Every company, big or small, uses some form of measurement to assess the performance of it’s people and the company as a whole. Measurements are of course the indicators of where we are and where we are supposed to go. Well, that’s not enough. We have to realize that we are dealing with human based organizations and the one factor that has the most power to impact behavior of people is the measurements. If you measure someone in an illogical way, don’t complain about his illogical behavior.
Therefore measurements should be put in such a way that they drive behaviors that are good for the company as a whole. This in fact can be used as a yardstick to evaluate each measurement that you choose. If a measurement triggers behavior that would jeopardize the system as a whole implies that the measurement is wrong.
Every business uses various measurements to assess it’s performance. Financial measurements, for example, measure the performance of the company as a whole. The profit and loss statement indicates net profit, the balance sheet indicates ROI etc. There is absolutely no issue with such measurements.
The trouble starts when we do not measure the company as a whole and dive in to measure performance internally. This invokes the need of measuring two sections. First is the section of policies for making decisions and the other one is the section of execution. Basically, how well the policies are followed in making decisions.
When it comes to measurements at execution level, we need to realize that we are dealing with human beings. If we are measuring a person on more than five measurements we are actually creating anarchy. Simply because with a whole lot of measurements the person gets a chance to work on whatever he wants and score on it rather than focus on what is good for the company as a whole. This impedes any efforts being put in by the organization towards bringing in ‘Ongoing Improvements’. The top managers today are more aware than ever that if they do not lead their company on a process of ‘Ongoing Improvement’ it would just be a matter of time when there would be no company to lead.
The question here is that if we want to inspire people to continuously improve, what should we concentrate on measuring when it comes to execution:
- Do we concentrate on measuring what is done properly OR
- Do we measure what is not done properly
The proclivity in general is to measure what is done right, score on it and incentivize on it. Perhaps because people like it this way and we want people to be happy. However if we want people to improve quickly, this kind of measurement will not trigger the right actions by people towards what we want viz. to improve quickly. Therefore it makes more sense to concentrate on measuring what is not done properly. If we now look at the list of things that are not done properly in an organization, it would be a long list.
What is actually needed here is a way to classify/categorize the things that are not done properly in a manner that there are no more that five categories, they should cover everything that is not done properly and should have zero overlap. Though this is no triviality yet the answer to this puzzle actually seems so obvious when revealed. Here is the answer, it’s just two categories:
- Things that should have been done and were not (Cat A)
- Things that should not have been done but nevertheless were done (Cat B)
These categories may seem very generic and obvious but they have a much deeper connotation. Each of these categories actually represent a very important characteristic or personality trait that people should have to contribute to the organizational objectives.
The first category ‘things that should have been done but were not’ denotes ‘Reliability’. Simply put, if a person is doing a lot of this would you consider him as reliable or unreliable? Of course unreliable.
The second category ‘things that should not have been done but nevertheless were’ in a way implies that the person in not lazy, he is busy doing something but he is doing things that are not actually contributing to the desired results. This has nothing to do with efficiency. It is about being ‘ineffective’. Though the person is sweating a lot but for things that do not matter. Thus the second category denotes ‘Effectiveness’.
Every manager knows that Reliability and Effectiveness are essential to deliver results. The need therefore arises to quantify these traits, put numbers to them in order to measure them. Intuitively we know that these soft traits cannot be measured in financial terms. It would be prudent therefore to understand the impact of each category on the company as a whole to arrive at some measurement.
The unavoidable result on the system, as a whole, of not doing things that should have been (Cat A) is that the company is not be able to meet customer commitments on time. It will not be able to ship in time. Do businesses measure shipment on time? Of course, every organization measures shipment on time and generally the unit of measurement is ‘percentage’. For example, if the company had ten orders for the month and nine were shipped, the resultant ‘due date performance’ (or whatever term is used) is 90%.
As we have said earlier, the measurement should induce behaviors that we want, let us examine if percentage is the right unit of measurement for ‘due date performance’. In the example above if the total value of the nine orders delivered was Rs 9000 and the value of order not delivered was Rs 9,00,000, the percentage still remains 90% but the revenue achieved is just 0.99% of the total order value! Perhaps it was tougher to meet the bigger order so in order to maintain a high percentage lesser value orders were delivered. Is this really the behavior we want? We all know the answer, which brings us to the conclusion that the ‘order value’ has to be part of the reliability measurement.
The other aspect is delays in delivery. If an order is delayed by a day or two, the customer may not even notice it but if a shipment is delayed by two months, we won’t have the customer any more. ‘Time’ is therefore the second important factor that needs to be considered in assessing reliability. In measuring due date performance in percentage, delay time is also not a factor. Based on factors of value and time, the equation for measuring reliability would be:
Summing it up for all the orders that are late is the total measure of reliability. The lesser the number, the more reliable you are. A division factor can be used to bring the number between 1 -10 which would be easier to understand instead of a figure running into lakhs or crores.
What is the global impact of things that should not have been done but nevertheless were (Cat B)? It creates inventory. By doing things too early the net effect is piling up of inventory. Once again every business measures inventory and it’s done either in terms of value or number of days or maybe both. A company might say they have Rs 35 lakhs worth of inventory or it might say they have about 45 days of inventory. This does not necessarily imply that the company has 45 days of inventory for all the products. The bulk of it would be for one or two products and when there is a pressure to reduce inventory, it’s this bulk that is dissolved first which in turn affects customer service moving ahead.
Piles of inventory gets created when either a supervisor wants to save setup time on a machine and runs a large batch or when the plant manager wants to show higher efficiency to the management (because the management demands it). The result of measuring and insisting on ‘efficiency’ drives behaviors that create mountains of inventory blocking finances. Yet, almost every management, every manager and every supervisor insists on efficiency simply because of the common belief that an idle resource is a criminal waste. The resource could be the worker or the machine.
If efficiency induces the wrong behavior what should then be measured that will induce the right behavior? Efficiency is creating inventory and reducing effectiveness. The motive however is to increase effectiveness while reducing inventory. The measurement that can help achieve this balance is:
(summed up for all finished products)
Measured in this way, a supervisor will never save setup on a machine to run a larger than required batch, a manager will not be pulling his hair out if a resource is idle because now they would know that running after efficiency will only create a heavy burden of “E” on their shoulder that they will have to carry for a long time.
A more appropriate way of representing ‘R’ is Sales Value Days (SVD) and for ‘E’ it is Inventory Value Days (IVD).
Thus if we want effective measurement for each subsystem we should use Sales Value Days (SVD), Inventory Value Days (IVD) and LOE i.e Local Operating Expense. LOE is of the least importance because the changes there are the smallest. LOE basically means all the operating expenses that are under the control of the sub-unit being measured. For example, the salaries of workers and overtime are under the control of production department which they can try to reduce.
At times, organizations are known to load expenses of support departments like accounts, Computer Department etc. also on production. The logic given is that they are a user of these departments but one should bear in mind that the production department can do nothing to control the costs of these departments. It is a wrong practice to measure someone for something that is not under his control. Such loading should therefore never be resorted to.
One of the major concerns of a business is product quality. A number of processes, procedures and parameters are established to ensure quality. Often times there is an overspend by companies on this aspect as they obsess about brand image. In principle the businesses are not wrong. Quality is important, brand image is important for sustained business and and profitability.
This emphasis on quality however, transfers complete ownership of maintaining quality to the quality assurance department resulting in a situation where a worker or supervisor begins to consider quality as something outside the scope of their responsibilities.
So if quality is so important is there a need for another measurement? The answer is no. This is the beauty of the SVD & IVD measurement system. To understand how, let’s say a company has delivered a defective product to a client and the client finds about it three weeks after delivery. He comes back to the company complaining about the problem with the product. With the SVD system of measurement what would it imply? It implies that the company has not delivered yet. The goods are returned and now lie with the company. As long as the issue is not fixed, the product is not delivered.
In terms of SVD, the company is still carrying the selling price of this product multiplied by all the time it takes to fix the issue. That is a huge amount of SVD on the shoulders of all departments involved. It would also affect any incentives they were supposed to receive. The quality issue will therefore trigger a behavior where people will jump through hoops to ensure quality before items are shipped rather than the client finding out about it weeks later. SVD in effect ensures what Edward W. Deming has tried to teach industry all this time, Quality at Source.
With SVD therefore, there is no need to have another measurement for quality. The team will ensure quality even if it means generating half a percent of scrap more (scrap is a LOE), and in turn the business delivers customer satisfaction and maintains it’s brand reputation.
If you give it a deeper thought it would be easy to see that these measurements encompass every aspect of the business holistically. It would in fact not be an exaggeration to claim that these measurements are the only thing that will solve the problem of coordination, communication and collaboration between silos.
The policies that a company defines or the rules that it makes are based on the assumptions made behind such policies and rules. Even when things do not go according to expectations, assumptions are rarely challenged. Social science has provided enough evidence of human tendencies towards “if everyone does it, it must be right’ and to ‘compromise’ when there is inconsistency. Hard sciences however go the opposite way. If something does not give the desired or consistent result, they dive straight into the underlying assumption to find the root of the problem.
SVD and IVD as measurements that induce desired behaviors are a result of challenging the assumptions of measuring local efficiency and local improvements. Local efficiency and improvements are important too but the impact these have on the company as a whole is very small. In human based organizations, the process of Ongoing Improvement can only be sustained if the measurements used by the organization induce behaviors in people that are good for the organization holistically.
* Acknowledgement: The work and research of Dr EM Goldratt published in his books, The Goal, Critical Chain and Beyond The Goal *