Do I Really Need a Quantitative Model for Decision Making?

This is a common question I heard from small business owners. Some felt that their business environment is so complicated and full of uncertainties. No model can capture the essence and provide a good suggestion. Others think their years of experience has equipped them with the ability to call a decision using their mental processing power. Both have their merit but also miss a key point.

For the former, the claim is true for higher level decisions, especially those at the strategic level.  Decisions for a longer horizon tend to involve higher level of uncertainty and factors harder to quantify. However, at the operating level, decisions are made for the short term, in which all relevant pieces are more certain, for example, developing a employee work-schedule.  Even for a long term plan, a model can help depict the general relationship between key factors and provide some insights.  Also, certain types of long term decisions, by their nature, are number driven and have well developed models available, for example, capital budgeting.

For those who said experience is good enough for them to make decision, it might be true. However, imagine if a manager can model her decision-making thought process into a model. It reduces the time and effort when making a similar decision next time.  The model, in a sense, preserves the manager’s decision process in the model as artificial intelligence. This releases some burden from the manager and allows her to focus on more critical and strategic issues where more human intelligence is still needed.     

In summary, a quantitative tool as primitive as  factor rating table or as sophisticated as a machine learning model can serve as a good starting point for a decision process.   Managers can exercise their judgement to accept it, modify it, or disregard it.  It, nevertheless, provides a basis for discussion. Decision making with the support of data analysis makes the final decision more robust, as W. Edwards Deming said, “Without data, you’re just another person with an opinion.”