Decision Making: Simple Rules to Improve Analytics and Intuition

The increased use of analytics versus intuition in decision-making has been significant in improving the understanding and results of decision-making.

The growth has been particularly affected by the growth and confidence in behavioral economics fostered by authors like Daniel Hahnemann, Richard Thaler, and Michael Lewis. This article argues that there is no simple resolution, but there are simple rules to improve the decision process with both.

A great example of the tradeoff occurred several decades ago, when a poultry equipment company I was consulting with was moving their facilities from Michigan to Georgia to save money before everyone went to China. They had engineers analyze every phase of the manufacturing process, and successfully moved everything, with only a few people, to the Georgia facility.

When they shipped their first order, the installation manager called and said everything is great, except there is nothing to put the cages together with. The bottom line of the story is that the last step of shipping an order was the shipping manager grabbing handfuls of screws and nails etc. and just putting them in the shipment based on his intuition. They were never in the specs and thus never written down and his intuition had always been accurate.

Analytics is simply the increased use of models, probability, risk, numbers, and analysis to improve decision-making. In some simple cases, it has proved to be a valuable tool to understand and improve decisions or simply validate prior intuition – particularly, where there are lots of stability and historical data. For example, we have significantly improved results by helping clients focus on the 20 percent of customers or products that account for 80 percent of the sales.

As the bias switches from “we have always done it that way” to analytics, it can become easy to underestimate the value of intuition.

It is generally fast, experienced-based, and actually considers many factors (like the screws in the example above) that analysis often misses. It is almost required in cases where change, instability, uncertainty, and creativity are present. As George Bernard Shaw said, “The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man.” Similarly Steve Jobs said if he asked customers what they wanted, it be obsolete before he got it on the shelves.

There are two statistical concepts – regression to the mean versus outliers – that illustrate intuition versus analytics. Simply stated, regression to the mean says the odds are very slim at winning the lottery. An outlier is someone does win despite the odds. Most people spend lots of energy trying to be an outlier.

Our approach is to consider both approaches, and consider the appropriateness and effectiveness in each situation. Some simple tips are:

The greater the certainty, the more analytics can improve decisions. The greater the uncertainty, the more intuition is necessary to at least develop alternatives.

Intuition is essential in developing ideas and hypotheses. In contrast, analytics can be especially useful in analyzing results and developing new modifications.

In using analytics, be sure to consider the validity of the data, sample size, bias, uncertainty, and risk in making decisions that can undermine the use of analytics.

Remember the analytics are only as good as the least reliable variable. You need to understand the dynamics of how volume, price, and profit interact with marketing models. You can’t make it up with volume if you lose money on every transaction.

In using intuition consider how change, demographics, and the environment can affect tried and true beliefs.

Develop simple models like our profitability model that allow you to analyze the interaction of various factors with different intuitive alternatives.

In summary, the most important aspect of this discussion is to understand the use of analytics versus intuition in your decision processes. The assumptions, results, effort, and process can be greatly aided. In general more analytics is generally useful for small businesses; however, one must be sure the foundation, reliability, data, and processes of the analytics have a firm base.