Measuring customer experience

Q: I am trying to develop appropriate customer experience metrics for service delivery touch points at a financial institution, such as automated teller machines, internet banking, point of sale, mobile banking and bank branches. These metrics should be linkable to appropriate business financial performance outcomes, such as profitability, deposit liability growth and value of new accounts. Can someone who has worked on this help?

Fajimi Oladapo

A: In today’s environment — where every dollar not only counts, but is absolutely vital to long-term viability — you want to make sure that expenditures are moving the appropriate dials. In this case, it sounds like you want to make sure that the expense designated to improve customer experience not only improves the customer experience, but also improves some associated financial measurements.

Here is something important to remember: There are no silver bullets when it comes to metrics. More important than a catalog of potential metrics is a process to generate the metrics that you need, not a list of metrics that has worked for someone else. You need to know how to uncover and establish your own metrics.

There are two important tasks that must be completed: a current-state assessment and a future-state design. There are at least two schools of thought here. One position is not to bother with the current state. It will only serve to put you in a box consisting of existing paradigms. The other way to think about this is in terms of a gap assessment. It is easy to debate that without the current state and future state, it is impossible to know what gaps need closing.

So where do you start? Consider establishing your current-state baseline. As part of this exercise, it is critical to know the current performance levels for all dimensions, not just customer experience. In the past, I have taken a substantial matrix of data, correlated the data against itself and evaluated it for cause and effect. This narrows down the data from the trivial many to the vital few.

Next, or perhaps even concurrently, consider asking your customers what is important to them in terms of customer experience. Conduct surveys, interviews, focus groups or workshops to obtain the voice of the customer (VOC). You may want to consider a benchmarking exercise and research to find out what others are doing in this space.

At this point, you have your current state, potentially know what the future state looks like, and you have the VOC. You have a nice data set that you can leverage for conversation. Bring in functional leaders and start asking what they really need to run their area. Sometimes, leaders struggle with understanding the outputs (the “Y” variables). Spend time with them to make sure they get this part, as it is crucial to the success of the initiative. You are now ready to start testing for alignment between the identified metrics and the associated levers. This is where you learn whether you have the correct data so that you know the needle will go in the correct direction when you turn a dial. Here is an example of applying this thought process:

You find out that your customers are not happy with the number of times they are handed off during one of their calls to your organization. Your data also show there is a correlation between the number of handoffs and the number of times a cross-selling opportunity is successful. You decide to invest in training to reduce the number of handoffs by increasing the breadth of knowledge your call center personnel have. After a period of time, the cross-selling rate improves.

In this example, you were able to turn the dials (take action to reduce handoffs) and see a positive resulting outcome.

This response to your question does not give a listing of metrics. Rather, it gives an approach to get the metrics you need. Too many times, individuals look for silver bullets that just do not exist. A good process is better than a silver bullet any day.

Keith Wagoner
AVP Partner Solutions
Lincoln Financial Group
Greensboro, NC

SPC for low-volume assembly

Q: How can assembly manufacturing organizations meaningfully implement statistical process control (SPC) if they are purchasing all parts and assembling only a few units per month?

A: Let’s first interpret “a few units” as 30 or less per month. For the purchased parts, assume there is a mixture of commodity items and custom parts built to supplied drawings. As for the finished item, assume it is complex, such as a large medical instrument or an aircraft.

In this scenario, the primary focus must be on the parts and working with the suppliers. While the onus is technically on the suppliers to demonstrate capability of their processes to make your parts, you must work with them in partnership. It is more about supplier relationships than it is about SPC.

At the macro level, when choosing a supplier, verify it has a certified quality system or one that meets your organization’s requirements. At the detail level, you must identify critical dimensions or performance requirements of the parts or part drawings, measurement methods must be agreed upon and suppliers must provide inspection data via certificates of conformance.

Incoming part inspection must be conducted for verification until there is confidence in the suppliers and the parts. This is often termed “item certification.” Your organization will have to define just how much data are required before incoming inspection can be reduced or eliminated for a specific part. This may be a function of part criticality. Some high-priced custom parts may always require incoming inspection while, for others, inspection may be reduced or eliminated after a period of time. Commodity items may require minimal or no inspection.

Within your operation, subassembly testing results and final testing results can be charted over time. What is the nonconforming rate of subassemblies or finished units when initially tested? P-charts can be used. This can be done based simply on number of subassemblies or finished units, or it can be done with consideration for part complexity.

For example, a subassembly may have 20 parts. Count the critical dimensions on each of the 20 part drawings and total them. This can be termed “opportunities for error.” (There are other ways to count opportunities, but we’ll use critical dimensions for the purpose of this discussion.)

Say there are 100 critical dimensions (opportunities) in the parts of a certain subassembly. The nonconforming rate can be expressed as the number of failures divided by the number of opportunities. For example, if there were two failures in 10 subassemblies, the nonconforming or defect rate for that subassembly for that week or month could be expressed as:

2 / (10 subassemblies x 100 opportunities) or 2 / 1,000 or 0.2% or 2,000 defects per million opportunities (DPMO).

The same could be done for other subassemblies with results presented in Pareto fashion. If desired, an adjusted Pareto can be developed factoring in dollar value per subassembly. Resources may then be directed accordingly.

This example is best understood when thinking about putting together an initial subassembly from parts. But as subassemblies are put together into higher-level assemblies, the method still can be applied.

At the initial subassembly or higher assembly level, in addition to critical dimensions, the opportunity count can include performance requirements such as mechanical actions or electrical requirements.

But don’t get too caught up in the math. In almost all assembly or manufacturing operations, if you really want to know what the problems are, just ask the people doing the work. They will tell you.

Peter E. Pylipow
Principal engineer
Vistakon — Johnson and Johnson Vision
Care Inc.
Jacksonville, FL

For More Information

  1. Peter E. Pylipow, “My Supplier’s Capability is What?” Quality Progress, May 2003, pp. 60–64, http://asq.org/pub/qualityprogress/past/0503/qp0503pylipow.pdf.
  2. S.K. Vermani, “Capability Analysis of Complex Parts,” Quality Progress, July 2003, pp. 65–71, http://asq.org/data/subscriptions/qp/2003/0703/qp0703vermani.pdf.
  3. T.M. Kubiak, “Perusing Process Performance Metrics,” Quality Progress, August 2009, pp. 52–55, http://asq.org/quality-progress/2009/08/34-per-million/perusing-process-performance-metrics.pdf.

Average Rating


Out of 0 Ratings
Rate this article

Add Comments

View comments
Comments FAQ

Featured advertisers