Why Procurement Rocks!

My first order of business (after my coffee) every day is to check the blogs I subscribe to. One that I recently ran across is called The Vendor Chronicles, and a recent post there caught my attention. Check out Darcy’s take on Why Procurement Rocks.

I still need to look at all her earlier posts but, so far, this one’s worth a subscription – to go with your morning coffee.

Advertisements

Contracts, the law, and vendor performance

Gillian Hatfield makes some interesting points over on his blog, Commitment Matters, about the relationship between contracts and the law. My take on his post is that the law is important but, for our purposes, it’s at least as important that the relationship between the firms is well thought out and documented.

It’s been my experience that far too many companies allow the stakeholders to negotiate and approve contracts, with review (of the legal terms and conditions) by the lawyers. When purchasing tries to get involved the response is that “We’ve done business with these guys for years and they’ve always been great” or “We don’t have time to go through all those negotiations – need that (fill in the blank) now.”

It’s always interesting that trying to pin down “great” with a supplier, through the use of performance measures inserted into a contract, leads the supplier to try and red line those clauses relating to performance – and to remedies associated with non-performance. Hey, if the supplier is telling you they’ll perform at a certain level what’s the problem with putting it on paper? But, most times, the suppliers fight these clauses.

In many cases this negotiation exercise leads to a better understanding of what both companies REALLY intend to do once they’ve started doing business. And that understanding may not, unfortunately, be what the stakeholders had in mind.

Time is the enemy for putting good agreements into place. Give your purchasing department the time to research the supplier and negotiate the deal. It will benefit everyone in the end through better supplier performance, shorter lead times, more certain deliveries, services that truly meet your needs, and less vendor churn.

Forecasting – get out the crystal ball

My first job was one I really didn’t want. I’d landed a job in purchasing at an aircraft parts distributor and was in training with an old timer who was getting ready to retire. The 30 day training program consisted of him telling me that the job was really easy – just get your lead times and forecasts right and everything else takes care of itself. Little did I know.

We’ll leave lead time for another post and take a look at forecasting today. On paper it really does look fairly straight forward, assuming you remember what you learned in that statistics class. Just get some good history of product sales (or issues), crunch the numbers, and let the system drive itself! If it were only that easy we wouldn’t see so many different forecasting models out there – and so many times when otherwise great ERP systems fail us. As so often happens in the world, we lose sight of the basics. Let’s take a look.

The first step, of course, is to admit that our memories are really not that good. Even if you could accurately gauge the movement of a few products you simply can’t keep track of eight or ten or fifty thousand SKUs. The salesman who tells you “we sell those all the time” is probably recalling the one or two instances when you were out of stock and he got burned by losing a sale. It’s human nature. So, the first item of business is to accumulate accurate data on product movement.

Ideally you’d have this for a minimum of two years, by month. You’d want the quantities returned deducted from the quantities issued, and you’d want to show the numbers based on the time they were required – not at the time the product was actually shipped or invoiced. There’s an ongoing controversy about what to do with requirements booked – and then cancelled – when you’re out of stock. My inclination to count them even if they’re cancelled because the probability is that they would have been shipped if there had been inventory available. In any case, once you come up with some history you can start the process.

At this point it does become a numbers game. Drag out the statistics book and look up the formulas for calculating a least squares line. I won’t go into the actual formula because it’s readily available. The idea, though, is to “fit” a straight line into the sales quantities (points on a chart) across a time span. Doing this will give you a straight line that trends up or down based on the history. Taking it to the next step you can calculate what the projected quantities are for the months in the future.

If you’re not dealing with seasonality, and assuming that nothing changes (at least too much) in the market, those forecasts will be reasonably good. But, sometimes reasonably good isn’t good enough. We’re always trying to make the forecast better – or at least good enough that you can use it for most of the products you’re controlling while you spend your time worrying about the oddballs out there.

As good as any forecast can be the true professional knows when to throw it out and go with the numbers provided by the production or sales guys, market intelligence, and the gut. Keep the phone working and reach out to your own team, your supplies, and your contacts in the market to know what’s going on outside your four walls.

Your ERP system doesn’t have a forecasting module? I’ve worked in situations where I was told forecasting wasn’t needed because “we buy to requirements from the production department.” My answer to that was to ask why they carried inventory in the first place. The reply was always to the effect that any inventory needs to be forecast. It’s not too difficult and, as I’ve said, forecasting by the seat of your pants isn’t a good alternative. The good news is that inexpensive software can get you where you need to be. I’ve done it several times and look forward to doing it again.

Lead time

Lead time’s pretty straightforward. If I want to know when to order something, I’ll need to have a pretty good idea of how long it takes to get that something. So, it’s the number of days (or weeks, or months) that will elapse between the time I realize I need it, and the time ┬áit’s received.

Pretty simple? It really is, until it isn’t. The vendor tells me his lead time is 4 weeks. To that we’ll need to add the time it takes for the PO to be approved (assuming the order needs to go through an approval process). So, now we have a number that can be used to help calculate the re-order point.

How many times can you really trust the supplier? Are they giving you the best case scenario? If their real lead time is actually six weeks you’ll be out of stock two weeks before your shipment arrives. On the other hand, the supplier might want to give himself time – just in case. You order based on the expectation that delivery will be in a month and the material shows up in a week. Now you have 21 days of excess inventory on your hands. This isn’t a big deal if you regularly use the product, or if it’s inexpensive, but can be a problem when the unit cost goes up.

A world class system will monitor several things – and report on them in different ways. Here are the numbers that should be tracked – on an item, buyer, and supplier basis:

  1. Requested lead time. This is the date the buyer asks for delivery.
  2. Quoted (promised) lead time. This is the date when the supplier says you’ll have your order.
  3. Actual lead time. This is the time it has historically taken for the product to show up.

Putting them together you can compare:

  1. Requested vs quoted: If you consistently ask for product faster (or later) than your supplier can deliver it indicates a bad planning system.
  2. Requested vs actual: Again, continually asking for early delivery shows some serious wishful thinking on the basis of your planning or purchasing departments!
  3. Quoted vs actual: When a supplier can’t give you a good idea of how long it will take to get our order filled you need to do some serious work with your supplier.

I worked with a supplier who was favored by management because their rep visited us on a regular basis. He was a good guy and everyone liked him. His prices were slightly better than those from his competition. Unfortunately, the only thing he could deliver on time were the doughnuts he brought in. An analysis of his performance revealed that we’d do better buying from the higher priced supplier because we experienced fewer stockouts, less excess inventory, and much smoother operations.

A final word on lead times. It’s not enough to compile lead time history. You’ll need to watch the consistency of any variation. A large coefficient of variation on historical lead time indicates the need for larger safety stock (the “buffer” inventory you’ll carry to cover unexpected demands or delivery variation).

A word about buying from manufacturers. Manufacturers can make an inventory planner absolutely crazy because the manufacturer doesn’t put the product into production until they’ve reached the point where they can economically produce a batch of the product. If you’re lucky enough to place your order when they have inventory to ship delivery is simply transportation time. If they’ve just run out of stock and are building backorders in preparation of a production run you may be waiting months! This is where you’ll need to stay in touch with the manufacturer to balance everyone’s needs, go through a stocking distributor (and hope they know what they’re doing), or be content with carrying a lot of inventory.

Watch the numbers, stay in touch with the manufacturers and distributors, now your own market (stay in touch with your sales people or engineers to understand upcoming jobs), and keep the lead time numbers – and reorder point calculation up to date.

My first purchasing mentor told me it was all really easy. Just make sure your forecast and lead time numbers are right and you’ll be OK. If only it really were that simple.

Does your purchasing / inventory planning give you these numbers? I may be able to help.

ABCs (and XYZs) of Inventory – Continued

My first two posts talked about ABC classification of inventory. These strategies help you understand where your effort (either in forecasting and negotiating, or in cycle counting) should be. The second traditional measure is XYZ.

Imagine that you had an item where you’ve used 10 per month for the past 24 months. You’d expect to use 10 each month for the foreseeable future. Now imagine another part, also with an average consumption of 10/month, but where the monthly history is just all over the place: January 0, February 25, March 0, April 0, May 50, June 5, July, 1, and so on. How many do you expect you’ll need in stock next month? You have an average, but using only an average will result in months where demand greatly exceeds supply.

By running an XYZ analysis you’ll know which of your thousands of stock items fall into the X class (very smooth and predictable consumption), Y (somewhat “lumpy” history), and Z (who knows WHAT”S going on with it!). The X items can safely be kept at a minimum level, this frees up cash to invest in the Y and (more so) in the Z materials.

I won’t go into the formula for calculating XYZ numbers but it’s easily derived using the mean, standard deviation, and coefficient of variation of historical demand. You should have 12 months of consumption history (24 would be better) to get valid numbers. You’ll also want to include returns to stock as negative numbers.

By using the financial ABC calculation (my first ABC example) and the XYZ numbers, you can set up a matrix to specify what your “model” stock levels should be.

AX: High annual dollar consumption and very predictable consumption. You’ll want to carry the lowest number of units in this category. You might specify that the minimum should be zero, and the maximum should be only 1 month’s supply.

CZ: Low annual dollar consumption and very erratic consumption. You’ve conserved cash on the AX items so you’ll be able to invest more heavily in this class of inventory. CZ items might have a specified low of six months with a high of a year’s worth of inventory.

AZ: Would be high dollar consumption and very erratic demand. This is the worst of all worlds and requires the most inventory management. In this group you’ll want to negotiate for consignment stock or some other way of pushing the inventory risk off onto the supplier. If you’re in a manufacturing setting these may be the components that are (at least to some extent) purchased only as required.

CX: Low value turnover and highly reliable forecasts. In this case the concept of economic order quantity comes into play. You don’t need to carry a large inventory but it might be more economical, given the fixed costs of creating a purchase order, paying an invoice, and shelving the receipts, to simply place large orders when inventory gets low. Running out of stock is not a big risk, nor is carrying “cash on the shelf” – the cost of placing the order becomes a more important consideration.

Once you’ve determined the ABC and XYZ classifications, you can begin analyzing your inventory quality – a macro measurement of how your efforts are paying off.

Parts of the XYZ calculation will also be used when calculating safety stock (a component of your re-order point).

Questions or suggestions? Please leave a comment.

Ned

ABC analysis – another way of looking at it

In an earlier post I talked about setting up an ABC analysis. This traditional model stratifies your inventory based on the extended “cost of sales” of all the materials. The idea behind this model is that the items with the highest cost of goods sold are those you want to most keep track of. Unfortunately you may see items with a very high individual cost and very low “sales” classified as A items – with the really inexpensive (and most used) items classed as B or C. Counting these high usage items every six months or every year opens you up to potential stockouts.

There may be a better way. If we consider that the items being most used – those with the largest number of individual “consumptions” – we’re looking at those with the highest number of potential stockouts.

The process is the same, just use different numbers. So, instead of calculating the total item sales per year, total the actual number of sales of each item. That way, the paint brush that costs a dollar – but is used 500 times a year (500 x $1.00 =$500.00) becomes an A item while the $5,000.00 pump that’s only been used once may be a C item. In the traditional model you’d count that pump four times a year – “Yep, it’s still sitting there on the shelf, just like it was three months ago.” The paint brushes would only be counted annually. During that year between counts you’ve actually sold the paint brush 500 times. That’s 500 opportunities for being out of stock.

If the idea of cycle counts is to reduce the number of stockouts this model, of using the number of actual sales, rather than total dollars sold, might be your best bet.

Thus, the ABC calculation of inventory can be used as a basis for looking at inventory quality (more on that later) and ABC of consumption is used as a trigger for cycle counts.

ABCs (and XYZs) of Inventory

If you carry inventory you have money going to waste. Let’s face it, inventory is there to be used or sold and, if it’s still on the shelf it’s doesn’t fall into either of those categories. Your only hope is that it WILL be used – soon. Meanwhile, you need to keep track of it. The more material on the shelf the more difficult it is to manage. So, how to analyze it all?

The good news is that there is a way. The bad news is that many companies don’t use the tools that have been around for years – ABC and XYZ analysis. Let’s take a look at ABC first.

This idea goes back to the old 80/20 rule. 80% of your inventory dollars are tied up in just 20% of the product. If you have limited time that’s the part of your inventory you’ll want to concentrate on. So, every month you run a report showing, by item, the quantity on hand times the inventory cost. Next step is to rank them, highest extended dollar amount to lowest. You’ll end up with a spreadsheet showing all your inventory. Now, start running a cumulative total on your spreadsheet. Total for the first line on line 1, Total for the fItrst and second lines on line 2 and so on. The dollar amount associated with the last item should be the total dollar value of your inventory.

Nest, calculate the percentage of that running total to the total amount. As you run down the list you’ll see that the percentage is getting bigger until, at the last line, it’s 100%. The “A” items will be those up to and including that 80% amount. The “B” items might be those up to the 95% figure, and the “C” items are those above 95% up to 100%.

Sure enough, you’ll see fewer A items than B items, and far fewer than C items. Those As and Bs are where you spend your time and energy. You you watch how many are ordered more closely, you work on your forecasts, watch the lead times, and negotiate harder with the vendors on those.

ABC analysis should be done monthly because the numbers will change. But, doing it on a monthly basis ensures that you’re on top of your investment.

Classical ABC analysis uses this stratification as the basis for cycle counting. You’ll notice that I did not include that function as an outcome of this process. That’s another ABC analysis altogether, and we’ll get to that!

Next up – another way of looking at the ABCs.