Is the Model Forward Prescriptive or Backward Diagnostic?

At a very fundamental level, there are two ways of building mental models of cause and effect. We can build them from left to right—that is, from the cause to the effect, something we refer to as forward prescriptive. Or we can build them from right to left, that is, from effect to causes, something that we refer to as backward diagnostic. While both approaches should ultimately link causes and effects, a natural question to ask is whether there would be a difference between the two. One can certainly argue that mathematically speaking, the forward prescriptive models and backward diagnostic models should converge. Further, would the data not tell the same story whether we think one way or another? Perhaps they would. However, we find that the two approaches lead to a vastly different set of questions that managers ask themselves and very different sets of places where they seek answers.

Generally, we find that forward prescriptive models are the norm as far as mental models go. Managers tend to believe that certain paths will lead to successful outcomes and plot strategies in order to move forward in that direction. Very often, these paths come from a silo-based perspective, where managers attempt to link the metrics within their silos to the metrics of overall firm performance. For instance, marketers might try to show the relationship between measures such as customer satisfaction and customer retention with measures of corporate profitability. Similarly, human resource managers might attempt to demonstrate the positive relationship between measures of employee engagement and overall firm profitability. Forward prescriptive models tend to have built-in intermediate markers that are assumed to be measures of success. Organizational effort then focuses on two things. The first is to invest in initiatives that are consistent with the model that will drive the levels of the intermediate markers to an acceptable level. And the second is to examine whether achieving these desired levels has the expected effect on the final outcome of interest, such as financial performance. The following are typical sets of questions that managers working with some common forward prescriptive models will ask themselves:

Brand Equity–Based Forward Prescriptive Model

1. Do our brands have high or low equity at this moment?
2. How do we build equity in our brands?
3. Does increasing brand equity increase our market share, pricing power, and profits?

Customer Satisfaction–Based Forward Prescriptive Model

1. How satisfied are our customers with our products and services?
2. What can we do to increase customer satisfaction?
3. Does increasing customer satisfaction increase our market share, pricing power, and profits?

A similar set of questions is put together for other forward prescriptive models that are built around customer loyalty, operational efficiency, competence, employee orientation, and other similar paradigms. What is noteworthy is that these paradigms rest on a key intermediate metric such as equity or customer satisfaction that is the focus of the entire forward-based thinking. Connections are then made from organizational initiatives to this metric and now increasingly from the metric to its consequences.

Backward diagnostic models start from the right-hand side and work their way to the left-hand side through flexible paths. Our experience is that these models are not as popular as their forward prescriptive counterparts, in part because they require a deviation from silo-based decision making. Their implementation is not tied to achieving targeted levels on preset metrics. Under this paradigm, one starts from looking at the most effective driver of the ultimate or terminal performance and the metrics that capture these drivers. The next steps then successively unfold these drivers to their simplest and most disaggregate form. This disaggregate set of actions constitutes the portfolio that would be expected to drive the ultimate financial performance. The following types of questions would be on the table using a backward diagnostic approach:

1. What are the drivers of margins in the business?
2. What is the relationship between market share and gross margins?
3. What are the drivers of market share across segments?
4. Should market share be increased across all segments or should certain segments be vacated?
5. What is the relationship between product design and unit prices and market share and margins?

As one will notice, no single paradigm sets the bounds for the series of questions, and the exploration can lead to very different sets of answers and solutions depending on what the circumstances are and what the data reveal. There are no universal metrics whose value needs to be maximized as a matter of rule. On the contrary, the process of discovering the appropriate metrics moves in a reverse direction, guided by the likely impact that they would have on the downstream measures of interest. In other words, if margin is a downstream metric of interest, then the backward diagnostic approach is completely flexible with regard to discovering the upstream metrics that drive margins. They might include high market share and low prices or vice versa. To that extent, the backward diagnostic approach relies on a data-driven discovery process to determine both the key metrics and the direction in which to drive them.

Overall, when one executes strategy, it might be worthwhile to explicitly ask whether one’s mental model is forward prescriptive or backward diagnostic. A mere consciousness of the type of model one is working under will help realize the extent of its flexibility and its ability to discover paths to profitability.