Wednesday, September 28, 2011

Prediction: There's No Single Method.

Share this ARTICLE with your colleagues on LinkedIn .



In my most recent post about using simple linear extrapolation as one of many possible tools which could prove useful in forecasting and preparing for certain events, I'm afraid that I made the terrible mistake (it isn't my first, and will not be my last) of failing to issue a whole roster of circumstances where this approach would not be ideal (if applicable at all), with an accompanying list of limiting parameters where this approach might certainly need to be adjusted significantly if not discarded altogether.

My intention [and perhaps the road to Hell is paved with good intentions] was to cite one very simple approach to discerning a trending pattern and where it might lead. I apparently mislead at least one reader into believing that I thought that linear extrapolation was the absolute best means to make predictions. Now, I hang my head in shame. The linear extrapolation approach to predicting the future has the following limitations, to state just several of a vast number:

1) It assumes a constant (linear) rate of growth or decline over time, which is seldom the case -- the patterns of history do not necessarily repeat themselves, rates of growth and decline are subject to variation, as are many other items. These factors may render our initial projection assumptions invalid;

2) It does not provide for changes in technology, resource availability, demand, competition, regulation, obsolescence, economic markets (fundraising, the cost of capital, the money supply, interest rates, and the like), natural disasters, amazing breakthroughs or the intervention on the part of the futurist making the prognostication in influencing the results via a mechanism correlative with the notion of a self-fulfilling prophesy;

3) It  does not provide for limited markets, holding capacities, or changes in either of these.

In a comment that I received regarding my now infamous post, Xamuel (whose insightful quote is printed below, unretouched) set me straight. Both The Global Futurist and Douglas E Castle humble ourselves before you, and apologize for misleading any of our readers. We also offer our sincere thanks for having this pointed out to us and our readers. Your input, negative, positive or neutral, is welcomed and encouraged. We learn from our mistakes.
---------------
Xamuel said...
Linear extrapolation might work sometimes in, say, climatology. It's not going to predict the future worth a hill of beans, though. It completely misses any revolutionary new development-- by its very nature! Even besides that, it has a bad habit of ignoring holding capacities (say), and predicting (e.g.) that a sample of bacteria will rapidly fill up the entire universe.
Although I have been told never to make generalizations, I will take exception with this advice with reference to Futurists (and those people who love them): There is no single method for prediction. There are many approaches, but each has its own inherent limitations. My approach to drawing a futurescape involves a great deal of imagination, a great deal of data collection, and a combination of a large number of approaches, each with its own respective strengths and weaknesses.
Sadder still is the fact that unpredictable events happen during the course of the extrapolation period (which makes any forecast a moving target) which cause a need for us to re-evaluate our assumptions and our methods to be changed dynamically. Forecasting requires constant monitoring, adjustment and correction for the effects of reactive intervention or interference, self-fulfilling prophesies, and other recursive variables that blur the line between predicting the future and causing the future.

Tuesday, September 27, 2011

Forecasting Simplified - Growth Rate Extrapolation

Share this ARTICLE with your colleagues on LinkedIn .



























In generating predictions about the future, one of the mechanical mathematical operations we often use is simple linear extrapolation. This approach is very elegant, very conservative, and is broadly applicable to a myriad of  forecasting situations involving relative (differing) rates of growth. For example, if we are attempting to project what our company's gross profit margin on sales will be in three years, we might collect historical data about our price per unit sold (one variable) and about our direct cost per unit sold (the other variable).

Based upon prior data, we might determine either an average or anticipated rate of growth per year in each of these two "competing" variables. If we simply graph the respective rates of growth in each of these two competitive variables (usually I prefer to use a compound interest growth formula) we can see the difference between the two at any given time by superimposing the graphs.

For an example of this, you might wish to visit my Braintenance Blog post, located at:
http://braintenance.blogspot.com/2011/09/real-economics-lesson-growth-rates.html .

Simply hit the "BACK" button on your browser to return when you've finished.

The reason that this growth rate extrapolation so useful is its simplicity in terms of citing and illustrating trends involving two or more variables. Generally speaking, the greater the difference in annual growth rates, and the greater the passage of time, the larger will be the gap between the two. In some situations the two lines will eventually converge and cross each other -- in others, the two distance between the two lines at any given point on one axis will simply continue to increase, and never converge.

If we now enjoy a gross profit margin of 30% (for illustrative purposes), but our price per unit can only be increased by 3% annually due to competitive or economic concerns while our direct materials supplier is one of the only sources of supply accessible to us, and that company has been consistently (despite our pleas for mercy, and our whining about college tuition for the kids, confiding about Aunt Tillie's desperately-needed life saving surgery, sharing our worst fears about how we'll soon actually be losing money on our sales of our principal product, and the like) raising costs at an average of 7.5% annually, we can estimate the amount of decrease in our margin at any given time in the future during the forecast timeframe, and we can even determine at what point our gross profit margin will simply be too small to continue our manufacturing and sales of that particular product.

The notion and application of differential growth rates is one of the most crucial steps to making any predictive analysis.

This type of procedure is incredibly simple -- the greatest challenge to the prognosticator is determining which growth rates are most appropriate [and sadly, in the real world, they are seldom linear], while the greatest challenges to leadership and management are to monitor these changes and to prepare an actionable agenda to prevent the fulfillment of these prophesies.

In The Global Futurist, we paint scenarios of the future if trends or tendencies are permitted to continue along their current paths. A proactive strategist can change the future by his or her early intervention. Free will and the prospect of precipitous action are the "Human Factor Wild Cards" that can alter the sequence and path of events that lead to the future. The future is never certain, as long as 1) forecasting is not an exacting science, and as long as 2) Human Beings can choose to act upon their expectations of the future and take the initiative to change outcomes.

Douglas E Castle [http://www.linkedin.com/in/douglascastle]

http://SendingSignals.blogspot.com

Monday, September 19, 2011

The Global Futurist: New Resources And Tools

Share this ARTICLE with your colleagues on LinkedIn .



Dear Readers:

When you are able to find a quick moment, I would be deeply appreciative if you would be kind enough to visit the actual site for this blog, located at http://TheGlobalFuturist.blogspot.com and take some time to explore the site and look through our extensive library of links and our trends, breaking news and RSS feeds sections. It is our intention to provide excellent articles (original content which you'll not find elsewhere, but also to be a resource, research and reference center for you to bookmark as one of your favorities.

If you have additional links to resources (websites or blogs) or to RSS feeds which yoiu think would be of benefit to other of our readers and colleagues, please suggest them in a comment of this post, and I will see to it that they are given serious attention. Your input is heartily encouraged.

As always, thank you for reading me, and for your referral of this site to your friends and associates.

Faithfully,

Douglas E Castle

Blog Archive

Bookmark and Share