Corona and the Limits of Marketing Controlling
We are all feeling the economic effects of the coronavirus crisis. These effects are severely impacting business models that we had previously relied on. In this VUKA (volatile, uncertain, complex, and ambivalent) world, advertisers need fast and reliable planning scenarios for their marketing strategies. The models used have to be fit for purpose and able to map the complexity of the marketing efforts and their underlying media strategies – taking into account the impact on higher-level business objectives. Only then can strategic business decisions be successfully adapted to the dynamic and complex economic situation.
Why Standard Models do not work in Exceptional Situations
Many traditional modeling approaches used in marketing controlling are now reaching their limits. The enormous effort involved in manual calculations and data collection means that they are neither flexible enough nor fast enough. Additionally, traditional approaches to media mix models are constrained in a rigid and largely linear model world, which means that they cannot meet current quality requirements relating to analysis and forecasting.
However, marketing mix models and multi-touch attribution (MTA) are still considered the standard for measuring marketing performance. Although both approaches have their strengths, the current unprecedented circumstances have exposed serious weaknesses. Media mix models are capable of incorporating online and offline channels as well as the impact of factors, such as seasonality and trends. But their calculations are based on highly aggregated data, which usually involves a lot of manual effort. This makes it impossible to provide, let alone utilize the data required for strategic business planning quickly enough. Instead, this complexity means that traditional media mix models are only recalibrated quarterly at best; the necessary workflows for faster model adaptation are simply not available in many cases.
Do we really want to use Yesterday's Methods to make Forecasts for the Future?
In the current coronavirus crisis, parameters are changing rapidly. Under these circumstances, a traditional media mix model is only able to map the situation with several months’ delay. Similarly, standalone MTA models at user level are unable to factor in the effect of long-term impacts (such as structural demand, seasonality, etc.) on sales.
Exactag's Case Recovery Modeling allows CMOs to successfully navigate the Crisis
Exactag combines the advantages of standard modeling methods in a comprehensive model. This allows a holistic view that integrates the latest developments as well as long-term effects. To achieve the speed and flexibility required for strategic adjustments, significant changes to traditional approaches are needed. One such change is the automation of the baseline calculation, the other is the expansion of calculations to incorporate additional modeling layers. Both changes are made during attribution and this is then referred to as Case Recovery Modeling.
Jörn Grunert, Managing Director of Exactag and responsible for the strategic direction of the company, explains the challenges for CMOs as follows: "Due to the fast pace of the market, it is now crucial for marketing success to keep the time frame between modeling and strategic decisions as short as possible. We have, therefore, extended our baseline models to include a market simulation as a decision support feature for fast and optimal scenario calculation. An automated process for data acquisition and processing is indispensable in reducing the time required for model adjustment and recalibration to just a few days.”
Exactag's research shows that external factors such as seasonality, vacation periods, etc. have a significant impact on online sales. Factors that are attributed to this so-called baseline, i.e. structural demand, are responsible for more than 80% of online sales in some companies. The extent of these external influences depends on the industry, product, and brand recognition. Long-term impacts also play a decisive role. Changes in purchasing behavior happen particularly in highly competitive markets and under changing macroeconomic conditions (e.g. disruptive business models, trends such as sustainability, etc.). Depending on the industry, these long-term impacts account for between 20% and 48% of online sales (see Fig. 1 and Fig. 2).
Managing the Crisis instead of being managed
Even before the coronavirus crisis, many markets were already facing disruption created by technology and digitization. Companies that had faced up to these developments have clearly profited from realigning their business and adapting to market conditions during the coronavirus crisis. Figure 3 shows that the coronavirus crisis is emerging as an additional driver of digitization processes. Exactag's research reveals significant differences in performance between companies with a strategic reorientation compared to companies without appropriate strategies (see Fig. 3).
Marketing Strategies for Companies in the Recovery Phase
For strategic marketing planning during and after the coronavirus crisis, two factors are of vital importance: the recovery of demand, and also the stabilization or recovery of the company’s own productivity (see Fig. 4).
If there is no significant recovery in demand in the markets relevant to the company, marketing strategies with a lot of advertising pressure and a strong presence are likely to have less impact on sales. However, if company productivity does not recover from the exogenous shock created by the coronavirus crisis, these market participants will be temporarily unable to meet demand in some market segments. This would also lead to a reduction in advertising efficiency as the company's offering would no longer be able to specifically map out and satisfy its customers' needs.
Exactag’s Case Recovery Modeling is intended to counter this scenario. The first step involved in model extension is to set up a scenario funnel in which several different simulation results are examined based on the company's plans. These simulations produce several scenarios based on extrapolating underlying data trends on the one hand, and the impact of the demand shock (dramatic reduction or increase in structural demand) and the recovery model on the other.
The model parameters for this market simulation are determined automatically based on the tracking data from Exactag, but the company’s existing restructuring plans and the forecasts of industry associations regarding the effects of the coronavirus crisis on demand can also be used to improve planning. The principle at work here is that the faster a company can reactivate its assets, the faster it can recover (see Figure 5). However, the greater the impact of the coronavirus crisis on demand, the less likely it is that the baseline, i.e., structural demand, will recover to pre-crisis levels in the foreseeable future (see Fig. 6).
To enable as many companies as possible to access this technology and scenario planning, Exactag has extended the algorithm for calculating baselines for its customers by adding automatic adjustment and recalibration. The model extension is available to all customers and will be accessible in the user interface as usual. These findings and adjustments also flow into the scenario calculations in Exactag's AdSpend Optimizer and are automatically used as the basis for each calculation.
The Importance of Strategic Business Planning
The shifts in these effects have an enormous impact on advertising efficiency. Studies by Exactag show that advertising campaigns in periods of low market demand generate some 42% fewer online conversions per euro spent than in periods of average demand. In comparison, if we look at periods of high demand, the advertising efficiency per euro spent is 30% higher than in periods of average demand.
The media channel used dictates whether the decline in advertising efficiency in periods of low demand will be greater or smaller. For example, the efficiency of display campaigns falls slightly less than the average. The efficiency of retargeting campaigns, in contrast, suffers massively from decreasing demand and is on average 63% less efficient than in periods with average demand (see Fig. 7). Purchasing interest tends to yield to an interest in information, which does not result in corresponding online conversions.
To translate these effects into concrete planning scenarios for the marketing and media strategy, the calibration of the AdSpend Optimizer has been extended to include the findings from the recovery modelling and the automated baseline calibration. This means that the parameters for recovery from the coronavirus crisis can be estimated weekly based on current sales figures. The baseline models, which determine structural demand, are automatically calculated and adjusted based on the latest developments. By doing this, Exactag ensures that the planned scenarios are still applicable. The automation and the rapid adjustment of forecasts create a solid basis for making effective strategic business decisions on marketing investments both during and after the crisis.
A Positive Outlook for the Future
In summary it is clear that the coronavirus crisis is turning companies’ marketing strategies and their current sales modeling practices upside down. But Jörn Grunert sees reason to be hopeful: "Although mastering these challenges is no easy feat, it is also not irresolvable. New technologies and the automation of processes, ranging from data collection and model adaptation to recommendation, already enable Exactag to overcome the biggest challenges today", he says.
For companies to use these models as quickly and meaningfully as possible, they will need to be adapted to standard models and modelling procedures. The following steps have to be considered:
1. Extension and automation of the standard baseline analyses
a) Implementation of automated models for decision making that quickly adapt to situations and shorten the decision-making process
b) Extension of baseline modeling to include recovery modelling so that the effects of the coronavirus crisis can be included
c) Automated calibration of model parameters based on available company data.
2. Simulation of possible scenarios in a scenario funnel, taking into account current sales data, plans for the recovery of the company's financial strength and liquidity as well as the development of consumer demand.
3. Consideration of possible changes in advertising efficiency and effectiveness, which can be explained by changes in consumer behavior.