menu icon
close

Why Managing by Exception is Key to Supply Chain Efficiency

By Bernard Milian
warning symbol - supply chain efficiency

Every planner today has hundreds or even thousands of references to manage. Often, these references must be synchronized during the assembly of products with complex bills of materials, where the slightest shortfall stops the flow. If you manufacture an electronic assembly with hundreds of components, a simple diode that costs nothing can stop everything. How can you prevent this from happening while maximizing supply chain efficiency?

Going through every part number to make the best decision is no longer an option. You must get straight to the situations that are most likely to cause problems.

The difficulty is that this must be done in a very dynamic, fluid, and uncertain context, because today’s demand and supply projections could change significantly tomorrow. 

Generating the right level of alerts without generating unwanted noise is a challenge. If you want to alert your planners to all the risks, you will drown them in alerts. Have you ever tried to exhaustively process exception messages in an MRP system?

Moreover, the sensitivity of the risk detection must be adapted to the horizon on which we work. 

Execution Alerts Balance Sensitivity and Discipline in the Short Term

In the short term, for three hours from now, for tomorrow, for next week, you can/should be sensitive and precise, without becoming hypochondriac. This is the area of “buffer management” at the heart of the Theory of Constraints.

With each buffer — especially the inventory and time buffers — articulated in red, yellow, and green zones, planners have a gradation of alerts, in order to really focus on situations that deserve attention.

In addition to this basic DDMRP functionality, we have defined alert criticality levels and a review process that allows us to highlight only a limited number of situations each day, whether on MTS or MTO items, to avoid false positives.

For the process to be effective and efficient, it must be followed in a disciplined and orchestrated way. Everyone has a defined role — from the equipment operator to the planner — and must deal with alerts as they arise. Once this discipline is in place, the level of disruption and stress is greatly reduced, as only new risks and a limited number of items need to be addressed.

One of our customers recently told us: “Since we started Replenishment+ we have reduced the number of alerts by a factor of 10.” By having 10 times fewer alerts, planners gain peace of mind, focus on the essentials, and contribute to improvement.

This example customer, with a portfolio of 4,000 items, has to process about 30-40 critical alerts in a day — in a particularly difficult procurement context for that company. 

DDS&OP Identifies Longer-Term Risks

If you are looking at the medium term, on an S&OP (or Demand Driven S&OP) horizon, you need to identify at macro level the risks of stocks being too low or too high, lead times being too long, and capacity risks, to prepare your operating model. There is no need to work at a fine level: things will change, you just need to make sure you have the bandwidth to handle the change. 

Don’t ask your planners for precise three-month forecasts, or your supply planners for a firm six-week production plan. They will boil the ocean for not much benefit in the end!

Based on your forecast scenarios, you will identify risk situations — that is, items or groups of items whose stocks are likely to fall too low or rise too high, and situations when the load would be incompatible with capacity. To limit these alerts and focus on the essentials, you will only consider materials that are in difficulty for a significant period. 

Ignore the risk of a three-day outage in three months; things will change. On the other hand, a shortage risk for three weeks in three months deserves to be analyzed, and if the forecast scenario is plausible, the buffers should be adapted.

Statistical techniques and BI approaches help adjust the sensitivity of the analyses to really focus on the essentials. For example, the amplitude of the fluctuations of the buffers projected in the simulations of our DDS&OP module will facilitate the detection of forecast fluctuations and generate the right decision process.

At this point, do not try to manipulate the buffers too much. Before considering adjustment factors, check that your buffer calculation method is relevant. For example, it is more relevant to calculate an article with high seasonality or high growth on a forecast basis, and you will reduce the need for compensation via zone adjustment factors.

If your buffers are designed and updated in the right way, 95% of your projected items probably won’t deserve more attention. 

Boost Supply Chain Efficiency by Increasing the Signal to Noise Ratio

Managing alerts and recording your day-to-day buffer statuses allows you to identify recurrences and root causes — and feed the improvement of the model. Over time, this learning process allows you to further reduce the noise level, for more targeted, relevant alerts, dealt within the adequate time horizon.

The more unstable the environment, the more essential it is to be able to focus teams on the real priorities.

In his book The Haystack Syndrome, published in 1990, Eli Goldratt underlined the importance of finding the few relevant pieces of information in the middle of the ocean of data in which we evolve. This is even more important today, and it is at the heart of the management by exception that we implement with our solutions.

Get in Touch

Share This Story, Choose Your Platform!

Facebook
Twitter
LinkedIn

Recent Posts

Sign up to our Newsletter

You may also enjoy