Gaea Global Technologies

Service parts business has unique requirements, so much so that I thought it worthwhile  to devote a couple of blogs to this topic. Gaea Global is currently implementing WMS for one of the largest automobile companies in the world, in their service parts business and we have seen pretty rigorous requirements for WMS to manage service parts fulfillment.  We have helped TTI Floorcare (one of the largest floor care companies in the world, maker of the Dirt Devil and Hoover brands) reconfigure their warehouse in order to better handle the service parts side of their business.  Therefore, we have a fair bit of experience in this. First a quick recap of what I have observed as being some of the unique requirements of service parts business:

1. Large number of service parts: Obviously J. If you are one of those guys driving around a 1996 clunker for which you also happened to purchase an extended warranty, the auto manufacturer better have the service parts needed for your car. So the end result is that the service parts warehouse must stock a large number of parts. We are talking about 100,000+ parts here.
2. Low Demand: Most of these 100,000+ parts are “Slow Moving” parts. This one is obvious too. The demand for service parts declines a few years after the product matures.
3. Single Line orders: Service parts business get a lot of volume due to 1 or 2 part orders that are needed for one-off repairs. Due to the small order cube, the orders are mostly parcel shipments.
4. Prevalence of “hot” orders:  Service parts needs are often needed urgently and I have seen more than my fair share of “hot” orders flow through the warehouse on a given day. Sometimes these orders are overnight shipments because the manufacturer has a guaranteed SLA for servicing the part requirement.
5. Very High cost of shipping the wrong product:  Shipment accuracy is even more important for a service parts business. Further complicating the matter is that many of the parts look alike and often the part description is also very similar. Often there is a customer who is willing to pay for overnight shipment to get a service part on an urgent basis and getting a wrong part will not be appreciated.  Correcting such a mistake after the fact turns out to be very costly.

Stocking thousands of parts is no mean feat.  The service parts warehouse needs thousands of bin locations to store parts. The parts could vary from a few inches in length to a few feet in length. You need a large physical warehouse area to house these bins. A large area means a longer pick path. The challenge is to limit travel time to maximize operator productivity and at the same time ensure that a “hot order” gets shipped ASAP to minimize premium shipment service. Since accuracy is important, you need the operator to scan and maybe even rescan a part to ensure that the part you are shipping is the right one.  Clearly reliance on Auto-ID is important. Since parcel shipment is important in service parts fulfillment, integration with one of the parcel manifest software like Kewill, Connectship, Precision, etc. becomes important.

In my next post, I will review some of the features in Oracle WMS that addresses the unique needs of service parts fulfillment operations.

Have been meaning to write this one for a while, but kept putting it off. Recently though, I have been hearing this requirement quite a lot (no pun intended :-) ) – the ability to pick a lot that is convenient to the picker as opposed to the actual allocated lot. First, some background: Lot allocation based on lot attributes is a powerful feature in WMS. A common use case is a “strict” FIFO policy that some warehouses use to handle perishable products. The way it works currently, the WMS rules engine checks the lot expiry date and then directs the picker to clean out a lot that is the oldest in the warehouse. Obviously a very useful feature provided you are dealing with perishable products where strict lot control is essential to minimize product obsolescence.

The problem arises when you have a lot controlled product primarily for traceability but do not really need lot control for shelf life control. In other words, you want the system to recommend lots to pick based on loose FIFO but not enforce them during picking. Unfortunately, Oracle WMS gives you an “all or nothing” solution. If you decide to make your products lot controlled, you have no option but to enforce lot allocation and only pick the lots that WMS tells you to. This is often a cause for grief and heart burn at many warehouses because this could drastically impact picking.

Imagine a scenario where each pallet is lot controlled and in a 2-deep warehouse slot. WMS directs the operator to pick a pallet that is currently sitting at the back of the rack. Obviously this will have implications for warehouse productivity since the operators will spend considerable time sorting through cases or pallets to pick the right lot. If the warehouse operators want to beat the system, this may also adversely affect the inventory accuracy and lot traceability i.e. the operators confirm the lots recommended by WMS but in reality, the operators physically pick the lot that is most convenient to pick. Either way this is a grave situation.

Here are some ideas and thoughts around this situation:

Do you really need lot control?

The essential point for having lot-control is traceability and shelf life control. If your product is not perishable or has a defined shelf life or if you do not require lot attributes or full traceability of products from manufacturing to shipment and don’t do product recalls, you do not need lot control.  Consuming products in FIFO mode because it’s a good business practice is not a good reason for Lot control. You can do a loose FIFO without having full lot control. You can use the receipt date and allocate down to the locator level using the rules engine without using lots.

Should you commingle lots?

Unless business reasons warrant it i.e. lack of shelf space, avoid lot commingling. If you have to commingle, at least commingle lots that have expiry dates within a tight cluster say within 30-days of each other. This way you can have a good trade-off between consuming products in FIFO mode, tighter space usage without the need to rummage through multiple cases or pallets to pick the right lot.

Is lot substitution the answer?

Maybe.  WMS allows lot substitution in 11.5.10. Oracle in its infinite wisdom has chosen not to release the patch for R12. Lot substitution allows an operator to pick a different lot than what the system is recommending. It seems like a good solution but unfortunately it does not solve the problem.  The reason is that the substituted lot must be freely available and not allocated for some other task. This is not a realistic assumption in a busy warehouse.

Is customization the answer?

What you really need is lot swapping -  i.e.  the ability to freely pick another lot regardless of its allocation status. Unfortunately you really need a customization to achieve it.  Extension rather – we don’t do customizations, we do extensions.  Like I said earlier, we have heard this from a few customers and are actually busy at work prototyping this and should be releasing this fairly soon. Stay tuned!

Daniel Soosai of Gaea will be co-presenting with Sree Kakanate of Cisco at the Northern California Oracle Application Users Group 2010 Training Days session on “Collaborative Distribution Planning: An Innovative Framework” at the Santa Clara Convention Center on January 19, 2010.

This presentation highlights innovative channel management strategies and distribution planning tools that improve lead time predictability and forecast accuracy leading to enhanced service levels while reducing the working capital.

For more information on the NorCal OAUG Training Days event, Northern California OAUG please see their website – http://www.norcaloaug.com/