Horseshoes, Hand Grenades, and...

"Numbers are pure and true, counting never is"
-Michael Blastland and Andrew Dilnot from their book The Numbers Game



There was a great line from an episode of the Simpsons many years ago. In this episode Marge had taken a job as a real estate agent. Initially she struggled to close any deals due to her honesty and integrity, leading to a confrontation with her boss played by the weasely Lionel Hutz. Lionel Hutz instructed Marge to bend the truth a little when making a sale, at one point he added that there’s The Truth(all serious and focused) and then there’s "the truth" (whimsical and head-bobbing). Lionel proceeded to tell Marge that she’d be fired
if she didn’t make a sale. This moment seems to describe a verycommon aspect of the working experience. From the outside we see want a company wants us to see. But behind that surface of shiny gloss and sheen lays the mechanics of how things operate, of the ugly brutality of business, and in some cases the layers of fraud that stay hidden from view. No matter how a company operates on the inside, it's important for a business to maintain a spotless public persona, especially to those on the outside looking in, who don’t have the access to see what really goes on. The phony, glamorous façade seems to be essential to business today, and one can see countless examples of this in the ever omnipresent advertising that exists today. Yet there seems to be one rather easy way to get past this layer of sheen, and that’s by working, by becoming employed with a company. In a lot of walks of life, there seems to be a Lionel Hutz moment at a certain stage in a job, when one realizes that things aren’t what they appear to be from the outside, when one sees how a business truly operates, and when one realizes ultimately what their goal is, to make the company they work for successful and profitable, and if they don’t they’ll be gone. The concept of the phony public facade is no stranger to the world of inventory counting. It might be more prevalent here than in most industries. I once applied for work with an inventory company called Quantum Services, the HR person who I interviewed with brought along some pamphlets and company information that she actually referred to as "propaganda"; a description that was not entirely inaccurate. Similar forms of propaganda exists on inventory services’ websites, and on the slogans of their inventory tags, from phrases like "Accuracy is our Primary Concern" to "A Company Committed To Service", to "The way it should be!". I have spent a decade working in the inventory service industry, and the biggest example of that phony façade that I see is in the notion of counting. No doubt that if Lionel Hutz were an inventory counter, he’d be talking to an unspoiled new hire saying that there’s counting and then there’s "counting".

The official public policy of auditing companies everywhere is that estimation is wrong and unethical and not to be tolerated. In a monthly letter sent out by Quantum Services they talk about several things that shouldn’t be done in an audit; taking short cuts, plugging numbers, estimation, and collusion. In other materials they offer tips like ‘never compromise your accuracy and integrity.’ Over the years I’ve collected numerous handbooks from RGIS and every single one has contained the same phrase:
"Unauthorized estimating or batching has never been, and never will be, permitted or tolerated in any event/inventory performed by RGIS"
Older versions of the handbook, used to have another phrase on the inside cover that read "NEVER ESTIMATE". Jack Henry, a Loss Prevention expert who once spent 15 years working in the inventory counting business, has a more realistic take on estimation. From a series of articles linked from the NAAIS website, Henry writes, "An inventory service that tells you they "never" estimate should not be hired. Not only do they estimate, they are lying to you as well". Okay, let's establish one thing right now just to get it out of the way; estimating is a part of the counting process. I myself have done it, and so has virtually everyone who’s ever held a counting machine in their hand. Anybody who's spent any decent amount of time counting inventories and who says that they never estimate is lying. Some level of estimation exists, and more importantly it’s tolerated in spite of what the propaganda says. The most disconcerting word from RGIS’ credo is "Unauthorized", this seems to allow some wiggle room for ‘authorized’ estimation to occur. There's a great blog called "Tales of a RGIS Auditor" written by a former counter for RGIS out in California identified only as the Misfit. Even on his blog you’ll see people talk about the right way and wrong way to batch. Quantum will talk about not using shortcuts, but in their propaganda will advocate a method of counting where one counts what isn’t there, which has the feel of a shortcut. Henry goes on to say that "Reputable inventory services do permit some estimating, yet stay within their code of ethics". I find this statement by Henry to be a little scarier than the first. Granted one has to be somewhat realistic in reagrds to inventory counting. Counting is not easy, and there's going to be a lot of challenging things to count in any store. The "never estimate" policy is extremely difficult to uphold. All one has to do is look at a chest full of ice bags, or a pile full of mulch to understand that even the most hardcore of counters will be forced to take a more pragmatic approach to counting every once in awhile. Estimating sometimes occurs out of necessacity, and shouldn’t always be interpreted as a nefarious act. An inventory service shouldn’t boast about never estimating, rather they should claim to estimate ‘less’ than any other service. The biggest issue with estimation is the frequency in which it should be used, where do we separate the "unauthorized" from the "authorized"? How 'pragmatic' should an inventory counter be? Where this line gets drawn depends for the most part on the service and the counters themselves, and too often it gets drawn primarily to serve the inventory service more than anybody else.

If an audit is a process that measures the amount of shrink in a store, then estimation is the purposeful inclusion of measurement error. The main reason estimation is done is to expediate the audit process in areas where exactness and great precision are not necessary. The most common and accepted form of estimation comes in the way sales adjustments are figured. When doing financial audits (especially in C-stores), counts will take place while the store is open for business, creating a need for some way to account for the sales that take place during a count. The most accurate way to adjust for sales is to run a sales report for a certain category or department, then go out and count that department entirely, further adjusting for sales since the sales report was run. At the end of this activity you will have a perfect sales adjustment to go along with the on-hand count. The problem with this method is that it’s time consuming and for certain departments it’s going to be impossible to adjust for sales perfectly do to the distribution of merchandise throughout the store, or due to high sales volumes. For instance, doing this for a category like "grocery" would be almost impossible. Plus if doing this for one department is time consuming, repeating this process for the rest of the departments in the store results in an absolutely insane way to go about doing an inventory count. The most common way to adjust for sales is to utilize the half-sales method. In this method a sales tape for all departments is run at the beginning of the count and at the end of the count. The adjustment takes all of the sales from the beginning tape and half of the sales during the count to account for the fact that some of the merchandise sold was counted. Mathematically this total can be achieved by adding the beginning and ending sales figures and dividing the sum in half. The level of error caused by this method is probably minimal for most departments. Although for departments like cigarettes and lottery, the perfect sales adjustment may still be preferred due to the high volume of sales for these departments, not to mention that these departments are usually kept behind the checkout counter, making it easy for the counter to monitor the sales activities during the count. But ultimately the half-sales estimation technique is accepted because it has been proven to be accurate, it is easy to implement into a counting process, and it allows the counter more freedom in determining how to count a store more efficiently.

From this we could judge an estimation technique by determining the resultant drop-off in accuracy caused by its utilization. The half-sales method in most cases won't affect the accuracy of the audit a great deal, but what about techniques employed in the counting process itself. Consider what information gets captured during an inventory count. In a financial inventory a counter needs to capture the department (or category), price, and quantity on all the items in a store. In a scanning inventory, barcodes can provide information like departments and prices, leaving the counter to only worry about the on-hand quantity. Estimation of quantities is as old as inventory counting itself. The idea that counters are going to do whatever it takes to capture an accurate count on frozen food, plumbing parts, boxes of Jell-O, candy bars or other hard to count items is just nonsense. Techniques for estimating quantities can include judging the size of the item relative to the depth of the shelf or bin that they’re contained in, or using the quantity of the box when full or using case counts, this gets done a lot when someone counts candy. But quantities aren’t the only thing that gets guessed at. There are numerous of instances were items are not priced, and counters generally don't do price checks on every single unmarked item. In financial inventories price estimation may be even more prevalent then quantity estimation. Unmarked items present another situation where estimation is used in the name of efficiency. In some instances, it would require too much effort and time to verify the price of all unmarked items in a store, plus given the notion that one can reasonable guess the price of an item based on past counting experience or similar items in the store that are priced, would suggest that this estimation method wouldn't affect accuracy that much. When working for Quantum, they actually instructed us not to use the store’s price scanners for one of our clients, arguing that they would slow us down too much. This was said in spite of the fact that we had never counted for this client before, were not well versed on the client's typical pricing standards, and also the fact that their price scanners were handheld devices that could go anywhere we could go, making it easier to implement price checks into the counting process. For a financial inventory, if the store doesn’t mark the price, they are essentially letting the counter decide how much it’s worth, and the overall accuracy of the audit depends on how well the counter can guess the prices of merchandise. Granted at times these can be educated guesses, but they're still guesses nonetheless. Also with unmarked items, it’s usually not one or two items that are unmarked, it’s a condition found throughout the store, especially in storage areas. The question then becomes how much of the store will end up getting estimated in this manner? Can a counter get to a point where they've estimated too much?

Another criteria by which we can judge estimation techniques, is by looking at the amount of merchandise that gets estimated in a store. It would be reasonable to say that the less stuff that gets estimated in a store, the more accurate the overall audit will be. If someone walked into a store and looked around the sales floor, backrooms, counter area, back cooler, and other areas of the store and then surmised a total of $120,000 (or any other total for that matter) with the proper breakdown for each department, this would probably represent the apex of unauthorized estimating. If a counter flat out estimates a given section of a store by either plugging a number or by pulling a number out of thin air based on the way a section looks, this also would be considered unacceptable. If a counter however guesses the quantity on a partially full box of $.25 pieces of candy this typically is considered okay. It’s easy for an inventory to defend this practice, because again it doesn’t affect the accuracy of the audit that much. The question though is how often does this technique get utilized by an inventory counter during an audit. Even further how many different types of estimation techniques are used and how much estimation is there in an inventory count. Most inventory services wind up doing what might be called scattershot estimation, where an inventory is estimated in small portions, but done multiple times throughout a store, so some stuff in one area gets estimated, some stuff over here gets estimated, some stuff over there gets estimated and the question becomes what is the net result of all this on the overall effect of the audit. At what point does the dose of estimation become toxic? How does one go about criticizing the overall accuracy of an audit without getting into specifics instances that the inventory service may be able to properly defend? Sure "glance" counting a box of candy will not make or break any audit, but when is that ever the issue? If that was the only time in an inventory count whan an item was estimated, then truly it wouldn't matter that much. But how often is this the case? How often does an inventory counter resort to estimation just once in an entire store? Counters are creatures of habit they usually wind up counting the same type of merchandise over and over again, and estimation becomes a habit that gets developed by the need for "pragmatism". Estimating inventory is akin to smoking a cigarette in that the real danger is in the habit, not in the singular act itself. A smart inventory service would seem to want to spread their estimation techniques out thin enough so that they will never be guilty of committing any major act of malfecious, and yet can leave counts that still look good enough to be accepted by the client. This seems to the objective of the inventory service, to produce numbers that are close enough to be considered correct and accurate, it’s not necessary to nail a count on any section, in order to get the client to accept your work. Why kill yourself to produce numbers that are dead on accurate, when you can get by with numbers that are ‘reasonable’?

Of course does accuracy really matter, when determining weather estimation techniques should be used or not? Accuracy is not the only issue in play here. Suppose an inventory crew gets to the end of the count and discovers that one section in the store has been missed. A counter could look at past inventory figures and plug in a number based on that. The figure that gets plugged in theory could be fairly accurate, hell it may even be spot on, the fact of the matter is this type of behavior will still be considered unethical and immoral and the reason for this has nothing to do with the accuracy of the end result. Instead it has to do with the concept of integrity. With integrity, the main concern is the manner in which merchandise is counted. How does the integrity of the counting process get judged?
One possibilty could be in the repeated and consistent accuracy and quality that a certain process brings. A low integrity technique can be highly accurate in a given instance, the question here is how often will a technique produce the same high level of accuracy. The notion that a technique can be successful repeatedly again and again is what’s key here. A counter can take a glimpse of a box of candy and estimate that it has 25 pieces in it, and this count may be accurate, but if this method was used again on other boxes, would those counts be accurate? If this method is used again on the next inventory for various boxes, how many of those counts would be accurate? If this method was used the next day, and each subsequent day in every inventory how often will this method be accurate? Each method or technique will come with a certain degree of error. If I employ a method where I pull the box off the shelf and take the candy out to count it, this method will be accurate but more importantly it'll be accurate far more times when it is used. Now there may exist some people who claim they could look at a box of candy and tell you how many there and do this accurately over and over again. Honestly I wouldn’t deny the existence of such people, being able to do that is a real talent, and there are probably are people who possess this ability. But with the idea of integrity, we can also consider the human factor. Just because counter A can look at a box of candy and accurately guess how much is in there, and can subsequently repeat this ability many times over, doesn’t mean that counter B can. The notion of repeated accuracy should also be applied across different counters, as well as across different inventories. A counting method that has integrity is one that produces accurate and quality results time and time again even with different counters doing the count. Even though we may use accuracy to gauge the integrity of a process, ultimately integrity isn’t about what number you come up with, but how you go about obtaining it. Integrity is about an attitude. It's is about having the desire to do things right, you have to want to get down and dirty to pull the candy out of the box, to open boxes to see if they’re full, to check to make sure you have the right price on an item. Integrity is more about doing what’s right or staying within a code of ethics, it’s a belief that this is the way things ought to be done. In any store there will be area that challenge and expose the integrity of the counter, weather it be the candy aisle, pluming parts in a hardware store, drawers full of undergarments, whatever. Maintaining one’s integrity in areas like these is difficult and hard, and more importantly time-consuming. And for the inventory service time-consuming is synanmous with money-consuming. The standard methods of counting usually encouraged by inventory services aren’t the ones that might be considered the most fundamentally sound or with the lowest degree of error, but rather the methods that encourage counters to be productive and most important profitable. Inventory services want to employ people like counter A, who possess the talent to estimate with great accuracy and precision. Honestly inventory services ought to have a boxed candy test for potential new hires. They can just sit new hires down in front of a few boxes of partially full candy, and have them "determine" how much is in each one. This is the way these companies want their people to count, and obviously the best guessers are the ones they want to put out in the stores. But ultimately no matter how much integrity one talks about or attempts to have in the counting process, estimation will always exist, it’s unavoidable. There is no perfect counter, always counting everything with absolute accuracy and precision, never taking shortcuts. Henry is right when he says that reputable inventory services do permit some estimating, the problem is unreputable services will permit even more.