How do you feel about trees? Some people find them to be too much upkeep. They don’t want to rake piles of leaves in the fall or deal with root systems that may damage foundations or sprinkler systems. Others enjoy the shade trees provide and the sound made when the wind rustles their leaves. I am most definitely in the latter group. Unfortunately, I live in a newly developed area that used to be ranch land and there were very few existing trees in the area when development began. My lot had no trees on it until my builder placed three very slow-growing trees in my front yard. I quickly tackled the goal of creating a wooded lot by planting more trees in the front and back yards.
My parents also enjoy trees. They live close to me and they received the same three strategically placed trees in their front yard. Like me, they planted additional trees to enjoy and provide shade for their house in the hot summer months. A few years ago they planted a hybrid poplar tree in their back yard. The fast-growing tree was only a few feet tall when planted but it quickly reached a height of about ten feet and a trunk diameter of about five inches. It had a beautiful shape and was providing shade for the east side of my parents’ home. Early one morning I was shocked to find the beautiful poplar tree lying on the ground. What had happened?
I was quite puzzled because there had been no storm or high winds the night before and trees don’t just fall over. Further examination revealed a pile of wood chips around the base of the tree. Later when our entire family was out staring at the tree in disbelief, we concluded that a beaver from a nearby lake must have destroyed the tree during the night. One of the lower branches was missing and we presume that the animal hauled it off between sections of a wrought iron fence surrounding the back yard.
The loss of the tree was quite distressing to us all. One day it stood majestic, tall, and healthy providing shade for their home and the next morning…..gone. Naturally we all began to think about how the incident may have been prevented and the “if only” thoughts started forming. If only we had placed landscape fencing around the base of the tree or one of the bags used for watering. Or perhaps if we had painted the base of the tree with a mixture of paint and sand this tragedy could have been avoided. Of course, these were all moot points because the tree was gone…here today and gone tomorrow.
Sometimes this happens in business. A company is thriving one day and almost overnight circumstances change and the business is gone. We would all like to prevent business failure but can it be done? Is there a way to predict that your company is in trouble and facing impending death? Many theorists in various disciplines have tackled this question.
Two researchers who have analyzed various methods of predicting business failure are Sofie Balcaen and Hubert Ooghe. Their article “35 years of studies on business failure: an overview of the classic statistical methodologies and their related problems” published in The British Accounting Review describes various types of strategies for predicting business failure. According to Balcaen and Ooghe, the classic approach involves cross-sectional statistical methods in which a classification system categorizes firms as either failing or non-failing. Numerous methodologies utilize this classical approach including conditional probability models, univariate models, risk index models and multiple discriminant analysis (MDA models), the most widely used approach.
Although these classic statistical methods all approach the prediction of business failure somewhat differently, they have similar problems. One problem has to do with the ‘classical paradigm’ used in these approaches. The classical paradigm purports that with a set of companies that have known descriptor variables as well as known outcome class membership, a rule can be devised that enables other businesses to be assigned to a specific outcome class on the basis of its descriptor variables. The paradigm does not account for some crucial aspects of predicting business failure which causes problems that bring about misleading results because the models are sample specific and unstable.
One problem with the classical paradigm is the use of arbitrary definitions of failure in which firms are separated into non-failing and failing. Some studies use ‘financial distress’ to define failure, others bankruptcy and others use failure-related events such as loan default or cash insolvency. Firms are also artificially separated into non-failing and failing in such a way that the categories are only valid in a particular time period. Such categorization is also inappropriate because in the real world the failure of businesses is not a well-defined dichotomy.
Non-stationarity and data instability are also issues with the classical models. In order to accurately predict business failure the relationships among the variables examined must not only be stable over time but must remain the same in future samples. An assumption of the classical paradigm is that the distribution of variables in the model does not change over time but we all know that we do business in a dynamic environment that leads to ‘data instability,’ also called ‘population drift.’
In response to these and other problems with classical models of predicting business failure, other prediction models have been developed. In their article titled “A hybrid financial analysis model for business failure prediction,” published in the journal Expert Systems with Applications, authors Shi-Ming Huang, Chih-Fong Tsai, David C. Yen, and Yin-Lin Cheng propose the utilization of a hybrid prediction model that incorporates both classical statistical models such as those described above, as well as a back-propagation neural network (BPN) model. Huang and colleagues provide data from a study of four datasets indicating that the combined use of both statistical models and machine learning techniques is more accurate than the use of classical models alone. However, in spite of evidence to indicate the hybrid model is superior to previously used methods, it is not without its shortcomings and is not a perfect predictive tool.
So where does that leave the executive who wants to predict business failure? Should you place any stock in the available models? It is helpful to utilize such models, but remember that many factors play a role in the success or failure of a business and no existing model can predict failure with 100% accuracy. In addition to classical or hybrid approaches, my recommendation is to focus attention on one of your company’s most valuable assets—people. Your employees and customers are the lifeblood of your business and play a vital role in its success or failure. Knowledge of their perceptions of your company gives you the power to make changes that can prevent the downfall of your company.
In the difficult economic times in which we live, it makes sense to cut back on unnecessary expenses. While some executives may find it tempting to decrease expenses by eliminating employee and customer surveys, such a decision can have detrimental effects. Lack of knowledge about the attitudes and perceptions of your employees and customers leaves your company vulnerable. Let’s look at what one company found when they decided to begin conducting employee surveys.
Recently a large gaming company realized the need to gain knowledge of their employees’ attitudes and perceptions because employees impact the perceptions and attitudes of customers. The company hired the National Business Research Institute (NBRI) to conduct a scientific survey to reveal the key drivers of their employees’ perceptions and to learn how their company compared to others in their industry. The results were a wakeup call. The gaming company learned that when the attitudes and perceptions of their employees’ were compared to those of employees at other companies in their industry, their overall score was below average. I have yet to meet an executive or business owner who strives to have an “average” company ranking in their industry, much less anyone who would be happy with a below average ranking. The ranking was not the only problem; there was more bad news. A topical analysis of the survey items revealed that the company did not have a single topic scoring in the “strength” category (from the 75th to 100th percentiles). Instead, the majority of the topics scored in the “weakness” category (49th to 25th percentiles).
While the company’s ranking and scores were certainly disappointing there was also good news: they learned that taking action to change employee perceptions of just four key drivers could bring about improved ratings on 60% of the items in the survey. NBRI’s identification of the root causes eliminated the uncertainty faced by many companies after receiving the results of a survey. When numerous items have received low scores, where do you begin? If you hire a firm that uses predictive statistical techniques to analyze your survey data you know exactly where to concentrate your efforts and resources. Armed with this knowledge, the gaming company can now focus on just a few areas and should experience a noticeable increase in the scores and ranking on the next survey.
In today’s economic climate, conducting employee and customer surveys is something you can’t afford not to do. If you would like to learn more about how NBRI can help your business avoid being here today and gone tomorrow, call today at 800-756-6168. Taking action today can help your business fall into the success, rather than failure side of the prediction dichotomy.
By Dr. Cynthia K. S. Reed, Ph.D., Ph.D.
Organizational Psychologist & Sociologist
The National Business Research Institute