Forecasting: Gaps and Plugs

It has been raining here in Dallas. It has actually been raining a lot here in Dallas. It is week twenty of a fifty two week year and we have now officially had more rain in the last twenty weeks than we had ALL of last year. The drought that has plagued us for the last four years is now officially over. The lakes are full. It can now stop raining. Please.

The reason I bring all this up is because it points out that at least for the last little while the weather forecasters in this area have had it pretty easy. All they had to do was mention the word “rain” in their forecasts somewhere and they were golden. They were going to be right. It was probably going to rain.

But even that didn’t seem to be good enough. It now became a forecasting contest to see who could be the most accurate in predicting the amount of rain we were going to get each week with each ensuing storm. Even this doesn’t seem too difficult. It’s week twenty and we have had almost twenty two inches of rain. I’m usually a pretty good numbers person, but this one should be pretty easy for everyone. We are getting a little more than an inch of rain a week here.

Forecasting. Go figure.

If only forecasting orders in business could be so easy.

For the most part it should be. We know how much rain (or how many orders) we already have. These amounts are called “actuals”. As an example, in Dallas we have actually had twenty two inches of rain so far this year, or “year to date” as we like to say. Management likes to work with trends. If we are trending at a little more than an inch of rain a week, they will more than likely expect at least another thirty three inches of rain over the next thirty two weeks. (Did you catch that? I told you I was good at math).

Knowing that everyone expects their fourth quarter to be their best quarter (for orders) they could conceivably expect even more rain, but we won’t get into that at this point.

Extending this example a little further, we could now say that we have a “gap” between the twenty two inches of rain that we “actually” have year to date and managements desired target amount of fifty five inches of rain for the year. This “gap” is obviously the target amount less the current actual amount. This is a pretty straight forward system and process. Take what you have and subtract it from what you want and there is your gap.

It works the same way for orders. Take the amount of orders you have (actual) and subtract it from what you want (plan) and there is your gap.

As time passes and more rain (hopefully) falls, the amount you have, your “actual” amount of rain, should grow and through the wonders of mathematics your “gap” to your desired annual rainfall target or plan should reduce.

But we have a slight problem. Despite what we have seen for the first twenty weeks of the year here in Dallas, we know that rain will not continue to fall at the rate of slightly more than an inch per week. We have a time here in Dallas that is known as “summer”. This is the time when you can replace the word “rain” in the forecast with the words “hot” or “heat”, and again be referred to as a brilliant and accurate forecaster. The only problem is that when it is “hot” in Dallas, it usually doesn’t “rain” much. It is usually dry.

Despite management’s belief and demand that twenty consistent weeks of rain performance does constitute an unbreakable trend, nature does not usually pay attention to these management expectations. There will inevitably be weeks where it does not rain.

Here is where weather forecasters and order forecasters begin to diverge. Weather forecasters would continue to look at the “actual” amounts of rain, compare it to the desired or “plan” amount of rain and calculate the “gap” or amount of rain needs to reach the plan. Order forecasters have developed this concept called a “plug”. A “plug” is something that is inserted into your forecast so that management can feel better about the team’s ability to reach the goal.

Dropping back to our weather and rain example, it can be expected to be both hot and dry in Dallas through most of June, July and August. It might rain occasionally, but it won’t rain at the afore mentioned rate of a little over one inch a week. It will be nowhere near that amount. Somewhere in late July or early August you can reliably expect the ground to dry out and start cracking due to the lack of rain. It would be fair to estimate that instead of the twelve or more inches of rain that the trend would show you to get, and that management would want, we might more realistically expect about two inches of rain in this period.

If this is truly the case, we would then expect to miss the annual rainfall plan of fifty five inches by as much as ten inches. From a weather and rainfall point of view this miss will probably elicit a collective “so what” from those of us who live here and see that the lakes are already full anyway.

If we are talking about orders however, this is unacceptable, unless you really want to invite a significant amount of management attention and assistance in your efforts to get more orders. So what happens here is that the orders forecasters understand that more orders are usually generated in the fourth quarter of the year than in the other times of the year, so what they will do is forecast ten weeks with more than two inches of rain in the fourth quarter.

Do they know if it will in fact rain this much? No. Is the plug amount twice as much as the rate for what is already one of the wettest years in a very long time? Yes. Do they have any idea as to if this increased performance rate is attainable?

In short a plug is something that is inserted into a forecast in order to make sure that the forecast ends up balancing with the desired annual target or plan. It is an as yet unidentified event or opportunity that is going to hopefully bring more rain after a dry period. In this example the forecasters do not know where they are actually going to find that extra / desired ten inches of rain that was missed in the summer, they are just committing to do it, somehow. It may not have any real substantiation, but it is now in the forecast so everyone now feels more comfortable about meeting the rainfall target.

The problem with putting plugs in a forecast is that they have a tendency to hide or mask an issue. As the actual performance diverges from the desired trend line, plugs have a tendency to be inserted. This may reassure some people that the target is still the target, but it does not solve the issue. Plugs are very good a defining what the problem is. The issue to be solved is how the lack of desired incremental rain is going to be obtained. Where is the incremental rain that is needed to reach the goal going to come from?

A “plug” in a forecast should be an alarm to anyone that sees it. A plug usually appears when there is a gap between actual performance and the desired goal. It is usually put in place to acknowledge that there is a gap and that there is every intention to try and close it. What it does is obscure whether the gap can in reality be closed and the goal attained. It provides the illusion of goal attainment when the reality may call for other actions to be planned or implemented.

There is a reason that weather forecasters don’t insert plugs into their weather forecasts. The lakes here in Texas for the last three to four years were well under their capacities. Water conservation measures were enacted to limit use so that things didn’t get any worse. Every year that the drought went on, the conservation efforts increasingly limited water use. Can you imagine what would have happened if rainfall “plugs” were inserted into the weather forecast in order to let everyone think that it was acceptable to continue to use water at the accelerated rate? When the shortage was finally acknowledged, it would have probably been too late and even more draconian measures would have been required.

Such is the case with orders plugs too. It is always best to acknowledge orders gaps and try to close them than it is to obscure them with plugs and have to deal with any potential shortfall consequences.

Goals and Processes

Whenever I find myself casting around for a topic to write about, I seem to always migrate toward one of my favorite conundrums: Goals and Processes. Is it goals that drive business processes, or is it processes that enable business goals? Because I have a little time, I think that I’ll go ahead and address this one some more.

Almost everything I see and read these days on this topic seems to be focused on the Process side of this question. It is interesting how the focus and primacy of management ideas and structures ebb and flow over time. In the past it did not seem to be such a Process focused set of business literature. I guess William Deming was one of the first to work in this area, and he did some ground breaking work. Having a Process focus is good from a predictability and repeatability point of view. Businesses like a predictable and repeatable outcome.

A good business process is like a security blanket. If you don’t know what to do, you can fall back on the process and hopefully expect to end up relatively close to where you were aiming to be at the end. Having a process reduces the risk and can remove uncertainty from the business.

At the risk of sounding like some sort of business contrarian I need to openly admit that I do not particularly ascribe to this way of leadership or business thinking.

For me an over reliance on process removes the value of people from the equation. They don’t Plan-Do-Study-Act as Deming said, they just follow the process. If they are following an industry “best practice” they probably are not even encouraged to think. They are part of a production line-like process. This seems to promote a very risk averse position for people. They can’t be wrong if they are following the process, and if they are wrong it is the processes fault not theirs.

I think a process is an extremely efficient and effective way to codify something that you have already done. That means that some way, somehow somewhere someone has already achieved the goal, and that the process has become the documented method that they used to achieve it. If you have been successful in manufacturing the first widget, then a production line process for all subsequent similar widgets would be called for.

When Sir Edmund Hillary climbed Mount Everest, there was no known process associated for a successful summiting attempt. People had been attempting the summit since 1921, but it was not until 1953 that is was actually accomplished. It was only after he was successful that the process of creating a series of ever higher camps, and the selection of which specific routes provided the greatest probability of success started to coalesce. In fact it has now evolved to a situation where the process for climbing Everest is so well defined that even novice non-climbers are now being taken up the mountain escorted by seasoned mountain climbing guides.

I think the cost for the Everest “guided expedition” is approximately sixty five thousand dollars ($65,000) and takes several months to prepare and execute.

I think this is a little bit of hyperbole, but it does illustrate my point. It took more than thirty years to achieve the goal. There was no defined process that anyone could fall back on. Every attempt was breaking new ground. We also need to recognize that just because there is a process does not guarantee that each climb will be successful.

To date there have been about eleven thousand (11,000) expedition attempts to summit Everest, with the vast majority of them since the year 2000. This would indicate that the process is reasonably well defined. However there have only been three thousand (3,000) expeditions that have ended in success. These successful expeditions resulted in only approximately five thousand (5,000) individuals that have actually stood on the summit.

In the mean time it should also be noted that more than 260 people have died while trying to climb Everest. That means that for every 20 individuals that succeeded in climbing Everest, 1 paid the ultimate price.

Thank goodness for the internet and Wikipedia. Where else can you get facts and statistics like that so easily?

So, what does all this have to do with the discussion of Goals and Processes in business? As I said, I think it illustrates several points:

First, if something has never been done before, there is probably no defined process available for doing it. People knew the process for climbing mountains. They had been doing it for years. It took more than thirty years and many unsuccessful attempts before they climbed Everest. They ended up creating a new process in order to do it. If you are trying to do something in business that has been done before, you had better come up with a faster, better cheaper way of doing, otherwise being the second successful one probably won’t get you too much.

Second, it was not the “process” of climbing Everest that captured people’s imaginations. It was the “goal” of climbing Everest that did. It is difficult to get people committed to a process. It is far easier to get them committed to a goal. The same goes in business. It is the goal that drives people to succeed, not the following of a process.

Third, just because you have a process that has proven to be successful in the past does not mean that it will continue deliver success every time. Less than a third of the expeditions attempting Everest are successful. Even well defined processes can fall victim to external environmental issues or potential team issues from within. Expecting to follow a process to get you to a goal without being prepared to deal with the unexpected or unforeseen enhances the probability of not being successful.

Fourth, even if only one or two out of the entire expedition actually get to the top of Everest, the entire expedition is considered a success. The goal is to get someone from the expedition on top of the mountain. If the goal was to get everyone to the top of the mountain, no expedition would be considered a success, and we would still be searching for the process to do it. The idea is to make sure that success is clearly defined and that everyone can participate in it.

I think it is reasonably apparent that it is goals that inspire people and it is the attainment of those goals that most people are measured against. Processes are good in that they provide a guideline on how to go about achieving the goal. But just like the weather on Everest, or the makeup and capabilities of the team attempting the summit, there are always variables associated with achieving the goal that cannot be accounted for in the process.

Processes are at their best and most useful when they are simple and allow for variances based on the environment surrounding the goal. It is only in the most repetitive of manufacturing production lines that a process can be fully relied upon, and even then it is subject to the vagaries of the humans doing the work.

As an example of this I would point to automobiles. They are produced primarily in a production line; however the quality of some cars can vary significantly. They are all produced by the same process, but some are acknowledged to be significantly worse than others. Hence the concept of getting a “lemon”, and the creation of “lemon laws” to protect the consumers unfortunate enough to have purchased one of “those cars”.

Business is lead and inspired through the use of goals. Processes can be of assistance in attaining those goals, but it is the goal and the measurement of progress against that goal that is important in generating progress. It is when business supplants goals with processes as its primary focus that the business will start to lose its way.