Internal Decision Support Modeling: How Much Causality is Enough? And not Enough?

Causality, or cause and effect, ranges from strong to none in a relationship.  It is the primary principle for internal decision support modeling.  How much effort should be put into identifying causality for an internal decision support model?

First, let’s get back to basics money isn’t the source of causality.  Business processes and resources are the source of causal relationships. This means you need to look beyond the money when you think about causality.  An operational model and the characteristics of resources and processes should always be the basis for a monetary model. You should not divide money further than you can divide a resource…even though money is infinitely divisible.

Second, causal relationships may only extend so far, not all the way to a saleable product or service.   For example, a product line has 150 products and 5 product groups.  The product line manager has 6 analysts that evaluate the performance of the product line’s business. The following questions should be considered:

·       Should the product line manager and the analysts’ costs be assigned to each of the 150 products?  

·       They are clearly a cost of the product line, but is there a clear causal relationship to each product or even a product group?

·       Perhaps time could be recorded to assign costs to the product groups, but what if they are evaluating new products?  

·       Would the analysts’ costs be reduced if 15 products were eliminated?

·       Would an analyst (or a part time analyst) be added if 15 products were added?  

I would argue the product line manager and analysts are a cost of the product line level of the business.   Assigning these costs to individual products would distort the marginal cost analytics and decision-making because there are no savings if the product sales volume goes down or the product is eliminated, and no increase costs if sales volume goes up or if a new product is added. 

The alternative argument is these costs should be included in the product cost so the product price and profitability can be evaluated more accurately.   Clearly, the costs of the product line must be covered before the product line is profitable.  If a “best fit” allocation isn’t made to individual products, then the product line costs are essentially allocated by revenue.   While this argument has some validity, including allocated, non-causal costs to an individual product or service unit level distorts more decisions than it enhances.

There is a solution to these 2 positions, the product line costs could be allocated using a “best fit” driver, but labeled clearly as fixed so they are not included in most marginal profitability decisions.  This also requires some training or sophistication among the users of a product or service cost.  One problem is that very few costing approaches or systems go to the effort to label costs a fixed and variable at the product cost level.

A third consideration is that not all costs have causal relationships with products and services.  In some cases, customers or customer behavior is the causal cost driver.   This requires multi-dimensional costing for an organization.  Many logistics, warehousing, order taking, customer service, marketing/sales, accounting, legal, and other organizational costs not typically assigned to saleable products or services have causal relationships to customer behavior, cost, and profitability.  These behaviors by customers and the associated internal process are manageable and can be improved if the causal relationships and processes are modeled operationally for performance evaluation and with the associated costs.

How much causality is enough?   To improve profitability and operations cost information must be causal (which also means very weak and non-causal costs cannot be allocated beyond clear causal relationships).  The real question is:  How much do you want to make cost and profitability improvements?   Because to do so, you will need causal operational performance and cost information.  And you will follow that trail through your organization as long as you are getting improvement.  You’ll know when you have enough and when you don’t.