From SCDigest's On-Target E-Magazine
March 9, 2011
Supply Chain News:How Much Technology is Needed for Lean Manufacturing?
Purists Say Not Much, but that isn't the Case for Many Manufacturers; Growing Complexity Requires more Support
SCDigest Editorial Staff
How much technology support is needed for Lean manufacturing? The topic can generate some fierce debate. Many Lean “purists” would argue that IT should play almost no role in Lean initiatives, and in fact can often get in the way of the simple “visual” based shop floor signals that drive classic Lean operations.
But a growing number of companies and observers are arguing that the role of technology in empowering Lean has been sold short, and in many cases is simply essential to scale Lean and make it work in complex environments. Who’s right? Well, as usual, it depends.
SCDigest Says: |
 |
Less debatable is the fact that as product mix and production process complexity increase, the �low tech� approach to Lean becomes increasingly challenged to keep up. |
|
What Do You Say?
|
|
|
|
The debate was fueled in part by a disconnect a decade or more ago between “no technology” Lean adherents, and the fact that most companies implementing Lean then were also using ERP/MRP software on the shop floor, solutions that at the time often did not well support Lean operations. Back then, the fundamental problem was that ERP/MRP were founded on a “push” supply chain model, while Lean is inherently pull-based.
That disconnect and other challenges led to inevitable conflicts, and helped foster the view that technology got in the way of Lean, which for many companies at the time it often did.
The hope that Lean can be implemented with little or no IT is of course also appealing to companies thinking they can achieve major benefits with minor investments in technology. But like many promises that seem too good to be true, the reality can often turn out to be a different story, as a number of companies are reporting.
A critical dimension in terms of the role of technology in Lean is the nature of a company’s manufacturing operations. In simple, straightforward manufacturing environments, it is relatively easy to tie together materials and information through use of simple visual “kanban” (replenishment) cards to signal the need for materials or parts from downstream operations, and to do basic scheduling manually. The simple kanban approach, in which cards and materials serve as a proxy for information, is especially workable in single-threaded production lines that have consistent flow paths and dedicated machinery, and/or in operations that deal with a small number of SKUs and relatively stable demand.
But that is not the world of a growing number of manufacturers today. Complexity is severing the connection between material and information in those environments, often necessitating a new set of tools to cope with that gap.
Why Many Experts See an Increased Need for Technology Support
There are a variety of factors driving the need for broader use of technology in Lean manufacturing.
The first – and perhaps most controversial – is that Lean initiatives in many companies have a tendency to thrive and sputter in cycles, as the focus on and knowledge of Lean come and go in the company. Because pure Lean is so dependent on people and their knowledge of Lean practices, its success can be subject to much variation over time. Toyota, in fact, used the recent slowdown in demand for vehicles to retrain thousands of idled workers on TPS concepts and practices, rather than laying them off.
As a result, many argue that putting in a technology foundation to drive Lean processes and information flow will serve to institutionalize Lean within the factory.
Less debatable is the fact that as product mix and production process complexity increase, the “low tech” approach to Lean becomes increasingly challenged to keep up. Many manufacturers deal with hundreds of products, with highly variable demand and complex routings. How can traditional low tech approaches possibly well manage that scheduling and routing complexity?
For instance, scheduling the so-called “pacemaker” operation (the driving step in the Lean process that pulls the rest of manufacturing activity, generally near the end of the line) for just a few dozen products using spreadsheets and manual techniques is one thing; doing for the hundreds of products many companies must support is quite another. Scheduling other processes around the pacemaker manually also becomes nearly impossible as material flow and routings become more complex and machines/cells handle multiple products. In complex environments, robust shop floor scheduling tools are required to achieve level scheduling and maximize capacity utilization.
The same challenges to the low tech approach are found in complex environments for managing kanban sizing and work-in-process strategies. Even traditional kanban-based replenishment approaches must be resized on an increasingly dynamic basis in complex environments.
Manual approaches, in which kanbans were resized once or twice per year, quickly lead to sub-optimal work-in=process inventories in those manufacturing operations. Other companies are using a blend of approaches for WIP inventories, combining kanban techniques with newer approaches such as CONWIP (from MIT), POLCA (from the University of Wisconsin), and other strategies that clearly require technology tools to manage successfully (see our Lean Manufacturing resources web page for more details on these practices).
SCDigest, in fact, spoke with one company that had hired a large consulting firm to help it implement Lean on the factory floor. Initially, the consultant positioned that the Lean implementation could be almost technology free. But after months of problems and challenges, the consultant company ultimately implemented at least a dozen “kanban calculators” on the shop floor to help operators optimize replenishment quantities.
Technology support was needed, and “home grown” technology like this can sometimes be used but has a challenge scaling across the operation.
(Manufacturing article continued below)
|