What has come to be known as Lean is essentially a process improvement discipline. As Shigeo Shingo outlined as early as the 1950’s when he began his relationship with Toyota, Lean focuses on processes, not individual operations. Shingo was not interested in improving individual efficiencies unless they could be seen to contribute to overall leadtime and inventory reduction, a theory he called ‘stockless production’. Shingo drew a matrix model of processes (flow) and operations (individual tasks) and insisted we focus on process improvement and flow.
As the Western concept of Lean developed in the 1990’s Mike Rother devised Value Stream Mapping as a way of clarifying this concept. In a value stream map we analyse two flows, those of materials and information and attempt to synchronise the two. In terms of material flow Rother utilises Little’s Law of Queues which states that leadtime through a process is equal to the quantity in the queue divided by the rate at which the process clears the queue, so that if we have 400 units of material in our operation which produces 40 finished units per hour then the leadtime is 400/40 = 10 hours. Our value adding percentage is then the ratio between the time spent adding value to each unit in individual operations, say 30 minutes, and the leadtime, 10 hours, giving a value adding percentage of 5%. It was the realisation that if we improve individual operations we are limiting ourselves to less than 5% improvement in leadtime which led Shingo to his process/operations insight.
Shingo’s insight was that there are only four categories of operation: transformation, inspection, storage and transportation. Of these, only transformation ie changing the properties of materials, is of any real value. Storage and transportation are completely non-value adding and, to Shingo’s mindset, inspection should be unnecessary if processes were mistake proof (Poka-Yoke). Personally I am willing to accept some value in inspection if the customer demands an inspection report as part of the delivered product, but I know that this was not Shingo’s view.
The other fundamental lean principle we apply is Jidoka, Toyota’s second pillar alongside JIT. In manufacturing operations Jidoka means separating the person from the machine so that people are not tied to individual operations, particularly to watching semi-automatic machines while they wait for the operation to be performed by the machine. In administrative process it means separating individuals from the process so that tasks can be rearranged and combined to optimise the process.
Ironically, Little’s Law was originally devised to calculate how many tellers were needed in a bank in order to keep queues to an acceptable level, so that in applying Lean concepts to financial services we are returning to the origin of the concept. Unfortunately in modern administrative processes it is unhelpful to distinguish between material and information flows as they are one and the same – the material is the information, and so Value Stream Mapping is of limited benefit.
Some years ago I was introduced by Boeing Corporation to a technique which they used to map administrative processes and which I have subsequently favoured over Value Stream Mapping, Swim Lane Mapping and other common process mapping techniques. I have used Four Lane Process Mapping in pharmaceutical research processes as well as financial service processes, having deviated somewhat from the original Boeing model, but have found it a very effective way of promoting process improvement in non-manufacturing applications.
The technique is based around two well-known approaches, firstly an insistence on standard operating procedures (SOPs) for all operations and secondly on the industrial engineering process of ECRS – Eliminate, Combine, Rearrange, Simplify. Using ECRS we first eliminate unnecessary operations, then rearrange and combine the remaining operations in the most efficient sequence before simplifying the remaining operations to reduce cost or time. ECRS is itself a simplified form of Osborn’s Checklist, but is perfectly adequate for most uses (Shingo’s fabled SMED technique is to all essential purposes ECRS with the addition of the internal/external changeover operation distinction).
One ancillary technique I use to discuss standardisation of operations with groups is Professor Fukuda’s Window Analysis, which asks whether there is a known standard which is followed, a known standard which is not used or no known standard. If these categories are used by two individuals or two teams in a 3 x 3 matrix then it generates four categories (see illustration). Where neither party can identify a standard, then this is a Standardisation problem – ie there isn’t one, and the first countermeasure is to create one. Where there is a standard which is not known by one or both parties then this is a Communication problem, which requires communication and training. Where both parties know the standard but one or both do not use it consistently then this is an Adherence problem and the countermeasure is to discover why and provide further guidance. Where both parties know and work to a standard but the results are not satisfactory, then there is a Reliability problem and we must work to make the standard capable. The insight here is that before we start trying to improve the standard we need to know whether there is a current standard which is being used correctly.
Four Lane Mapping works by describing a process on four horizontal lanes on a large roll of paper. The Four Lanes are: Process as Documented, Process as Witnessed, Process Issues and Process Improvements. In using this approach however I have found it best to start with Lane 2, Process as Witnessed. To create lane 2 we need a group of process users as well as other interested parties – CI specialists, managers and internal customers for the process. The process may encompass more than one department as information is processed to provide the result for the internal and external customers.
One example is of a financial services company providing secured loans to small companies for capital investments. Brokers would put forward customer proposals to the company, who would then have to provide a loan offer, often the same day or even within two hours. The process would involve assessing the value of the asset, the creditworthiness of the customer and level of risk and would pass through a number of departments, including underwriters. Once a proposal had been made this could be live for up to 180 days before acceptance by the customer, who then required funds to be released at short notice.
The project was to reduce leadtime from receipt of customer acceptance to release of funds as this was a major factor in customer satisfaction and the ability to generate repeat business. We examined the process right from opening post and receipt of email acceptances through to release of funds. Working with the team we mapped out lane 2 and were immediately struck by a number of issues relating to duplication (“why does your department do that? – we’ve already done it”), inefficiency (“We never knew you did that – that makes our job much harder”) and sequencing (“if you could let us know you’ve done that earlier, then we could start this task much earlier”).
At this point we then assemble all the documented standards and use lane 1 to cross-reference them to the process as witnessed. We are looking for divergence between the documented standard and the process as witnessed and also for gaps in the documented standards. By the end of this we had a huge number of issues documented in lane 3 of the map. As an improvement process note, I usually try to break for the day at this point (for a large process this will be at the end of day two) so that people get a chance to reflect overnight and come fresh to the process the next morning.
Assembling the next day we review the analysis, pull out common themes and then begin by applying ECRS and asking which tasks provide no value and can be eliminated. To do this we examine the function of each task, asking what its purpose is and why we do it. As we had already identified a number of duplicated tasks this serves to get the group moving on developing process improvements, which are documented in lane 4.
We then look at rearranging and combining tasks to give us the optimum sequence of operations, making sure we adhere to the Jidoka principle where possible and separate people from process. We may also introduce the idea of administrative kanban at this point to maintain First in First out (FiFo) through the process.
Once we have the optimum sequence sketched out we can then look at individual tasks and their documented standards to see if they can be simplified in any way. The output of a Four Lane Process Mapping workshop would be an optimised sequence of tasks to accomplish a particular process, improvement ideas to simplify particular tasks and an action plan to implement the improvements and develop the revised standard procedures.
In regulated environments such as financial services and pharmaceuticals this will also involve obtaining the appropriate approvals for process changes and one way to approach this is to divide the proposals into ‘Dos’ – things which can be done now without any external approval or additional resources, ‘Fixes’ – those things which require some resource planning, but not external approvals and ‘Redesigns’ – those things which require substantial resources or approvals and need to go through the management approval process.
I have found Four Lane Mapping to be a very useful and powerful process in a variety of non-manufacturing environments. An initial workshop can last from two days (the minimum to do any serious work) to four days (if the analysis is going to take longer than four days you might want to partition the project into multiple events). It is important to be very clear about the starting point and finishing point of the process under analysis in order to avoid ‘process creep’.
Lean tools such as Four Lane Mapping are just that, tools to be used in particular circumstances. There is an old saying ‘to a man with a hammer, everything looks like a nail’ and we have to be aware of the value and use of each tool in our armoury and use them as appropriate. Where we have separate material and information flows and physical inventory or work in process I would use a Value Stream Map. Four Lane Mapping is for reasonable complex administrative processes – to use another common image, using Four Lane Mapping on simple processes is using a sledgehammer to crack a nut.
What is also important is that the team you are working with understand the basic principles – process v operation, JIT, Jidoka, ECRS. These are often best learnt through simple simulation exercises which provide an enjoyable introduction before getting down to the hard work of examining your own processes.
My own experience of using this approach is that people are very often staggered by the level of improvement they can make to their processes and I am often likewise staggered by the creativity and ingenuity of the teams I am working with.