Helping You Manage Your Supply Chain More Effectively!
Cycle Counting: Down for the Count?
By Don Rice
Warren, buyer at Consolidated Industries, had just finished his first cup of coffee. He always started each day by reviewing the material planning exception reports generated by the planning software. It always seemed like the best place to start except those days when there was a line of people standing at his desk with emergencies. Warren felt that the smartest thing he could do each day was to look at the order messages for his purchased items and get them on order first. He had quickly found out that itís hard to expedite something if you donít have an order for it. Next, he would look at the expedites messages and then if he had time he would then look at the reschedule messages and anything else that might be needed.Ē
The first order message was on part number H297Z321BLK455. Warren quickly called up the material-planning screen and looked at the details.
In spite of all the complaints he had heard, Warren believed that the new software had made thinks a lot better and easier than before. Color monitors laser printers in every department, spreadsheet programs at your fingertips, and speed. If these young kids had been here when we used the punch cards they might learn to appreciate what the company had given them to work with.
The on-hand (OH) balance for H297Z321BLK455 was 874 and there had been an inventory adjustment made this morning adding 862 to the OH balance. Now Warren, the seasoned veteran, new there were problems with the inventory records. Even though the annual fiscal had just been finished two weeks ago, the problems still hadnít been purged from the inventory system yet. Warren hated those fiscal inventory counts, not that he didnít like the overtime pay, but the inventory records always seem to be worse after the count than before. So the first thing Warren always did before placing purchase orders or releases was to verify the on-hand balance. He had learned long ago that to last in purchasing or materials there was one simple rule,
Donít run out!
The buyers heíd known that had let a line shut down because of a shortage, always seem to have a short career path at Consolidated Industries. He had also heard all the preaching about having to much inventory, but that only seemed to be a concern a couple of times a year.
So Warren calls down to the warehouse and some new guy named Bill answers the phone. Warren explains about needing an updated count on part number H297Z321BLK455. The nerve of Bill, he actually suggested that Warren use the on-hand balance. Well, you have to overlook the new people until they get used to the way things work at Consolidated. But you know the worst thing that could happen now is to get stuck at a lunch room table with the Bill and have to listen how they did things at his old company. Sheesh, sometimes Warren gets so feed up hearing about where other people used to work, that he wants to tell them ďIf things worked so well there, why did he leave?Ē A few hours later, Bill called back with an update, 12 pieces. Well, thatís a lot different from the 874 in the OH balance. So far, Warren had accomplished two things, he didnít know how much H297Z321BLK455 was in inventory and two numbers (12 & 874) had been eliminated from consideration. So Warren calls back down to the warehouse and asks for Susie, she has been making the warehouse hum for the last 7 years. He calls in a favor and gets Susie to verify Bills count on the H297Z321BLK455s. It took about 40 minutes but Susie was able to get back to Warren with an answer, ďI found 245.Ē Notice Susie didnít commit to how many we have but to only the ones she could actually find and put her hands on. Because the H297Z321BLK455 cost so much and the lead-time is so long, Warren wasnít going to take a chance. So he heads down to the warehouse and after about 2 hours decides that he would use 200 units. So Warren goes back up to his office and using a spreadsheet he had custom designed just for this, recalculated the material requirements and place a purchase order.
Thank goodness we have this new computer system to help me with my job. In the old days this would have all been by hand.
What is wrong with this picture? Is the software, peopleís attitudes or the inventory process? Well letís start with a simple question?
Are your inventory records accurate? Every time I ask this question to a group of people, someone laughs. In many cases, they laugh because their own experience tells them that inventory records are never accurate.
An effective tool for getting accurate inventory records is a process called cycle counting. When done properly and completely, cycle counting will:
1) Document the inventory data quality level
2) Find data errors
3) Determine root causes of data errors;
4) Create and implement a plan of action to correct root causes;
5) Monitor results of corrective action plans
6) Achieve 95%+ data quality.
If you quickly review this list, you will notice that adjusting the inventory records is not listed. Thatís because an effective cycle counting process is focused on improving the inventory process not just correcting the data. This article will focus on how cycle counting will allow a company to achieve 95%+ data quality and keep it there.
Many companies have used or are currently using cycle counting. And the results have been less than stellar. The most common causes of poor performance with cycle counting programs is the focus on finding incorrect inventory records and adjusting the on-hand balance. The net effect is that in many companies the cycle counting process is focused on an activity that it was not designed to do and therefore cannot do effectively. Therefore, many companies with cycle counting programs have not achieved 95%+ data quality because they have overlooked several critical steps in the cycle counting process.
Establishing a working and effective cycle counting process requires a company to go through three phases of activities:
1) Benchmarking phase
2) 80/20 phase
3) Real World phase.
As discussed below, each phase is designed to accomplish a different portion of the cycle counting process. Skipping or doing poorly in any phase will most likely lead to poor results.
This phase determines if a problem exists with the current inventory data and process. It begins by selecting a representative sample of active inventory items. Active inventory items are those that have transactions on a regular and frequent basis, i.e. daily or weekly. A representative sample is typically 20-100 items, depending on how many inventory items are in the physical area to be cycle counted. If there are a few hundred items (<500) the low end is acceptable. If there are thousands (>2,000) the high end is needed. A record is made of the perpetual inventory systemís on-hand balance for each item in the sample. Each item is then physically counted. Outstanding inventory transactions are reconciled to the perpetual balance. A realistic " tolerance is applied to each perpetual balance to achieve the current " piece count tolerance for the current on-hand balance for that item. See the last page for a short discussion on setting tolerances.
A comparison is made between the recorded perpetual inventory count and the physical count. If the physical count is within the " tolerance for that item, this item is considered a HIT. If the physical count is outside the " tolerance for that item, this item is considered a MISS. Count the number of HITs and divide by the total sample size. This is the inventory data quality level. This result is the starting point and is used as a reference point to check the benefits of the later efforts. I recommend that a bulletin board be put up in the physical area that the sample is taken from. The expectation is that 95% of all perpetual inventory records will be within tolerance. If less than 95% accuracy is measured then implement phases 2 and 3.
In the example above, tolerances were established based on an ABC stratification where A items make up the top 10% of annual usage based on standard cost, B items make up the next 10%, and C items make up the remaining 80% of items. The tolerances for each class are:
The diagnostic phase is used to discover and fix the root causes of the " big dog " errors. These are the errors that occur on a frequent basis or occur on many items. The premise is that 80% of the data errors are caused by 20% of the process problems. These process problems are what I call the " big dogs. "
The first step in the 80/20 phase is to select a small control group made up of the same type of active inventory items used in phase 1. This group most likely will be smaller than the sample used in phase 1. It is not uncommon to have control groups made up of 5-10 items. The reason for the small size is due to the intense level of activity given each item in this phase. Every item in the control group is reviewed each day. A record is made of the perpetual inventory systems on-hand balance for each item in the control group. Each item is then physically counted. Outstanding inventory transactions are reconciled to the perpetual balance. A realistic " tolerance " is applied to each perpetual balance to achieve the current piece count tolerance for the current on-hand balance for that item.
A comparison is made between the recorded perpetual inventory count and the physical count. If the physical count is within the " tolerance for that item, then this item is considered a HIT. If the physical count is outside the " tolerance for that item, then this item is considered a MISS. Count the number of HITs and divide by the total sample size. This is the inventory data quality level and is the result posted to the bulletin board setup in phase one.
Every MISS is then reviewed in detail to determine the root cause of the inventory error. One of the keys to being successful in this phase is doing intensive research for process problem root causes. TQM tools are extremely valuable in this portion of this phase. In many cases, the lack of TQM skills could result in a collapse of this phase. After the root causes are identified, the next step is to establish action plans to correct the root causes. This is the overlooked secret of a successful cycle counting process,-finding and fixing root causes for inventory process problems. The corrective action plans must be monitored daily. This process is repeated daily until the control group is at 100% for five straight days. A new control group is then selected and the complete phase is repeated until this control group is at 100% for five straight days.
After two control groups have been taken to 100%, the cycle counting team can decide whether to repeat phase 2 for additional new control groups or to proceed to phase 3. It is critically important that phase 3 not be started until a high confidence level is achieved with the 80/20 phase. This means that most if not all of the " big dog " errors have been found and corrective action plans are at least in place, if not finished. In some cases this could require 10 or 20 different control groups (depending on the size of the control groups). If phase 3 is started too soon then the cycle counting team will be overwhelmed with a high daily error rate and not have the resources to find and fix the root causes of process errors each day.
Real World Phase
The objective of the ongoing phase is to verify that the inventory process is working correctly and to find and fix the root causes of the infrequent errors. This is where the remaining 80% of root causes are found and fixed. In this phase, daily counts are made of different inventory items. Most companies set up a rotation or cycle that will establish a frequency of counting for each inventory item based on pre-established criteria. A common pattern is based on ABC stratification of inventory based on value. Almost any pattern can be used that is understood by the cycle count team.
Each day a new cycle count group is selected. This group is a subset of all items in the perpetual inventory system, and most likely will be larger than the control groups used in phase 2. Every item in the daily cycle count group is reviewed that day. The diagnostic phase counting technique is used again. A record is made of the perpetual inventory systemís on-hand balance for each item in the group. Each item is then physically counted. Outstanding inventory transactions are reconciled to the perpetual balance. A " realistic " tolerance is applied to each perpetual balance to achieve the current piece count tolerance for the current on-hand balance for that item.
A comparison is made between the recorded perpetual inventory count and the physical count. If the physical count is within the tolerance for that item, this item is considered a HIT. If the physical count is outside the tolerance for that item, this item is considered a MISS. Count the number of HITs and divide by the total number of items counted that day. This is the inventory data quality level and is the number posted to the bulletin board. During the ongoing phase, some companies change from a daily posting to the bulletin board to a weekly posting.
Every MISS is then reviewed in detail to determine the root cause of the inventory error. One of the keys to being successful in this phase is doing intensive research for process problem root causes. Once again, the lack of TQM skills could result in a collapse of this phase. After the root causes are identified, the next step is to establish action plans to correct the root causes. The corrective action plans must be monitored daily. This process is repeated daily.
Tolerances can be an extremely emotional topic. Letís take a few minutes and look at how to use tolerances for measuring inventory accuracy. Letís start by lookingat one perspective on inventory record accuracy. In the figure below, 9 out of 10 records are different but the overall total is dead on. Who checks record accuracy this way? Right, the fiscal inventory process works just like this if only these numbers were dollars instead of units. It is possible and highly probable that the fiscal inventory records be highly accurate while the perpetual quantities are wrong
If we look at these records again and ignore the overall total what is the data accuracy?
Well, some people would say that these 10% of these records are accurate. Letís look at part number L456 in more detail. Itís obvious that the perpetual and the physical donít match. Does that make them wrong? This is a screw that we use 5000 of each day. It was counted last about 12 workdays ago. That means that we have used about 60,000 screws since the last count. What are the chances that we might be off by .1%? Letís look at part number A339. This is a 6-cylinder diesel engine. We seem to have 3 extras. They cost about $2500 each. What a nice surprise this would be for us. We use 5 engines a day and itís been about 3 days since the last count. Should we be off by 20% on $2,500 diesel engines? I am inclined to argue that we shouldnít be off any percentage on these?
This is why tolerances must be established and agreed to for each inventory item before we start measuring our performance. This prevents the accuracy numbers from being debated. Because and you can trust me on this, everyone is going to have an opinion the accuracy of the inventory records. Tolerances can be established using complex mathematical formulas or using general rules of thumb. Both techniques work, itís just that the first typically takes longer and cost more to do. And in the end, a lot of times most people donít understand them. I recommend that tolerances be established based on specific business criteria. Some examples are:
The key to effective tolerances is reasonableness. This is an area where " hand grenade close " is good enough.
The Keys to successful cycle counting are:
D. R. Rice Company
9326 Lake Shore Drive ē Brentwood, TN 37027 ē 615-221-2196 ē firstname.lastname@example.org