6 steps to a steep learning curve after project start-up

You have made it through to project start-up and, while it feels like the finish line is in sight, there is still the important matter of making the project pay.

The financial success of a project is very sensitive to how quickly it starts to repay the capital outlay. For complex production projects this means getting efficiency levels up as quickly as possible. Like climbing a mountain, climbing a steep learning curve should be done in a focused and prioritised way.

Like a good paint job, creating a steep learning curve is all in the preparation. This article has assumes that the project has had a robust governance process and that there has been the appropriate level of stakeholder input ensuring that the assets have been designed for optimum integration into the existing supply chain.

Step 1 – Create a process map for the new process

The Bullant Filters are sophisticated process maps consisting of 3 parts; Demand Filter, Capacity Filter and a Reliability Filter. For a project start-up you only need the Demand and Capacity Filters because there is no historical information to populate the Reliability Filter (reliability during start-up is addresses in step 3).

Step 2 – Identify where the pressure points will be during start-up

There will be plenty of targets during start-up and so its important to have a means of focusing effort. The Demand Filter and the Capacity Filter provide that focus by clearly identifying the most important running configurations and, for those running configurations, how the new system is designed to be able to deliver the correct performance.

Step 3 – Resolve to fix Special-Cause Variation before Common-Cause Variation during start-up

At start-up there will be extremely high levels of downtime due to the fledgling nature of the process and the occurrence of “infant mortality” of new components. During this time we will require a way of focusing our attention on the downtime that has the biggest impact on the learning curve.

There are 2 types of downtime or variation; Common-Cause Variation and Special-Cause Variation, we will be concentrating most of our efforts, at least in the early stages of start-up on Special Cause Variation.

  • Common-Cause Variation is downtime that can be expected and follows a statistical average. For example operators will be learning to run the equipment so their improving competency will follow a reasonably predictable Common-Cause path.
  • Special-Cause Variation is a special type of downtime that suggests a more systemic root cause. An example of a Special-Cause Variation may be a new bag sealer that keeps producing unsealed bags. The Special Cause may be a temperature set point that needs to be calibrated relative to the new bag material or a supplier that needs encouragement to deliver bag material to the agreed specification.

Fixing Special-Cause Variation during start-ups invariably has the biggest impact on efficiency, quality and learning curve. On the other hand, the statistical nature of Common-Cause Variation means that improvements here will have a slower effect on learning curve.

Step 4 – Prepare a information probing system ready for start-up

Of the 7 quality tools, the Pareto Chart is usually seized upon as the best tool for prioritising improvement efforts. This is usually true of continuous improvement efforts focussing on Common-Cause Variation, however Pareto Charts are better used as a supplementary tool during start-ups.

A better tool for start-ups, and Special-Cause Variation, is the Control Chart. A Control Chart is more useful for distinguishing between Common Cause and Special Cause events.

So, with steps 1 thru 4 completed we now have a suite of 4 information pools ready for start-up;

  1. Capacity Filter (identifies the expected performance characteris
  2. Demand Filter (identifies the major product families)
  3. tics of the system)
  4. Control Charts for major product families (identifies Special Cause Variation)
  5. Pareto Charts for major product families (prioritises Common Cause Variation)

These form the information backbone ready to help the start-up team prioritise their activities.

Step 5 – Prepare a Kanban system for activity prioritisation

During start-up there will be many activities that need to be completed as soon as possible. While these activities will be prioritised according to the data from the probing system, they still need to be matched with the available resources. The temptation here is to load up resources as much as possible however, there is a point where this actually reduces the speed at which activities get completed.

A project Kanban board is a great way of monitoring the progress of start-up activities while making sure that resources are loaded for maximum activity completion speed.

Step 6 – Finalise and train the start-up team

Its important that everybody on the start-up team understands the methods and terminology used in steps 1 to 5. There will be other activities going on in parallel such as validations, product testing and training so everyone should know how they fit in. Roles should be clearly identified and allocated as should a timetable of meetings.