Examining ERP

John Darlington of the Lean Enterprise Research Centre examines the rise of ERP, its foibles and the forces it brings to bear on manufacturing operations.

In the evolution of ERP we can see reflected the development of computing power via integrated, shared data systems for organisations.

The growth and adoption statistics for ERP are impressive. Often built on a centralised database and normally utilising a common computing platform, ERP systems encompass all business operations, marshalling them into a uniform and enterprise‐wide system. In plain English, this means you only have to enter data once and all users in all functions can see it and act upon it almost immediately.

This gleaming vision for the power ERP can bring to organisations is unfortunately sullied by the fact that they are breathtakingly expensive to install and prone to significant overruns of cost and time. ERP in manufacturing businesses are often justified on the basis benefits they bring in reduced inventories and lead times. Yet for all their promise too often they have failed to deliver, indeed ERP giants, SAP had some highly publicised lawsuits a few years back due to alleged system failures and Hopp and Spearman indentified a stagnation of inventory turns in the USA in the same period as a growth of MRP 2 adoption.

Why did these failures occur? In two significant instances basic ERP logic has been superseded; namely scheduling and costing. This is a case of uncorrected obsolescence rather than original error and we can explore what this means by considering how ERP systems deal with scheduling.

Scheduling ERP systems use fixed lead times to back off from the due date of the customer and calculate the start date of operations. This can be done by assigning lead times to each resource over which the part number passes (MOVE, QUEUE, SET UP AND RUN). Alternatively the lead time could be assigned to the part number itself.

The trouble is that there is no such thing on this earth as a fixed lead time. Common sense tells us that if there were no current orders in manufacturing the lead time through the factory would be short indeed. Conversely if the order book was experiencing rapid growth and the capacity were to remain the same lead times would extend for the each order taken. In ERP; both start dates of operations for the same part required on the same due date in these two polarised conditions would be the same.

Attempts to manage the resulting chaos (start dates can be in the past!) vary from ‘late to start’ reports which encourage piling even more work in progress into production, to glorified expediting.

Advanced Production Systems offering finite capacity scheduling can be a good potential alternative and can be specified within the suite of modules provided by big ERP vendors. That these systems have also been unable to make the sort of breakthrough that we might have anticipated is down to two reasons:

• Firstly; that the people who understand the software often have neither the time nor the inclination to try to understand the business environment including the product process characteristics of the plant they are supposed to help.

• Secondly; because within many businesses, one method of scheduling is recommended. A one model fits all approach is rarely suitable and many businesses have to cope with make to order and make to stock requirements over common production resources.

In addition, it has become popular in large multi‐location manufacturing businesses to specify a business template in an attempt at standardisation. This stifles local creativity and reduces the host site to compliance to a project roll out plan designed to minimise the cost of implementation rather than thinking about the business solution.

John Bicheno of LERC has observed that the justifications are often improbable whereas certainty resides in the employment of many expensive extra people for much longer than you thought at the outset.

You would have to have masochistic tendencies, considering the hernia‐inducing effort to get the damn thing up and running, to go back and revisit whether the original assumptions buried in the system are still relevant to the dynamic nature of your business but the fact is powerful IT makes people lazy. They rely upon the computer to do their thinking. Planners and schedulers become administrators of a flawed system and production spends their time chasing late to start reports.

There is nothing wrong with the architecture of ERP systems in manufacturing; I for one would not relish the prospect of manually exploding a deep bill of materials. Nor would I want to ‘net off’ demand against the parts already launched and in various stages of completion. For this alone I celebrate the invention of MRP and its evolution. Add in the data sharing abilities of the modern ERP system and you can appreciate power of good computing; one data base, speedy accessibility and the ability to do many transactions rapidly.

Start to think about increasing customer service however, with fewer inventories and improving response times, and real thought is required. It is not easy; it requires a thorough understanding of the business demands and production environment and an appreciation of the range of pull systems that might suit it. Good lean practitioners should know how to exploit ERP systems and how to make an intervention with an appropriate pull system, including, the vital ingredient of putting a cap on the work in progress.