Prediction and Execution
Yesterday we looked at the fine strands of supply of a single product. In our example, we looked at the combination of fulfillment time and transit time to choose the vendors that served our distribution centers. A few months have passed, and we have been hearing about some stocking problems and late deliveries at some of the distribution centers.
Luckily for us, our company tracks inbound supply performance, recording the dates of PO release, the date the vendor shipped the order, and the date the PO arrived at our DC. We go back through three months of data and examine the actual performance to learn that actual fulfillment times are consistent with what the vendors told us. Phoenix is still taking five days, and Orlando is still taking only one day.
There were a few times when Phoenix did not have stock to fill the order, and we used those occasions as opportunities to throw the vendor in Chicago a bone to see what their performance was. We did the same with an order to Orlando when neither Phoenix nor Chicago had the product. One time when Orlando was down because of a storm, we shipped a Memphis order out of Phoenix. Our contracts with the vendors stipulate that if they know they can’t fulfill the order in the contracted fulfillment time, they have to reject the PO so we can source from an alternative supplier. It appears that our two primary vendors are following the rules.
While vendor fulfillment performance lives up to expectations, transit times are a different story. The carrier that Orlando (2) uses to ship to Carlyle (C) consistently takes four days to deliver, not three. Our Phoenix vendor’s carrier takes an extra day to reach Memphis. Both Chicago and Orlando take an extra day to reach Stockton.
When the primary vendors can’t fulfill our orders, our lead times stretch out — sometimes significantly — because of a combination of fulfillment and transit times. When we have to fulfill from Phoenix, the lead times double and triple. While we don’t ship out of Phoenix to Memphis or Carlyle very often, we run the risk of running out of stock before a Phoenix shipment arrives. We can increase our safety stock to ensure that we stay in stock, or work with the Orlando vendor to ensure that they always accept the orders. Orlando’s performance is very good — rapid and consistent. The only time they have not shipped was when they had to shut down for a hurricane. However, we still should look at our safety stock levels and see what the risk is.
We first look at the safety stock level for our Orlando vendor. Because of the short lead time and rock-steady vendor performance, we carry three units of safety stock. Demand variance drives the need for safety stock for Memphis when we source out of Orlando. The short lead time that is less than weekly order pattern requires three units of additional stock.
Things get a little dicier when we consider shipping out of Phoenix. The lead time grows longer — triple what Orlando provides. We know that the carrier from Phoenix does not provide consistent service. About 15 percent of the time the shipments are a day later than promised, so we have more variability to lead time performance. Serving Memphis from Phoenix increases the safety stock requirement to 18 units of inventory with the same demand.
Fulfilling Memphis out of Phoenix is not a good idea. We need to make sure that the Orlando vendor stays consistent with their great service. We should consider using the Chicago vendor as our fill in for those few times that Orlando can’t ship.
Looking at the sourcing options for Memphis inspires another line of thinking. How much safety stock do we need if we ship to Stockton from Orlando? It did not make sense when we made our original decision about vendors. Now we know something new that could change our decision.
The demand out of Stockton is lower than Memphis. With the lower demand, there is more variability. The demand variability increases safety stock. Worse is the inbound variability. Again, the carrier that the Phoenix vendor uses does not have stellar performance, arriving a day late about 15 percent of the time. The effect is more safety stock — 14 units, to be exact. Shipping to Stockton out of Orlando is eight days. We don’t know how consistent the carrier is to Stockton, but we know that to Memphis and Carlyle the carrier does not miss a beat. With a lead time of eight days and rock-steady vendor performance, we see almost no variation in our inbound performance, so our calculated safety stock is 10 units, about four units less than what we need today in Stockton.
Should we change the source point for Stockton from Phoenix to Orlando? There are other questions for us to ask; we need to find out whether the Orlando vendor can support the additional volume. If they can’t, their failure to perform might affect the overall system safety stock.
Few procurement managers or merchandise buyers look at safety stock as a deciding factor for sourcing. They almost always make the decision based on cost, and then consider other “favors” that the vendors offer. This myopic decision- making seldom shows up in any of the measures of the buyer’s performance; the procurement manager is judged by the gross margin of the deal, not the inventory or working capital impacts. Tell me how you are going to evaluate me and I will tell you how I will behave. Perhaps the required safety stock should be one of the measurements of the deal.
Articles in This Series
Call Us! 877-674-7495 info@dksco1.com