3 phase approach to solid inventory record system o Design of the inventory record system and creation of the internal capabilities to put into place and keep it operating o Straightforward development of initial inventory balances o Ongoing operation and control of the inv record system, where responsibilities are defined and sources of errors can be located and eliminated. • measuring inventory accuracy needs to take into account tolerance. Might miss exact number on most of the items but might fall within the % variance of each item and then no problems. Thus overall inventory accuracy = (total accurate records / total records checked) *100 = % of accuracy • dollarizing value of firms inventory may be accurate to penny but that accuracy could hide appalling level of part specific inaccuracy. • tolerances may be set based on a products: o usage o dollar value o lead time o level in the bill of material o criticality o method of handling • most companies multiply usage by value in order to compute ‘dollars spent during the year for each part’. That typically will follow 80/20 rule. 80% of dollars spent per year will be on 20% of parts. Chart would be: ▪ Cost Class Parts Puchased % $ Spent % Tolerance % ▪ A 20 80 +/-0 ▪ B 30 15 +/-2 ▪ C 50 5 +/-5 • This however ignores the velocity of a part as well as the lead time • never should a tolerance exceed 5% • target should be to get to 95% inventory record accuracy with +/-5 percent tolerance on every item (would rank in the top 10% - 20% of all companies). • Three phase approach to improving and maintaining inventory record accuracy o Design and Preparation ▪ Design inventorying process • Stock versus work-in-process • Layout • Transactions ▪ Measure starting point ▪ Provide tools • people • systems • physical ▪ finalize policies and procedures ▪ train personnel ▪ education ▪ assign responsibility ▪ implement new policies and procedures o develop initial balances ▪ develop initial balances o cycle counting ▪ control group cycle counting ▪ ongoing cycle counting program • select method • assign responsibilities • correct error causes • Physical aspects of control o Standardize receiving, cycle counting, issuing of inventory o Limiting access to the inventory o Appropriate storage facility to store inventory o Appropriate containers, bins, racks o Correct equipment to handle inventory o Correct equipment for counting/tracking inventory (bar codes, weigh scales) • Tactics for controlling inventory o Made-to-stock, make-to-order (dell example – tactic depends on what you do – make to stock have to carry more finished inventory – assemble to order – more components but not finished inventory o Forecasting and demand planning with sales and ops planning – use forecasts to help determine optimal inventory to produce and hold o Inventory levels o Safety stock levels – formula is safety stock = stand deviation (determined from actual demand) * service factor. Service factor is 3.72=99.99%, 2.58=99.5%, 2.33=99%, 2.05=98%, 1.88=97%, 1.75=96%, 1.64=95% o Customer service levels o Supplier managed inventory o Customer managed inventory o Ownership of raw materials • Tactics must be determined first before aspects • Available inventory is on hand – allocated. Allocated still needs to be counted as available in the case that a cycle count for a location that has ordered product hasn’t been picked yet • Employees must be trained and someone must be responsible for the accuracy • Initial sampling to determine current numbers (look from both record books and actual locations for inaccuracies • Limited access • Problem with multiple storage locations for the same sku – best way to handle is have system that allows for piece count by multiple location which would show total of the sku and also the breakdown by location it resides in o Best to set up picking from location with smallest quantities to eliminate warehouse full of almost empty location • Randomly locating inventory provides 40% greater space utilization than does the assigning of permanent locations • Location convention – stock room, aisle, rack, shelf, opening o Y-A-03-B2 would be aisle A, 03 rack, B shelf, and 2nd opening on shelf. o Even / odd on different sides of aisle (picking efficiency since going down aisle only once o Start with A at the bottom of rack because expansion is never downward o Always build empty locations into naming logic • Reporting for maintaining accurate records o Stock status by part number o Stock status by location o Inventory transaction history by part number o Open order listing • Second phase is to calibrate the quantities (establish accurate initial balances). Three approaches to calibrating initial balances: o Taking physical inventories o Cycle counting o Systematic physical counting • Calibration process should be two things (accurate and quick). 1-3 weeks ideal, never longer than 4 weeks • Problems with physical inventory – problems identifying parts, correctly recording unit of measure, and accurately counting items themselves • To improve accurate physical inventory o Only people knowledgeable about parts being counted should be assigned o Everything should be counted twice and the time to count twice should be scheduled o Everyone involved with the physical inventory must receive instruction • How to maintain accuracy o Cycle counting o Two characteristics of a sample determine how well that sample represents the entire population. Size of sample relative to population. Second is stability of the sample, if as sample is increased, results increase at same result, sample is stable. o Errors found in cycle counting can be inferred to be true of rest of the population (i.e. sku is short in sealed box, may mean that other parts from that supplier are short. o Cycle counting highest value is in finding errors, and from those errors, their causes. • Method used to select a sample for cycle counting o Control group cycle count – counting same parts in same locations repeadetly within a short time frame. Typically only done in beginning to identify problems. Not ongoing type of cycle count. 5 steps: ▪ Identify control group – about 100 different locations recommended (should be cross section of inventory – big/small, fast/slow, expensive/inexpensive) ▪ Count control group ▪ Set inventory records equal to counts ▪ After 3-5 days, recount same locations ▪ Compare recounts to inventory records o Random sample cycle count ▪ Two types: • Constant population technique – items randomly selected this time are put back in as possible selections for the next time • Diminishing population technique- items not included until random sample has hit all items, then everything dumped back in o Abc cycle count o Process control cycle count ▪ Looking for incorrect parts ▪ Validating counts for low volume locations ▪ High volume locations only estimating that its accurate (if suppose to be 2914 and it looks like you have about 3000, then you move on without counting ▪ Interesting usage because it is biased like other options, but this one is biased towards errors (spending most time on suspicious locations) ▪ High volume locations end up getting counted when they get lower based on usage ▪ Other choices • Low volume counts • Upon receipt or pick (countback)

Big thing is don’t cycle count just to get accurate, use it as starting point to begin detective work into figuring out what caused error in first place

Problem with scale counts for huge quantity item is if you got 9900 instead of 10000 obviously no problem at beginning. But even if every transaction perfect, once you got down to 300 you would only actually have 200 and now it’s a MISS. Way to counteract this is to acknowledge item was never physically counted at receipt that way when you MISS you acknowledge it but understand probably cause of error.