and the projected lifetime value of the customer – and the supplier also has access to detailed information about clicks on e-mail newsletters, information gathered by service technicians, and browsing behavior on the firm’s website. These various data sources can be used to get an idea of which phase of the purchase journey the customer is in. For example, has the customer looked at information on solar panels before, and has the customer clicked on newsletter items about solar panels?
In addition, the data can be used to segment consumers and to determine which segment the consumer currently visiting the website belongs to. To create segments in the best possible way, it’s important to use data from a variety of sources. Service technicians, for example, can gather detailed information about a customer’s lifestyle. For each segment, the most appropriate purchase motive can be defined. These motives can be validated by a small-scale online survey. In the case of solar panels, purchase motives may include caring for the environment, lower energy bills, or being the first in the neighborhood to have solar panels.
Flexible media allow managers to act immediately on the insights. The feel or content of the website can be changed instantly (Hauser et al., 2009), the price may be adjusted based on the customer’s lifetime value, and different online ads may be shown after the consumer leaves the firm’s website.
3. Determine causality with field experiments
Finally, it is important to realize that highly targeted media may complicate the assessment of their effects, especially when cross-sectional data are used. For example, the finding that customers who saw an online ad are more likely to buy an insurance product is meaningless when these consumers were targeted on the basis of their propensity to buy. To establish clear causality, field experiments should be used – and, with the greater flexibility of media, they are easier than ever to execute (for an example of a field experiment in marketing, see Lambrecht and Tucker, 2013).
Randomized field experiments are the preferred option, unless it is impossible to link the performance metric to the treatment. For instance, when trying to establish the effect of online advertising exposure on offline purchase behavior, it may be impossible to determine whether or not the offline customer was exposed to the online advertising. In such situations, geo-experiments may provide a solution. In this case, to rule out alternative explanations for the results, it’s important to choose the areas in which the treatment is applied such that the customer characteristics in the treatment and the control areas are as similar as possible.
To conclude, new and rich data sources offer great opportunities to spend when marketing effectiveness is highest, to deliver the right message to each individual customer, and to test marketing initiatives in a clear-cut way.
To reap these benefits, it’s important that data be readily available and easily combined, that the firm commits itself to a data-driven strategy for the long-term, and that the firm be willing and able to act fast.
This article is an initiative of Trilations’ Advanced Analytics Center of Excellence, which employs the techniques described within the paper (among others as well) to support better decision-making and, therefore, decision-delivery.
For more information regarding this article or if you are interested in learning more about the services offered by our advanced analytics center of excellence, please don’t hesitate to contact us.
Author: Dr. Ernst C. Osinga is an Assistant Professor of Marketing at Tilburg University, the Netherlands. He obtained his PhD, cum laude, from the University of Groningen. His research interests include the development of dynamic models in the areas of pharmaceutical marketing, the marketing-finance interface, and (online) retailing. His research has been published in the Journal of Marketing and Journal of Marketing Research. Ernst provides guidance to major players in the pharmaceutical, advertising and retail industries.
Hanssens, D.M., Wang, F., and Zhang, X.P. (2014). Performance growth and vigilant marketing spending, Working Paper, UCLA.
Hauser, J.R., Urban, G.L., Liberali, G., and Braun, M. (2009). Website morphing. Marketing Science, 28(2), 202-223.
Lambrecht, A. and Tucker, C. (2013). When does retargeting work? Information specificity in online advertising. Journal of Marketing Research, 50(5), 561-576.
Osinga, E.C., Leeflang, P.S.H., and Wieringa, J.E. (2010). Early marketing matters: a time-varying parameter approach to persistence modeling. Journal of Marketing Research, 47(1), 173-185.
Verhoef, P.C. and Leeflang, P.S.H. (2009). Understanding the marketing department’s influence within the firm. Journal of marketing, 73(2), 14-37.