Why is so much customer satisfaction research so unsatisfying?

When I talk to research users, one of the most common themes I hear is dissatisfaction with customer satisfaction research (and its new expansion into customer experience). Key problems appear to be:

• It tends to be based on people’s memory rather than their current or recent experiences, resulting in inaccurate measurements.
• Halo effects can swamp actual, real-time experiences and distort the recalled evaluations. For example, if somebody likes an experience they are prone to say everything was good. However, if the experience was negative they are prone to describe most of the experience as negative – this makes the findings less actionable.
• The research is often targeted at phenomena that are of little interest to research participants. For example, questions about how a customer is greeted at the bank or supermarket can result in blunt and hazy measurements if users do not find that aspect salient to them.
• A company’s area of interest is often too wide to create a meaningful research instrument. One consequence of the breadth of interest has been the creation of ever longer customer satisfaction surveys, delivering ever lower levels of satisfaction to the users of the research.
• Customer satisfaction and experience is a lagged measure, it talks about what happened, not what is happening. The longer the delay between the measurement and the reporting, the more lagged the information is, meaning it can’t be used proactively in managing a brand or service.

One of drivers of the current problems with customer satisfaction research is that it is used for two incompatible purposes:

1. To find out what should be changed and what should be retained to maximise the success of the brand.
2. To measure the performance of the organisation, often to the extent of including customer satisfaction KPIs in pay and bonus calculations.

A system that helps improve a brand should be agile, finding out the current pain points, digging into them, helping solve them, and moving on. But a measure of performance needs to be consistent and needs to reflect the organisation’s goals (which are not necessarily customers’ issues).

My recommendation would be to use a very simple, inexpensive, and transparent model to evaluate the organisation – for example sales, profit, or NPS. This leaves the organisation free to use tactical approaches to tackle customer satisfaction and experience. One great resource for this sort of research is provided by insight communities, working with customers to optimise the brand.