Statistics and Performance

I always thought that Mark Twain was purported to be the author of one of my favorite quotes:

“There are three kinds of lies: lies, damned lies and statistics.”

I have come to find out that in his own autobiography Twain attributed this quote to a nineteenth century British statesman and former Prime Minister Benjamin Disraeli. This fact perplexed me slightly so I continued my research a little further using those modern bastions of all knowledge, Google and Wikipedia. After an exhaustive five minute search I found that the general consensus is that no one knows for certain who the author of one of my favorite quotes is. This fact will make it reasonably difficult to give attribution in the future should the opportunity to use it come up again.

Be that as it may, it may be time for me to take the slightest exception with one of my favorite quotes. When you are looking at the performance of a business, the numbers don’t lie. Now the way the numbers are arranged can sometimes be confusing or even misleading, so the business leader needs to be aware and careful.

I don’t say this too loud or too often, but I think I may understand numbers reasonably well. Physics, Mathematics, Finance, I have studied them all. And believe it or not, to one extent or another they are all numbers based disciplines. In addition to this degreed book learning, I have had what could almost be called a moderately useful stint of practical numerical application in the business world.

I sometimes use this unbridled numerical capability and familiarity to complete the most difficult of Sudoku number puzzles in the USA Today newspaper, or various in flight airline magazines. Just to stay in practice and make sure that I still “have it”.

Since it appears that this is going to be a quote based discussion, I might as well continue in that vein. Hippocrates, the ancient Greek physician (hence the source of the physicians “Hippocratic Oath”) said:

“There are in fact two things, science and opinion; the former begets knowledge, the latter ignorance.”

I think the corollary here is that if it cannot be expressed in numbers in business it is not fact, it is opinion. I have written and spoken in the past about the need for leaders in the business world to become increasingly more familiar with “numbers” of business as they matriculate up the leadership ladder. There may have been past instances of corporate wizardry where a leader intuitively knew what needed to be done (maybe Steve Jobs, or Bill Gates fall into this category), but their moves were invariably backed up by the analysis (numbers) justifying their moves. For the rest of us, as Robert McNamara said:

“Get the data.”

While I am muddling along focusing on equating science, facts and numbers, I probably should pay at least some heed to the power of opinion. While the truth may be out there, it will be people’s opinion of it that drives its valuation. As John Maynard Keynes, a man who is credited with some of the very foundations of economic theory (think of the source of Keynesian Theory) said:

“A study of the history of opinion is a necessary preliminary to the emancipation of the mind.”

Okay, enough quotation roulette. I hope you get my point, whatever it was.

Now back to the topic; in the business world, the proper statistics, when properly presented and interpreted are invariably a good indicator of business performance. That is correct. Statistics, which are the indicators associated with past performance, are usually good indicators of future performance. I understand that Mark Twain, Benjamin Disraeli and whoever actually created one of my favorite quotes regarding statistics are all probably collectively grumbling, wherever they are, but this is the case.

This brings us to probably the most common and clichéd quote in this dialog:

“Past Performance is Not Necessarily Indicative of Future Results”

We have all probably seen it or read it on just about any investment prospectus ever written. Why is it there? To make sure that we all know that just because an investment that has done well in the past does not mean that it will continue to perform, or do as well in the future. Investment firms don’t want anyone to cry “foul” if they have not interpreted the statistics properly or if a statistical performance anomaly occurs in the market. So with this in mind I now ask the following question:

When we have the choice of making an investment, or are reviewing the continuation of an existing investment, which investment alternatives do we choose: those that have done well in the past, or those that have done poorly in the past?

All things being equal, and pursuing an intelligent risk diversification portfolio, understanding long term investment return and interest rates, blah, blah, blah…..we almost always pick the one that has done well in the past expecting (hoping) that all things being equal it will continue to exhibit the same performance traits going forward. We very rarely select the one that has been performing relatively poorly expecting (hoping) that it is going to turn around. Why do you suppose that is?

Welcome to the world of statistics. You have taken the data points associated with performance in the past, extrapolated the line or curve forward and made a choice and prediction about the future. Statistically speaking that is probably the correct choice. But this isn’t a discussion about investments. It is a discussion about business. And the same exact concepts apply.

Herein lies a rub. Statistics can only be misleading if you don’t understand the underlying numbers. Hence my predilection and continued haranguing regarding the necessity of leaders being numerically literate. Remember there is another quote that may also be incorrectly attributed to Mark Twain:

“Figures don’t lie, but liars do figure.”

The author of this quote also appears to be shrouded in the past as well, but like all the other good quotes of the period, it was attributed to Twain.

Let’s look at a simple statistical example to make my point. Let’s say that in the next measurement period (it doesn’t matter how long the period is) that a business sells one more product unit than it did the previous measurement period. This is good right?

Basically the answer is “yes” selling more, any more is usually always good, but how good? If it is your first measurement period in business, it’s hard to say how good without more data (statistics). If you only sold one unit the last period, and then you sell two units this period, that is sales growth, but it is still difficult to evaluate without more data. If you sold a thousand units last period, then one more this period might not be statistically important.

On the other hand all of those responses could be changed depending on the cost and value of the product being sold. If you are selling nuclear power plants as opposed to canned hams, a single unit growth in sales could be seen as spectacular, whereas the sale of a single additional canned ham might not be a cause for much celebration.

Statistically speaking selling one more than nothing is infinite sales growth but it is still only one unit. Selling one more than one is one hundred percent growth, but it is still only two units. Selling one more than one thousand is only one tenth of a percent growth. All represent the same one unit growth, but can be represented significantly differently in the statistical growth example. The wary leader needs to always be aware of how statistics are being used and the story that is trying to be portrayed.

Again, when using data and statistics unless a business can specifically quantify what changes it is going to make and how those changes are going to be translated into performance, like your investment decisions discussed earlier, you would expect the business to perform very much in the future as it is doing today. It is through this process that the market valuates companies, and it is through this process that companies provide their future forecasts of performance.

Leaders always need to be aware that statistics are extensions of the data. They are the way that the data is being presented and interpreted. The data is the fact. It is the consistency of the statistic, the interpretation of the data that is the key. Understanding the underlying numbers, and the analysis and statistics associated with them is required and essential for the successful leadership of the business. To not be able to do so is to be at the mercy of those that do.

Because as Mark Twain also (and finally) said (and this one is actually directly attributed to him):

“Facts are stubborn things, but statistics are pliable.”

Leave a Reply