Direct Materials P2P -- Part 2: Relationship-Intensity
on Jun 3, 2014
In our research on direct materials Procure-to-Pay (P2P) processes, we saw tremendous variation between industries or types of companies in how they execute P2P. One major difference is the 'relationship-intensity;' how much negotiation and dialog is required at each stage of P2P.
Demand-to-Confirm—Demand is fed into MRP1 or project planning systems that calculate the quantities and timing of purchases, resulting in creation of POs. Orders are sent to the supplier who acknowledges and confirms their ability to deliver, or suggest alternatives.
Build-Change-Deliver—The supplier may ship from stock, assemble-/build-to-order, or engineer and build the items, which are delivered to the buyer’s requested location.
Receive-Inspect-Accept—The item is received and accepted or rejected by the buyer. Inspection may be done by the supplier, the buyer, or a third party.
Invoice-Reconcile-Pay—An invoice is issued, matched with what was ordered and received, and then paid.
Relationship-Intensity vs. Automation-Intensity
The nature of direct materials P2P varies tremendously by industry and each company’s way of doing business. One way to view this is by looking at how much buyer-supplier, person-to-person interaction is required at each stage, as shown in Figure 1 - Relationship-Intensity vs. Automation-Intensity Across P2P Stages.
Figure 1 - Relationship-Intensity vs. Automation-Intensity Across P2P Stages
At each stage, the process may involve a lot of dialog/negotiation/collaboration between the buyer and supplier, or may be highly automated with very little human intervention. There is of course a third very common alternative, which is a lot of manual processing of elements that could be automated—in the context of Figure 1, those are simply unrealized automation opportunities, i.e. they fit on the lower part of the diagram.
This notion of which parts of the P2P process are by nature relationship-intensive is one area where we see the differences between industries and companies in sharp relief. Further we see that the stage at which the relationship-intensity is high (i.e. where lots of collaboration/negotiation takes place) also varies from industry to industry (or company type to company type) depending on many factors like lead time, product complexity, amount of custom engineering, newness of product or supplier, and other attributes.
Typically demand is driven by a forecast, fed into an engine of some sort, which calculates what needs to be ordered, how much, and when. Smaller manufacturers often do this in a spreadsheet or even on paper or in someone’s head. Manufacturers above a certain size will typically use an MRP system. However, in large engineer-to-order equipment manufacturing, as well as in construction, they usually do not use an MRP system. Instead they will use the materials management and purchasing modules within a project planning system. These keep track of highly complex BOMs, schedules and interdependencies, lead times, delivery sequences and constraints (like lead time for permits), availability of specialized equipment and skills (e.g. rigger / special moving equipment), fabrication lead times, and more. It then generates POs, linked to the technical specs, drawings, requirements, and project plan.
As shown in Figure 2, the buyer often sends some sort of advanced forecast. This could be informal, like a simple email. However in some industries, such as automotive and high tech, they send periodic rolling forecasts, which also constitute firm orders when the orders fall within the agreed commit window. Thus the forecast may include the material release right in it. On the supplier side, they receive the orders and decide whether they can meet them. For a small shop, this may just be a person who keeps track of all the orders on paper or in a spreadsheet. Larger enterprises may have some sort of software system for managing and allocating production capacity/inventory and order commitments.
In some industries, build-to-stock is required because short lead times are demanded by buyers. An extreme version of that is the ‘proximity warehouses’ or ‘supplier hubs’ with consigned inventory located next to the buyers plant, with replenishment often managed by the supplier (aka VMI or SMI2 ). This model is increasingly common in high tech, automotive, and some industrial manufacturing.
Delivery Schedule Negotiations in Fabless Semiconductor Manufacturing
Generally the fabless semiconductor manufacturer will send a rolling forecast to their outsourced wafer fabricator on a periodic basis (often monthly). Then there is a flurry of negotiation about the dates of the wafer starts4 in the coming month. This includes ‘horse trading’ to push in or pull out the wafer start dates, trading off the projected demands of the buyer’s different product lines and the available capacity at the wafer fab. Once the plan is frozen, the creation and confirmation of the PO is often highly automated, since all of the negotiation and agreement has already been done.
Negotiating the Delivery Schedule
In other industries, there is a fair amount of negotiation about the schedule or even the specs between the time the PO is sent by the buyer and confirmed by the seller. In some industries, a lot of the schedule negotiation happens even before an order is placed, such as in semiconductor manufacturing where the wafer fabrication3 phase can have a 14-18 week lead time. This requires a lot of upfront collaboration on the demand plan and build schedule with the wafer fab (see sidebar: Delivery Schedule Negotiations in Fabless Semiconductor Manufacturing).
The next step after order confirmation depends a lot on whether the manufacturing model is ship-from-stock, assemble/build-to-order, or engineer-to-order. These different models, as well as the differences in order lead times, have a big influence on the degree and types of change orders that occur once the order is placed.
Figure 3 – Typical Build-Change-Deliver Processes
Story of a Pump—Engineer-to-Order P2P Dialogs
As an example of the back and forth dialog that happens in large complex engineer-to-order procurement, consider a large pump used in an oil refinery being built. These pumps are highly-engineered, custom machines that can require hundreds or thousands of horsepower, withstanding high pressures, temperatures, and harsh chemicals. Furthermore, they need to fit precisely into the overall layout of the plant. There are hundreds of parameters to consider including the size and spec of the bearings, flanges, nozzles, casing, mounting, pressure requirements, ability to move solids, flow volumes, all manner of physical dimensions, performance specs, and much more. The buyer sends requirements to the pump manufacturer, who may respond with some alternate suggestions. Ultimately they agree on a spec and the supplier sends detailed technical install instructions, precise dimensions, how it will connect to other equipment and pipes—specified in 3D CAD drawings. As the project unfolds in the field, the engineering firm (buyer) may see that adjustments are needed, to form factor, performance, or other specs, and a renegotiation occurs around delivery dates, price, and spec. In some cases, in spite of best efforts, adjustments may even have to be made in the field after the pump has been delivered, or it might have to be returned to the factory for modifications.
Short lead time items are often standardized (rather than custom-built) items, though they can be assembled-to-order. Because of this and the short window between order and shipment, change management is less of a relationship-intensive issue for these and is largely about shuffling quantities and delivery dates, or occasionally cancellations, but rarely about changes to the design and specification of the item. At the other extreme, for large engineer-to-order projects, such as the construction of a plant or commercial building or large complex machinery, there are often a lot of back-and-forth changes to the spec that happen after the order is placed, with many linkages to technical drawings and specifications (see “Story of a Pump—Engineer-to-Order P2P Dialogs” sidebar). The process of dealing with changes and adjustments often continues right up until the machine has been installed and is up and running and stable.
Another example where post-PO change management is important is in aerospace manufacturing or other industries that have very long lead time custom parts, such as forgings that can have lead times of up to a year. Since these changes can be costly, often the manufacturer measures things like the sources and reasons for defers, changes, and cancellations, to try and continually improve their own processes and their suppliers’ performance.
Auto-triggering Orders Based on Shipments
Another example of how each industry is different goes back to the fabless semiconductor manufacturer. As mentioned, their procurement consists of buying the output of each stage of production and sending it into the next stage of production (wafer fab → assembly → test → delivery). As such, the shipment alert sent by the supplier from one stage of production triggers an automatic release against the PO for the next stage. This automation can be taken a step further—the invoice from the supplier may be sent simultaneously with the shipment alert to the buyer, automatically matched against the inventory receipt by the buyer’s system, based on items received by the next-stage supplier/outsourced manufacturer, and then auto-paid. This requires a fair amount of system sophistication to integrate with the various suppliers’ systems and map the various internal and external material tracking codes.
Here we have explored some of the variations in the P2P processes between different industries for the first two stages of the P2P process (1. Demand-to-Confirm and 2. Build-Change-Deliver). In the next article of this P2P series, we will look at the final two stages of the P2P process (3. Receive-Inspect-Accept and 4. Invoice-Reconcile-Pay).
3 A Wafer fab takes thin slices of silicon (the wafer) and deposits integrated circuits on it, and then slices it into individual ‘dies’ (i.e. guts of the chip). This process is long and time consuming, hence the long lead times. These die are later put into various packages to make the actual chip. -- Return to article text above