Decision makers in marketing organizations continually face questions like: What level of overall spending is right? How should that spending be allocated across geographies, brands and channels? What specific messages should we send to whom and via which channels?
To answer these questions marketers need to understand the ROI of their marketing efforts, which is not a straightforward task. According to Hubspot 40 % of marketer report that proving the ROI of their marketing activities is the biggest challenge
So why is measuring the ROI of marketing efforts so challenging? The reason is that there are so many factors that needs to be considered in order to determine what effects are results of a specific marketing activity e.g.
- Money invested in marketing today will most likely have an impact on sales in the future
- The customer journey is non-linear and goes across multiple touch points
- Extraneous variables that are outside the control of marketing (e.g. sales reps activities, promotions, competition)
Fortunately, there are different methods that can handle these challenges - each with its own advantages and limitations. This article discusses the various aspects to consider when selecting methods and when to use them.
The simplest approach is the RCQ model (reach-cost-quality). In this method the different categories of marketing spend (digital, paid ads etc) and the impressions generated from those spends are determined. The quality of the spending categories is then estimated, using heuristics to generate a common currency across marketing touch points - an RCQ score per touch point (an estimated cost per reach).
RCQ is relatively easy to setup and provides a high-level comparison of ROI per category. The obvious drawback is that it relies on heuristics - which is too simplistic in most cases - and considers impressions rather than actual sales (e.g. the number of persons who saw a specific ad). Therefore, RCQ cannot be considered a fully data-driven approach. However, it can serve as a good starting point for marketers who currently relies on experience and rule of thumb rather than data when it comes to strategic and tactical planning.
Another, more sophisticated approach is attribution modeling. In this method credits are assigned to different digital touchpoints based on the clicks that the touchpoints have generated.
A common attribution model is the Last touch/Click attribution, which assigns all credit to the final touchpoint before the conversion. Multi-touch attribution models (MTA) based on advanced analytics are also used. MTA models use regression techniques to quantify how much influence that different touchpoint, across the customer journey, have on conversion.
MTA models based on advanced analytics can in some cases provide insights into the relative influence of different touchpoints. The method has however been critisized due to the big challenges in accurately tracking the customers through the entire, non-linear customer journey across multiple channels and touchpoints - a problem that has not been reduced with the increasing awareness of data privacy. Moreover, MTA does not consider offline sales, extraneous factors such as prices, seasonality and promotions and the important brand equity that is a result of the total sales and marketing efforts spent over time.
Marketing Mix Modeling
While MTA may be useful to gain insights into the effectiveness of different digital touchpoints, it cannot explain the impact of marketing for companies generating a major part of their sales offline. In these cases, Marketing-Mix Modeling (MMM) is useful.
MMM is an econometric modeling approach that correlates, through multivariate regression techniques, the marketing spend per aggregated category (digital, paid ads, print etc) and other sales drivers (promotions, seasonality, competition, macroeconomics etc) to sales. Industry and brand specific time lag between marketing spends and impact on sales as well as baseline sales are also considered with various degree of sophistication.
The challenge with MMM is that it, in some cases, requires extensive mapping of what data is needed and to collect data for several years back in time. Furthermore, MMM does normally not consider the relative importance of digital touchpoints (although it is possible if data is available) and cannot capture long-term effects of marketing spends nor the impact of marketing spends that are constant over time.
MMM, however, is very useful for answering the marketing ROI question. MMM also provide high-level insights into if, how much and when different categories of marketing spend impact sales, understanding of the relative importance of other factors, possibility to simulate various business scenarios for strategic planning and improve marketing budget allocation.
No model is perfect - and the right one depends on the business question at hand and the situation of the organization
Since each model has its own advantages and limitations, marketers should use a mix of models and use them to answer specific business questions. While MMM could be used to understand the marketing ROI and gain a high-level understanding of how different marketing strategy and tactics influence sales, MTA can create a more granular understanding of how different digital touchpoints drive sales. Even RCQ can be useful when data is limited or as a starting point in a transition to data-driven marketing.
As with any advanced modeling it is essential that marketers work closely with data scientists and analytic translators who understand both the business and analytics side to develop insightful models that answer the business question at hand. In addition, fine-tuning models is an iterative process that requires continuous input from the marketers. Change management and training is also a part of the equation to succeed with adoption since marketers need to understand and trust the model and adapt ways of working to make data-driven decisions concerning marketing tactics and strategy.
And as with any measurement methods, marketers need to start from the business objectives and company's current situation in order to identify what is important for them. Based on that, the most applicable method and approach to measuring marketing ROI can be chosen. Becoming more data-driven is a way for marketers, regardless of industry, to create a real competetive advantage for their organization.
Interested to know more about how your organization can start or improve measuring marketing ROI? We will help you. Contact us and we will help you with the next step for your organization.
And don't forget to share this post!