Pull Vs Push: The Lean-ERP Parado

Japanese production management (e.g. Schonberger, 1982; Schonberger, 2007) and lean production (e.g. Holweg, 2007; Krafcik, 1988; Womack et al., 1990); as well as materi- al requirements planning (MRP) and enterprise resource planning (ERP) systems (e.g. Browne et al., 1988; Orlicky, 1973; Ptak, 2004); are recurrent themes within the field of operations management, particularly when we consider the options for achieving com- petitive advantage in modern manufacturing. There is no doubt that lean production has been shown to lead to performance improvements (e.g. Womack and Jones, 1996; Krafcik, 1988; Shah and Ward, 2003; Sugimori et al., 1977; Womack et al., 1990). More recently, the application of ERP systems has also been shown to effectively im- prove the performance of manufacturing companies (e.g. Hitt et al., 2002; Laukkanen et

al., 2007; Murphy and Simon, 2002; Shang and Seddon, 2000; Tsai et al., 2007). Alt-

hough the application of lean and ERP are consistently rated as the main contributors to competitive advantage in manufacturing operations, there has been a recurring debate as to whether lean and ERP are compatible, or whether they are contradictory in nature. Whereas in the past, information technology such as ERP has been regarded as a source of waste by lean purists, in the current climate, the majority of manufacturers are using ERP systems to plan manufacturing operations whilst also developing a desire to realise the benefits associated with lean production. ERP systems have become a requirement for modern manufacturers, whose customers demand an ever-increasing portfolio of products, resulting in 100s if not 1000s of stock keeping units (SKUs). Managing such a wide range of parts is not a simple task, hence the growing number of ERP systems available today. In order to manage such an array of products also makes the elimina- tion of non-value added activity even more appealing to producers, hence the big ques- tion, ERP, lean, or both?

A common argument arising between lean production and ERP systems is that of pull vs. push. Benton and Shin (1998) suggest that there is a common agreement among re- searchers that a lean, Kanban controlled production system functions as a pull system, whereas those systems using MRP-logic in an ERP system are predominantly push. We can suggest that it is in fact the MRP-logic that is the source of such a paradox. For ex- ample, Rother and Shook (2003) suggest that to qualify as pull, parts must not be pro- duced or conveyed when there is no Kanban, and the quantity produced must be the same as specified on the Kanban. They suggest that the MRP system should be turned off to realise a future-state value stream based on Kanban and pull production (Rother and Shook, 2003 p.78).

When defined in terms of information flow, in a pull system, the physical flow of mate- rials is triggered by the local demand from the subsequent customer, often via Kanban cards. On the contrary, a push system uses global and centralized information stored within the central ERP system in order to drive all production stages (Olhager and Östland, 1990). This leads to the next contrast between lean and ERP.

Where lean strives for decentralized control of production through empowered workers, ERP remains a centralized planning and control database. Stadtler (2005) suggests that ERP systems are incapable of performing real time control of production operations at

the shopfloor. Rother and Shook (2003) also suggest that for lean production, a produc- er should get rid of those elements of an MRP system that try to schedule the different areas of a plant. A further contrast between the two approaches is that of the time- phased vs. rate-based decision (Alfnes, 2005). With lean, the aim is to achieve a level schedule of mixed-model production, synchronized with the rate of customer demand (takt-time). With ERP, the system often calculates an ‘economic batch quantity’ which is often based on machine utilization. Thus, it becomes apparent that the main discon- nect between lean production and ERP systems is that lean flow methods are used to control production activity over the short-term time horizon, and ERP in the form of the master production schedule (MPS) and materials requirement plan (MRP) work over the medium- to long-term. The lean-ERP paradox is summarized in Table 2.

Table 2: The Lean-ERP Paradox (Powell and Strandhagen, 2011)

Lean ERP

Production based on consumption (Pull) Production based on forecasts and machine utilization (Push)

Decentralized control & empowerment (Bot- tom-up approach)

Centralized planning and control (Top-down approach)

Rate-based, mixed model production Time-phased, batch production

Focus on maintaining flow Focus on tracking material movements

The lean-ERP paradox gives an overview of the classical differences between lean and ERP, which have arisen from the distinction between JIT and MRP. It is clear that the two have emerged from fundamentally different approaches to production management: Just-in-time, pull production from Japan (Sugimori et al., 1977); and MRP push from America and the West (Wight, 1984). However, due to extensive developments in the capabilities of ERP systems, it now appears that there is a potential synergy to be real- ized in combining the two. For example, Riezebos et al. (2009) suggest that ERP sys- tems can dramatically reduce the amount of time required to obtain information relating to products and processes, as well as helping to increase the speed and quality of man- agement decisions, whilst simultaneously reducing costs. Al-Mashari (2002) also states that the use of ERP can stimulate the adoption of standardised business processes throughout an organisation. These motivations and benefits are clearly well aligned with the principles of lean production. Furthermore, many lean companies are using ERP based approaches for communicating demand through the supply chain in order to facil- itate just-in-time delivery, to the point where lean control principles (such as Kanban) take over.

Martin (2010) argues that IT applications and lean can be synergistically integrated in two ways. Firstly, IT applications should be deployed effectively and efficiently to in- crease an organizations flexibility in responding to external demand within the con- straints of available resources, helping achieve the goals of lean production. Secondly, he suggests that IT systems can be modified to accelerate the deployment of lean pro- duction in order to improve operational performance. He suggests that MRP II systems are designed to push materials through a supply chain, but they can be modified to pull

materials through portions of the same supply chain to increase its flexibility and re- sponsiveness, reducing inventory and operational costs and increasing schedule attain- ment.

Benhabib (2003) suggests that MRP and JIT strategies are not competitive but can actu- ally be seen as complimentary inventory management strategies. He states that whilst JIT emphasizes the initiation of production only when a firm order is placed, MRP complements this by back-scheduling the start of production in order to avoid delays for lengthy production activities. According to Benhabib, one can easily see the natural place of JIT in manufacturing companies today, where orders are received via the Inter- net and passed on to the shopfloor as they arrive.

Cunningham and Jones (2007) suggest that having a centralized ERP system is enor- mously important for a lean company. A standardized ERP system provides a common toolset that simplifies decision making processes, as all employees can view, discuss, and make decisions from the same standardized information. They also suggest that two of the most complex and important areas for the ERP system within a lean environment are order entry and order management. This is because lean manufacturing does not re- quire products to be produced to forecast stocking levels nor to maximize the operating capacity of the facility. Therefore, a key discussion point is the integration of Kanban with the ERP system.

In lean production, the operator takes a Kanban item off the shelf and uses it; turns in the Kanban card and more items are ordered. Cunningham and Jones state that it doesn’t matter to the operator if the ERP system thinks there are 1 or 1000 items in stock, as he has the one he needs and he knows more are on order because Kanban is a manual and visual process. However, it is the fact that Kanban is a manual process that poses the greatest risk where there is a lack of disciplined operators. For example, when the op- erator forgets to place the Kanban card in the correct place for the replenishment pro- cess to begin, the whole system fails. This is again where the application of automatic identification and data capture technologies (AIDCs), such as RFID, can help to reduce risk. When a Kanban item is removed from stock, a replenishment signal can be auto- matically logged or sent to the supplier.

To summarize, Gibbons-Paul (2008) suggests that though many lean purists believe that information technologies such as ERP are incompatible with the discipline, nothing could be further from the truth. Because ERP systems can help increase the speed and quality of management decisions, Riezebos et al. (2009) suggest that they offer a satis- factory level of support for lean production, making computer-aided production man- agement and lean manufacturing complementary technologies. This helps to confirm the previously defined research questions that will guide this research project.