In This Issue
It's very easy to survey your customers: these days it can be as simple as opening a (free!) account with an online survey service, spending a few minutes writing some questions, and sending e-mails to a bunch of customers.
Building an effective feedback program requires considerably more thought. Usually, the long-term goal of a customer feedback program is to improve the overall customer experience. This requires enabling more effective decision making at management and executive levels, and changing the behavior of front-line employees.
Effective customer feedback programs have three attributes: they are Dependable, Actionable, and Credible. I call these the "Three Abilities."
The dependability of the feedback you’re collecting is a measure of the scientific accuracy of the data: how well it reflects reality, whether it is truly measuring what you think it is measuring, whether the process is biased in any way, and whether the data actually supports the conclusions being drawn.
Dependable data can be counted on to support decision-making, but data which is not dependable could be inconsistent, unreliable, or even flat-out wrong. There is no single measure of dependability, but some important considerations are:
- Sample size and Margin of Error: How many people complete the survey, and what is the resulting margin of error?
- Process Bias: Are some groups of customers more likely to complete the survey? Are there barriers which prevent certain people from taking the survey at all?
- Sampling Method: Are customers selected for the survey randomly? Do a large enough fraction of the selected customers actually take the survey?
- Survey Construction: Are the questions clear or confusing? Do customers interpret the questions the same way you do?
Actionable data is data which has enough detail to enable business analysis and decision-making. This is independent of dependability: you may be measuring key performance metrics with a high degree of precision and reliability, yet not have enough richness to understand why the numbers are what they are and how you could make improvements.
Actionable data allows you to go beyond measuring “top line” results and understand what the drivers are. For example, knowing that your top-box satisfaction score is 82% is nice, but by itself that tells you nothing about what you could do to make it 85% next month. A very general customer complaint like, “I hate your phone menus” is hard to do anything about without any suggestions about why he hates the phone menus or what might make the phone menus better.
Actionability generally comes from having a rich enough data set to enable deep root-cause analysis of individual customer problems or a class of common problems. Customer feedback can be made more actionable by:
- Tying individual customers’ feedback to detailed records of the customer’s interactions. In a call center this might be call recordings and data about who the customer spoke to and the disposition of the call. In a retail environment, this could be point-of-sale records including the store and salesperson.
- Including more than just satisfaction questions on a survey. Customer service surveys which also ask about what happened during the customer’s interaction enable you to connect the dots between specific ways of delivering service and customer opinions. Are customers upset because the wait for a representative is too long, or because their questions aren’t being answered? Does a store have low satisfaction scores because of poorly trained staff or because the shelves are always empty?
- Back-populating historical survey data with information about future customer behavior. If a customer takes a survey about a service experience, and then six months later buys again from you, that fact should be reflected in the survey data for future analysis. This will let you draw direct lines between the customer’s experience and future buying behavior.
Finally, data may be scientifically sound and very rich but still useless if nobody believes it. Customer feedback needs to convince others in the company to take action to improve, even if they might have conflicting interests.
Unfortunately, the factors which make data dependable and actionable are not the same factors which give it credibility. Raw statistics tend to leave people cold and unaffected, where stories carry much more emotional weight. People also tend to believe data which supports their existing opinions, and discount data which doesn’t. If an executive believes his company provides good customer service, he will likely discount a customer survey showing otherwise.
Obvious flaws in the process or the results will also undermine the credibility of the data, and make it easier to ignore uncomfortable conclusions. It’s very common for employees to try to manipulate customer surveys to give results more favorable to them. Clear examples of bad data (for example, the customer clearly did not understand the question, or was answering questions about some other experience) should be discarded.
The challenge is to align the data with the story and get everyone’s buy-in on the process rather than the outcome:
- Illustrate the statistical data with individual stories, and the more emotional impact the better. Audio or video recordings of customers are best, and first-person written narratives are also powerful. The individual customer stories should support the conclusions from the data; for example, if the survey shows that front-line employees need more training you can play a recording of a customer describing how nobody knew the answer to her question.
- Get buy-in on the feedback process before coming to any conclusions. People who had a say in haw the data was collected will be much less likely to challenge the results, so make sure key decision-makers are involved from the beginning.
- Keep collecting feedback. Most people will eventually come to accept feedback they don’t agree with if customers keep complaining about the same things over and over.
- Promote audits and appeals. Especially if employees are paid based on customer feedback, they want to feel like they are being treated absolutely fairly. Have a formal process for challenging feedback they feel is flawed, but only discard data in very narrow circumstances so employees don’t feel they can get any negative survey thrown out.
A survey process which is Dependable, Actionable, and Credible has the traits needed to drive improvement and well-informed decision making. There still needs to be effective reporting and a strategic desire to change, but an effective customer feedback process will form the foundation for improvement.