The Six Sigma tool SIPOC (supplier, input, process, output, customer) is used to map processes at a high level. SIPOC diagrams are typically included as a slide in project presentations – often as a required element of a presentation template – accompanied by a perfunctory description. Presenters may not know what do with the SIPOC beyond this use. However, there are more benefits to be gained from SIPOC; this article addresses some of those uses as seen in a business process outsourcing company.

Elements of a SIPOC

First, a quick review of the elements of a SIPOC:

S (supplier): Entity that provides input(s) to a process

I (input): All that is used (mostly as variables) to produce one or more outputs from a process. It is worthwhile to note that infrastructure may not be considered as inputs to a steady-state process since any variability induced by such elements remains fixed over longer periods of time. (Exceptions include new infrastructure being introduced or a greenfield project.)

P (process): Steps or activities carried out to convert inputs to one or more outputs. In a SIPOC, the process steps are shown at a high level.

O (output): One or more outcomes or physical products emerging from a process.

C (customer): Entity that uses the output(s) of a process.

Table 1 shows a sample SIPOC for the process of handling inbound calls at a call center in a business process outsourcing company.

Beyond SIPOC

A SIPOC diagram can be the springboard to other critical activities in process improvement, each building to the next step. In particular, it can be used as the starting point for:

1. Identifying sources of variation
2. Metric identification
3. Determining a relationship between the variables
4. Generate improvement opportunities and projects
5. Create control plans

Identifying Sources of Variation

In Six Sigma efforts, the primary objective of improvement is to reduce variation in process output. This requires identification of the causes of variation, which is where the SIPOC can be utilized. Every item under the SIPOC categories of input and process is a potential source of variation in the output to be considered or investigated. Once the SIPOC diagram is created, look at the items identified under each SIPOC category and consider how they can influence the process outputs (Table 2).

For example, the output “resolution/assurance of resolution to customer issue” (seen in Table 1) will vary due to variations in the corresponding inputs and process steps, such as “frequent changes in call quality guidelines leading to inconsistency.”

The above is not an exhaustive list of attributes/metrics identified against the items; however, a measurement plan can now be prepared to collect data on these items. The initial data collection will involve some work, but once the relationships between these metrics are established with the output measures, the list can be further refined and a final measurement plan prepared.

Even a “simple” SIPOC may reveal a significant number of parameters that need to be measured and controlled to ensure quality outputs.

Metric Identification

With the possible sources of variation defined, it is possible to start measuring them to 1) determine the significant sources and 2) develop plans to control or reduce the variations in them. Table 3 displays a list of attributes and metrics determined from the SIPOC shown in Table 1 and from the sources of variation shown in Table 2.

The variation due to frequent changes to call quality guidelines, for example, leads to disparate understanding of those guidelines by various call center operatives. If the variation in understanding among operatives can be reduced, reduced variation in the output – resolution/ assurance to customer issue – can be predicted.

Determining a Relationship Between the Variables

SIPOC also helps determine the relationship between variables related to a process. After identifying the various attributes and associated metrics under each SIPOC category, it is easy to then start establishing correlations between those elements. Where measurements are possible, statistical tools like regression analysis, factor analysis and logistic regression can be applied to establish a relationship or model between the attributes and metrics. For example, an equation between “end customer satisfaction score” (output metric) and “forecast adherence/deviation” (input metric) can be determined.

Once the relationships are established, metrics that, if improved, would significantly impact the outputs can be determined. A list of such metrics can be prepared and the measurement plan finalized. The rest can be eliminated from the measurement schema. The shortlist of critical input metrics can help clarify how the process needs to influence the supplier. The business process can develop and deploy a supplier quality assurance program to address the quality of critical inputs.

Consider an example referring to the SIPOC shown in Table 1.

Assume the relationship model derived through regression analysis between the end customer satisfaction score and other input and process factors looks like the following:

Y = (w₁ x F₁) + (w₂ x F₂) + (w₃ x F₃) + (w₄ x F₄) + ……..

where

Y: End customer satisfaction score

Fi: Factors (input or process) that are significant (F₁ = forecast adherence/deviation; F₂ = availability of SOP updates, etc.)

wi: Coefficients attached to each factor

Depending on the mathematical value of the coefficients attached to F₁ and F₂, the factor that has the relatively greater contribution to the Y can be determined and actions to ensure that the performance of these critical factors can be taken.

Generation of Improvement Opportunities

SIPOC has the ability to unveil several areas of improvement. In the process of establishing measures for the elements and identifying the sources of variation, there will also be opportunities to improve output metric performance. This becomes more clear with the understanding of relationships between the parameters under each SIPOC element.

While the performance of an output metric that is critical to the customer or process is the primary driver to launch improvement projects, these other capabilities can also help proactively identify weak areas in an apparently steady process. It can be easy to not delve into a process because the performance metrics are meeting stated targets, but there is always a possibility that the metric would fail if any of the input or process measures fail marginally. With the metrics in place and with data available for establishing relationships between them, a sensitivity analysis can be carried out and the level of risk the process maintains is exposed.

Next, the target levels of each critical factor can be determined to ensure achievement of the output metric Y. Assess the gap between the actual performance levels of the factors against the set target levels to decide which factor or input or process variables need to be improved. This requires regularly collecting data on the factors and the output metric.

For example,  consider that for the SIPOC described in Table 1, the company has established that the training effectiveness measure (input metric) has a direct relationship (strong correlation) with the repeat call measure (output metric). Additionally, a regression equation between these variables has been established:

Repeat call % = A + [B * (training effectiveness %)]

Here, A is a constant and B a coefficient for the variable training effectiveness %. These are determined based on data and statistical techniques.

The above equation can predict changes in repeat call % by varying the training effectiveness %. Additionally, considering the current value of repeat call % and its difference from the target, a safe range or threshold for training effectiveness % can be determined. With this in place, the organization can keep a close watch on the training effectiveness % values and control it within the safe threshold to ensure the repeat call % does not fail to meet the target.

Improvement projects can be taken up on output or input metrics whenever a performance issue is found and variations seen in the process. The improvement efforts should not only be directed internally to improve process capabilities but also externally to the suppliers who provide critical inputs.

Creation of Control Plans

The final capability of SIPOC stems from the above capabilities. The organization can now establish clear control plans for managing the suppliers, inputs and process effectively to ensure control over the outputs and to satisfy the customers. These plans will typically include actions to be taken, the frequency of the actions, responsibility, status check, and governance on the entire plan by the quality and internal audit functions.

In the instance of an inbound call handling process, the control plans will encompass all critical measures including the output metric (lag indicator) and the factors (lead indicators). A comprehensive plan may include:

• Data collection trackers
• Control charts for critical metrics
• Action plan trackers with responsibility and timelines to address out-of-control situations
• Review mechanism (internal with management and external with client)

It should also be noted that the process owner should validate the relationship equation over a regular interval and whenever there is a significant change introduced in the process. This is required to ensure constancy of direction and effort toward sustaining the levels of the outputs.

More than Process Mapping

SIPOC is not a mere process mapping tool. While that might be its primary objective, there are other benefits to be enjoyed with this tool. Some of these other benefits require additional data collection and statistical analysis, but all originate from a simple map that is drawn using simple methods combined with existing process knowledge and experience.