Melting the Tip of the Iceberg

This paper develops a Systems Thinking perspective regarding the interactions within a Help Desk operation. This particular Help Desk is providing computer support to a field sales operation.

We are by nature very myopic. We have a very real tendency to see just what's before our eyes, and seldom what's beyond the reflection or beneath the surface. We see what we choose to see. This myopic nature, coupled with our apparent penchant for reactive behavior, results in our missing some of the greatest opportunities for achieving progress in difficult situations.

This apparent natural character in conjunction with our well ingrained learning results in our experiencing situations, seeing them as problems, and reacting with perceived solutions. More often than not, we react with actions which are not really solutions at all, yet rather simply short term quick fixes which treat the symptoms, in fact ensuring that we will have to deal with the same situation again in the future.

We seldom actually deal with or solve true problems. We generally treat the symptoms until they subside to an extent where we no longer feel we need to treat them-and we foolishly think we have solved the problem, when we have in fact never even really addressed the problem. We are marvelous at deceiving ourselves into seeing and believing just what we want to see and believe.

The end result of all this myopic activity is that we miss substantial opportunities for creating the future. We spend our time treating symptoms, which alleviate the situation for some time, only to have the symptom return again, and again, and again.

What follows is not an analysis, and not a synthesis. It is both and neither. A term of recent origin which may be appropriate is anasynthis. What follows is intended to develop an awareness and understanding of a particular situation. The situation considered is a help desk operation where the call statistics are higher than desired. That is, the average time it takes to respond to a caller and resolve the callers situation is longer than desired. What follows is a synthesis of the character of the environment intended to provide sufficient understanding to design intervention which will enhance the current situation.

First, a note about reading the systems thinking diagrams which follow. An arrow between two entities indicates the direction of influence. The arrow from Current Call Stats to Stats Gap indicates Current Call Stats influences Stats Gap. On the influence arrow there will be either an S or an O, an S indicating the influenced variable changes in the same direction as the influencing variable, while an O indicates the influenced variable changes in the opposite direction. Thus, the arrow from Current Call Stats to Stats Gap with an S on it indicates that as Current Call Stats increase it influences Stats Gap to increase. The arrow from Desired Call Stats to Stats Gap, with an O on it, indicates that as Desired Call Stats increases it influence Stats Gap to decrease. Note that influences do not indicate the degree or magnitude of the change in the indicated direction.

Within a diagram an R indicates a reinforcing loop, and a B indicates a balancing loop. Balancing loops seek equilibrium and reinforcing loops are like a snowball rolling down hill. Also note that reinforcing loops may be virtuous, enhancing, or vicious, producing deterioration of the something.

Initially, the lower than desired call stats are most likely to be perceived as a result of a lower than appropriate level of analyst expertise, a situation well remedied by training.

This situation just described is depicted by a standard balancing loop system diagram showing the appropriate influences.

This diagram assumes that the initial condition that the current call stats are greater than the desired call stats. This variance between the current and desired call stats creates a stats gap. As the stats gap increases it will lead to an increased emphasis on training. The increased emphasis on training is based on the belief that training will enhance the composite level of analyst's expertise, enabling them to reduce the current response time, subsequently reducing the response time gap. An increase in training will, after some delay accomplish the desired result.

As the response time gap decreases a perception develops that the problem has been solved and the emphasis on training declines, with the expectation that the response time will maintain the newly achieved level. As it turns out, this is a very erroneous assumption.

The difficulty is embedded in that what is perceived as the problem is really only a symptom. Actions taken to resolve the problem really only alleviate the symptom which, generally after some delay, returns. Because of the time lag, the same fix that was previously applied is applied again, alleviating the symptom yet again. The net result of applying the fix is to ensure the fix will need to be applied again in the future. The burden has been shifted from finding and solving the real problem to training, and the organization develops an addiction to training. Thus, the more training is performed, the more training will be needed.

Now, as the newly achieved level of call stats is not sustained something else must be acting which causes the stats to decline. Actually there are several elements which cause the current call stats to decline, as anything which causes the expertise level to decline will influence the current call stats to rise.

The first influence is software changes as depicted in the following diagram. The environment being support by the help desk operation is in a state of continuing evolution. In addition to changing policies and procedures associated with the software being used by help desk clients there is an ongoing enhancement of the deployed software. These enhancements consist of electronically distributed changes to existing applications, as well as completely new applications.

When these software changes occur the expertise level of the help desk takes an immediate plunge, and, through training it rises back to its previous level.

With the realization of this interplay it is paramount that training be scheduled coincident with the software changes being fielded. Yet, even with this planning there are problems which will arise when the software is fielded, problems which no amount of planning and training will be able to prepare the help desk operation to accommodate. As these problems are experienced and the learning propagated throughout the population of help desk analysts the expertise level will return to its previous level, with a subsequent decline in current call stats.

Before considering turnover as the second element which has a major effect on expertise level it is probably best to consider additional effects of software changes.

An element closely linked to expertise level is call capacity, the volume of calls a number of analysts can respond to with a given expertise level. Call capacity is closely linked to expertise level, software changes, and software changes.

An increasing expertise level would serve to increase call capacity, ultimately producing a desired affect by lowering the current call stats. Yet, software changes serve to depress the expertise level, thus decreasing the effective call capacity, resulting in increased current call stats. Note the multiplier effect software changes have in that at the same time they depress the expertise level, they also serve to increase the call volume. And, the call capacity declines at the same time. Thus, software changes have three effects on the system each moving in an other than the desired direction.

Another very annoying interaction at the junction has to do with the way a decline in expertise level actually causes the call volume to increase. This happens because the resolutions to the client's situations are not as complete as they would be from a well seasoned analysts. Even worse, the responses provided may even be incorrect. Each of these situations results in additional calls from the client to obtain the desired level of resolve to their situation.

Well, I supposed its time we get back to turnover. Turnover actually comes in two parts, resignations and new hires. first the resignation part.

This diagram represents the addition of a reinforcing structure which should enhance the efforts of the balancing structure defined by training. As the expertise level increases call capacity increases and environmental chaos decreases. A decrease in environmental chaos influences an increase in morale. This increase in morale, after some delay will result in a decrease in turnover. The decrease in turnover causes the overall expertise level to increase thus reinforcing the cycle. This reinforcing structure should serve as a virtuous loop causing the efforts of training to be multiplied thus further increasing the expertise level. Yet, in the help desk operation being considered this was in fact not happening.

This points out the real nasty thing about reinforcing cycles. They may be virtuous, aiding movement in a desired direction, or vicious, hindering movement in the desired direction, or actually moving in a negative direction continuing to defeat ones best efforts, and frustrating the change agent. That is, they can multiply either positive or negative change. In this particular situation, resignations were depressing the overall expertise level, leading to an increase in the level of environmental chaos. This situation further depressed morale, leading to additional resignations.

The second part of the addition has to do with new hires. The addition of a new hire decreases the average expertise level, as the new hire has an expertise level less than the average. Secondly, addition of a new hire immediately decreases the call capacity because trained analysts must spend time training the new hire. After some time the new hire will come up to the appropriate expertise level and the call capacity will eventually reach a level above where it was before the new hire.

There is a question as to whether hiring actually depress call capacity directly or does it only acts via decreased expertise level? Initial perception is that since it removes resources from the task of responding to calls it is indirect.

By itself, a sufficient level of training should have stemmed the vicious operation of this loop and turned it into a virtuous cycle, yet this was not the case. Something else must be operating which is negating the potential gains from this reinforcing structure, but what?

There is a very deeply ingrained belief we hold regarding rules. This belief is one of the fundamental foundations of our entire political economic system. The belief is that if we establish rules and people follow them then things will be the way we want them to be. This shows up in the form of a balancing loop.

The logic is correct yet the premise is incorrect. If you established the appropriate rules and if people followed them, then yes, maybe things would get better. But, people don't follow the rules unless they choose to. The rules are only part of the structure in operation and when the rest of the structure is ignored the rules simply are not followed. In fact, they often have an effect just the opposite of what was intended. Would you expect a river to change course just be cause you told it to? Not likely! You have to alter the underlying physics of the structure within which the river flows, then it will change course.

With an increasing response time gap there will be an increasing tendency to establish rules of operation which are intended to remedy the situation. What quite often happens is that the development of additional rules of operation simply depresses morale still further as depicted in the following diagram.

This establishment of rules of operation fits into the architecture as a vicious reinforcing loop. With the establishment of more rules of operation morale declines. The decline in morale leads to increased turnover thus lowering the composite expertise level. The lower composite expertise level then results in an increase in call stats thus increasing the response time gap. And, this turns out to be just an added impetus to establish more rules of operation. Verily, the actions taken to resolve the situation only serve to make matters worse. Amazing isn't it! And we're not done yet.

As the response time gap increases, or even maintains at a substantial level for some period of time there is a tendency to increase the organizational emphasis on the numbers. The continuing emphasis on the numbers often plays into the structure in the following form. Note that a couple additional factors have been added in this diagram and they will be addressed shortly.

The implications of this interaction are very interesting. The continuing emphasis on the numbers results in a decline in the perceived meaningfulness of the operation by the members of the organization. As the perceived meaningfulness declines morale declines. What has been created is another vicious reinforcing loop, and the emphasis on the numbers leads to just the opposite result it was intended to. As the perceived meaningfulness declines, morale declines, increasing turnover, reducing expertise level, increasing the response time gap, finally driving an increased emphasis on the numbers, and we're back to where we started, only getting worse.

In the midst of this, the increased emphasis on the rules of operation also serves to decrease the perceived meaningfulness, thus establishing but another vicious reinforcing loop. This loop also serves to produce the opposite intended result as described in the previous paragraph.

What has been developed to this point only depicts a few of the influences interacting in the help desk environment being considered. Yet, the extent to which the situation has been elaborated should serve to show that organizational situations are anything but simple in their operation.

The normal response at this point is to ask, "so what's the answer?" This perspective is in fact part of our ongoing difficulty. As can be seen in the associated diagram, there is not "a problem" but rather a whole set of interaction which come together to create the situation being experienced. One of the fundamental understandings to come out of system science is that any intervention, if it is to have the desired effect, must be as complex as the situation which it is meant to affect. Also, there must be an admission that any understanding of the situation is only partial. With a continued deepening of understanding more appropriate interventions can be constructed which will have a higher likelihood of producing the desired results. One must continually strive to deepen the understanding, especially when thing are not emerging as expected. When this happens either something yet unidentified is influencing the situation or some previous assumption is incorrect.

For the current level of situation depiction appropriate intervention would be to decrease emphasis on the numbers, decrease the emphasis on rules of operation, and continue the emphasis on training. Also, training for to be introduced software changes has to be coordinated with the developing organization to ensure help desk analysts are as prepared as possible to support new releases of the software. The level of expertise needs to develop to the point where the morale loop becomes virtuous. Whatever it takes! If you only melt the tip of the iceberg there will simply be more iceberg to deal with later.

theWay of Systems * Feedback * Musings
Copyright © 2004 Gene Bellinger