The average service business spends about 30% of its time and money on activities that don’t contribute to the bottom line. Facilities managers are now turning to tools developed in the manufacturing sector to eliminate this waste and increase process efficiency.

Many service companies, including Jones Lang LaSalle, Xerox, American Express, General Electric and Canada’s banks have made effective use of Lean Six Sigma to introduce significant service improvements and increase their return on investment (ROI).

Six Sigma represents the combination of two powerful toolsets: Lean, used to reduce waste and improve throughput; and Six Sigma, used to increase consistency and remove the common causes of defects.

Toyota initially developed Lean as part of its Just-in-Time manufacturing and production system. By eliminating bottlenecks and reducing work in progress, the company found that it could increase the throughput of a system and, hence, its effective capacity. Lean is most effectively applied when the causes are fairly obvious, once examined properly.

Meanwhile Motorola developed Six Sigma in an attempt to create products with a very low defect rate. The term six sigma, or 6σ, comes from the symbol for standard deviation, which is a lower-case sigma, or σ.

This underscores that Six Sigma is about data and measurement, and reducing variation in processes. It is most effective when applied to problems for which the true cause isn’t obvious or where there are multiple linked causes that affect the outcome.


Proponents of the two systems spent many years trying to prove that theirs was the better approach, but more recently it’s been recognized that they are complementary, not competitive, and share many similar objectives. The fundamental similarity is that both Lean and Six Sigma are driven by the customer’s needs.

Lean is focused on “value-added” activities, based on what the customer would be willing to pay for. Six Sigma starts by asking what things are Critical to Cost, to Quality and to Delivery (CTC, CTQ and CTD), again based on what’s critical to the customer.

Of course, Robert Rodin, CEO of Marshall Industries, says it best: “What customers want is simple. They want it now, they want it perfect and they want it free.”

This is not news to facilities managers. For example, when overseeing moves to new locations, they are typically required to meet the schedule (usually tight) of the business units affected (Critical to Delivery), achieve efficiency of space and economy in procuring services, equipment and furniture (Critical to Cost), and meet the standards expected by the organization (Critical to Quality).

Both Lean and Six Sigma are also concerned with maximizing ROI. This is simple mathematics: spending $10 to save $50 is better off than only saving $20. This may seem obvious, but it’s easy to get distracted by trivial problems if the approach isn’t rigorous enough.


Lean Six Sigma (LSS) problem solving follows a five-step process: define,  measure, analyze, improve and control – DMAIC (pronounced duh-may-ick) for short.


Properly defining the problem is the most important step in the process, but is often not given its due. Albert Einstein stated it very eloquently: “If I had an hour to save the world, I would spend 59 minutes defining the problem and one minute finding solutions.” He recognized that if a problem is well defined the solutions are much easier to achieve.

There are several “must haves” from the define phase:

  • Project Charter
  • Current-state, value-stream map or swim lane diagram


The concept of a project charter should be familiar to anyone who has taken a project management course. It provides the foundation for the project, not only by defining it very specifically, but also by obtaining authorization and determining the commitment of stakeholders. The charter for an LSS project should, in addition to the scope and authorization, include the process history and a stakeholder assessment.

An important concept on a lean six sigma project charter, which would not typically appear in a PM charter, is the Voice of the Customer, which leads to the three CTs (Critical to Cost, to Delivery and to Quality).

The project charter describes how these will be improved by the project, as determined by the customer’s requirements (hence, the Voice of the Customer). Another fundamental difference from a project management charter is that in Lean Six Sigma, the Quality-Delivery-Cost triangle is not a zero-sum game. It’s not acceptable to reduce cost at the price of poorer quality. All of the CTs must either improve or at worst remain the same.

A value-stream map (see Figure 1) shows the process at a high level (usually 6-10 chunks). Completion and wait times and work-in-progress quantities are identified for each step in the process, as well as whether they are value-added or not. Non-value-added steps that aren’t required for regulatory or business purposes (called business non-value-added or mandatory non-value-added) can be eliminated at this stage. Removing non-value added activities early in the project also saves time that would be later wasted trying to improve them. Don’t fix it if you can throw it away.

Figure 1: Value-Stream Map

Swim lane diagrams identify the process at a more granular level and describe not only the detailed tasks for the process but also the role that performs each task. Figure 2 shows a swim lane diagram for a process used to vet and hire contractors. The responsible party is clearly identified for each step, as are the handoff points.

Figure 2: Swim lane Diagram

The SIPOC (Supplier-Input-Process-Output-Customer), on the other hand, is more interested in what each activity in the process requires as inputs and what it delivers. For each step in the process, it describes who supports and is supported and what tangibles are delivered to or by the process. Figure 3 shows the same activities as Figure 2, but in a very different form.

Figure 3: SIPOC Diagram

Once the process is mapped, it’s efficient to identify and eliminate obvious non-value-added activities. Keeping with the philosophy of narrowing the scope of the problem as much as possible, if the obvious problems are eliminated, then there will be less wasted effort in completing DMAIC.


The next step is to measure. Once the process has been reduced to the essential steps, serious data-gathering can begin. The information gathered must reflect both the potential causes as well as the desired results. Deciding what to measure requires careful thought, but it is key to obtaining meaningful results. Measurement can come from a variety of sources, but in general it is considered better to generate fresh data. Existing data is often inaccurate or missing critical data, and hence may be misleading or require considerable rework.

The measure phase should result in the following:

  • Hypothesis about critical causes
  • Baseline performance of the process
  • Data collection plan (Xs and Y)
  • Measurement Systems Analysis (MSA)


A hypothesis is a statement of the possible causes of the problem, before any testing or measuring. This usually comes from developing a cause and effect diagram, with input from subject-matter experts, to understand the most likely causes of the problem.

Baseline performance is the current state before any improvements. So, if the problem you are trying to solve is poor temperature control in an area of your building, the baseline performance would be the temperature range and average. Gathering baseline information is key to demonstrating improvements later.

A data collection plan is very important. Gathering the wrong data can be costly to your time and schedule. Common tools for gathering data include checklists, tally sheets and surveys. System information can also be used, but care is required as it is often full of erroneous or misleading information.

MSA is most commonly associated with calibrating and testing gauges and physical measuring devices, but it means, generally, ensuring that measurements are meaningful and accurate. In a service application, this might include considering whether or not instructions are likely to be misunderstood, whether people are going to have the time to fill in checklists properly, and the impact other activities (sick leave or vacations for example) would have on the results.

A cause and effect diagram (Figure 4) is a great brainstorming tool and is useful for a wide variety of problem-solving techniques. It’s sometimes also called a fishbone or Ishikawa diagram. The example shows a common starting point for a fishbone diagram. Depending on the type of process, different categories may be used.

Figure 4: Cause & Effect Diagram

When using this, the idea is to generate as many ideas as possible and not to discard duplicates on different parts of the diagram – these actually point to common causes and may be keys to the solution.

A technique called “5-Whys” is very useful here. Starting with the observed effect, ask “Why did this occur?”, which will generate a range of ideas. Apply “Why…?” to each of these. Eventually root-level answers will start to appear. These are called common causes and may be worth further consideration.


Once measurements are gathered, they have to be analyzed to determine which ones have a significant effect on the result. This is typically carried out by the Lean Six Sigma Black Belt, who will apply a variety of statistical tools to the data.

For the lay practitioner, there are a few simple tools that can be helpful even without going to this level of effort. Histograms, Pareto diagrams and Box plots are very useful for identifying gross areas of concern. By segmenting the data using these tools, patterns emerge which can be used for more detailed examination.

The following charts show how this might proceed. They look at maintenance and service requests generated over a five-month period earlier this year. The objective was to identify and address any problems with time to complete tasks.

The first step is to summarize the data so that it’s meaningful. In this instance, a Box plot, as shown in Figure 5, was created showing the distribution of duration by month.

Figure 5: Box plot of Time to Complete Work Requests


A Box plot is a powerful tool that provides a lot of information in an easy-to-read form:

  • The gray boxes show the middle quartiles (i.e. the 25% to 75% values of the list) and the median (line through the middle – this is the middle value when all of the values are sorted in ascending order).
  • The lines up and down from each box, or whiskers, show the first and fourth quartile values.
  • Asterisks indicate outliers – these are data points that are too far from the median. Outliers usually indicate either some sort of unique problem or erroneous data.

Additional information we get from the Box plot comes from how tight the boxes are. A small box, like the one shown for March, indicates that the process is producing similar results and can be considered fairly consistent. A large box, such as that shown for May, indicates a process that’s producing inconsistent results.

A good next step is to find out what types of requests are driving the longest times. A useful tool to use for this is the Pareto chart (Figure 6), a kind of histogram that shows both the contributors sorted highest to lowest and a cumulative percent curve. This is typically used to identify the top 80% of problems.

As the chart shows, the biggest contributors to jobs taking over five days were the movers and HVAC problems. Lights, locks and janitorial requests make up the rest of the top 80%.

Figure 6: Pareto Graph


The improvement phase then begins to address actual solutions. It may seem like a lot of work to get to this point, but it avoids the greater inefficiencies of implementing a solution that has no effect on the problem. Better solutions are attained by understanding the problem well.

In this example, the solution was to start by meeting with the move supervisor to ask what happened in May. The problem with May was fairly simple: tickets weren’t being closed. However, further discussion revealed that he was spending a considerable amount of time managing other paperwork – in most cases redundant – and not closing the tickets.

A common approach is to brainstorm solutions based on the analysis, perform a risk assessment and test solutions. Improvements are often tested in a pilot environment before being applied to the entire organization. This way, the improvements can be finely tuned.

Risk assessments can be conducted in a variety of ways, but a simple yet powerful tool for this is the Failure Modes and Effects Analysis, or FMEA. The steps for completing an FMEA are:

  • Describe possible failure modes
  • Describe the effects of the failure
  • Estimate likelihood (frequency) and severity of occurrence
  • Identify controls for most significant effects and estimate likelihood of detection
  • Determine actions and estimate new RPN
  • Calculate RPN after actions applied


Determining figures for likelihood and severity and the likelihood for detection usually follows a five-point scale, of which there are many depending on the application. One example appears in Tables 1, 2 and 3.


The results are entered into a simple table and the risk product number (RPN) is calculated by multiplying the three factors together. The risks with the highest value will require the most consideration. Once actions are identified to mitigate the risks, the new RPN values are calculated, then again once the actions are applied.


Sustaining the improvements is essential or all of the hard work will be for naught. When improvements are made, results must be measured and tracked so that old habits don’t return. At this point, the process map should be rebuilt to reflect the improvements – and published. The process inputs must be managed and the output measured and recorded on something like a control chart.

A control chart (Figure 7) is a simple yet powerful tool. It simply documents the performance measure against a control range. Wayward trends can be quickly identified and reined in before things get out of control.

Figure 7: Control Chart


Thus, the DMAIC process is the core of Lean Six Sigma methodology. Although it’s presented here as a straight-through process, it is, in fact, iterative and can loop back to an earlier step anywhere along the way. This is a strength of the process as it ensures that errors or omissions will be corrected before faulty improvements are put into place on a wide scale.

Such projects will generally be overseen by a Lean Six Sigma Black Belt. Black Belts receive a significant level of training in LSS and all of the various statistical toolsets, and are roughly equivalent to a certified project manager. They, in turn, may be working under the direction of a Master Black Belt who would provide guidance to the Black Belts as well as to the LSS implementation strategy.

The rigor of Lean Six Sigma methods can uncover and eliminate process problems and avoid solutions that cost more than the problems they address. Properly used, LSS techniques can lead to dramatic improvements in the way a business operates.




Chris Wheeldon is Industry Manager  for Cisco Systems  He holds a Master’s certificate in Six Sigma (Black Belt) from the Schulich School of Business.