Saturday, February 9, 2008

Don't Survey Your Customers!



Scaled customer surveys are among the most widely used tools in business. Unfortunately, they're also some of the worst. There's nothing evil about surveys, but they can turn an inherently simple task, such as gathering customer feedback, into something complex and unwieldy. When that happens, there's a good chance it won't satisfy its original purpose, which in this case is making improvements. Why exactly are surveys the wrong tool for most organizations? Let's explore the reasons and then consider an alternative approach that's far more appropriate.
Surveys don't produce timely data

Most traditional customer surveys are sent out periodically to a sampling of an organization's customers, typically once or twice a year. This is a manageable frequency from an administrative standpoint because implementing a survey requires a significant dedication of time and effort. The downside is that by the time the organization receives the feedback, the information is at least six months old and usually much older. The information is almost worthless because customers have already acted on their perceptions before the organization has had a chance to do so. Customers don't wait around to tell you what they're going to do before they do it. If you're not tuned into your customers on a regular basis, you'll never know what hit you.

It makes more sense to gather customer feedback continually. Make the customer feedback process something that's always happening, not a grand event that occurs once or twice a year. This consumes far fewer resources, and it also ensures that the information is current. If you can't take action on customer perceptions within a few weeks of the perceptions being formed, there's a strong chance that you will lose your window of opportunity.

Too many questions

Another downfall of most surveys is that they try to do too much. They probe the customer experience from every imaginable angle. Although admirable, this approach results in long, unwieldy surveys that most customers run away from as fast as they can. I have personally gotten into the habit of scrawling "TOO LONG" in huge block letters on these kinds of surveys. I'm providing feedback, but not exactly the kind expected. Most people don't even bother to provide this much; they simply toss the long surveys into the trash.

The key to successful customer feedback is to ask about the few aspects of the customer experience that matter the most. By asking about everything under the sun, you're establishing the expectation that you'll take action on everything, which is impossible. You're also telling your customer, "Your time isn't very valuable, so the imposition of this long and boring survey should be no problem for you." Focus on a few vital issues, and these obstacles go away. The dilemma is that most organizations don't know what the few vital issues are--thus the need for long surveys. Your organization must back up and get its arms around the things that really matter to your customers.

Difficult to design

If you like defusing explosives, you'll love creating surveys. They include so many failure modes that they're nearly impossible to design correctly unless you do it for a living. Why are they so hard? Let's examine two of their more challenging aspects: questions and scales.

Most surveys comprise a series of questions or statements, followed by a response scale. The response scale usually represents degrees of satisfaction (e.g., "very satisfied," "satisfied," "neutral," etc.) or degrees of agreement (e.g., "strongly agree," "agree," "neutral," etc.). Both of these scales present huge challenges. Most people don't have the writing skills to craft clear, unambiguous survey questions. The result is that the questions don't accurately reflect the attribute that's being queried. In the spirit of getting the job done, customers will often take a guess at what the questions really mean. Like all guessing games, sometimes they'll be right and other times wrong. At best, your data will be 50-percent valid--not a very good percentage.

In the unlikely event that the survey questions are clear, there's still the obstacle of designing a logical response scale. This would seem to be an easy task, but it's extremely complex. Typical errors I've observed are scales that aren't balanced, scales that are biased, scales that don't have equal intervals between the points, scales that don't match the question or statement, and scales that have too many degrees of resolution. If the scale is flawed, then the data that come from it are also flawed. Junk in, junk out, as the saying goes.

No direction for improvement

The fourth downfall of traditional surveys is that they don't provide much guidance for improvement. Sure, they provide data, but what actions are you going to take based on those data? For example, let's imagine that you've asked customers to rate the technical knowledge of your sales force. The average response is 3.4 on a five-point scale, roughly halfway between "neither good nor bad" and "good." What does this number really mean? Even more important, what are you going to do about it? The data help you produce fancy charts, but they probably won't steer you toward specific improvements.

If you aren't able to take action on survey data, their value as improvement tools is zero. Had you asked customers an open-ended question such as, "What do you think is our biggest customer service weakness?" you might have received some feedback that provided a clear path for improvement. In the world of customer perceptions, data don't always rule. Perceptions are by their nature qualitative and subjective, and the attempt to produce data from such a fuzzy source can be misleading. It's better simply to get actionable information than to attempt to turn human beings into precise measuring instruments. If you capture customer perceptions, analyze the trends and take action, you've accomplished a great deal.

So, what should you do?

I've made the case that a traditional customer survey using scaled responses probably isn't the best way for most organizations to capture feedback. If not a survey, what should you do? Here's what I recommend:

Step 1: Examine your existing customer interactions. Your interactions with your customers are limitless. These contacts are conducted via telephone, e-mail, mail, fax and in person. Because you already have numerous contacts with your customers, there's no reason to invent a new contact for the sake of collecting customer feedback. Make use of the connections you already have, and all parties will generally be much happier.

Step 2: Choose an interaction suited to collecting feedback. Not all customer contacts are created equally. Certain conditions should be met when you decide which contacts will be leveraged for collecting customer feedback. In general, the contact should be neutral, routine and candid. Here's what each of these mean:

• Neutral: The contact isn't related to an existing problem or complaint. Attempting to collect feedback when a customer already has a problem is obnoxious and counterproductive. Use an interaction that's neutral in tone and purpose, such as a query or order placement.

• Routine: The contact happens on a regular, routine basis. Feedback collected from routine interactions of this sort is likely to be fresher.

• Candid: The contact occurs between parties that trust one another and are willing to communicate freely. A candid relationship is key to collecting accurate and representative perceptions.

Step 3: Develop a tool that's matched to the customer interaction. Choosing the right tool for the job is critical in every endeavor, and that goes for collecting feedback, too. Once the organization has selected an appropriate customer interaction for collecting feedback, it must develop a tool that works in that context. This is a subjective task, and certain guidelines can assist in knowing what tools work best in different situations:

• Telephone contact: An unobtrusive tool that's conducted at the end of a routine telephone call. Brevity is critical with this kind of tool because most people are anxious to get off the phone once their business is completed.

• In-person visit: A tool that enables the company to see its product in use, just as the customer experiences the product or service. The tool should also enable different people to be queried, depending on the nature of the feedback sought.

• After service or consumption: A tool that enables the customer to conveniently provide "flash feedback." Make the return of this feedback seamless. If the customer
has to expend any effort to return the feedback, it probably won't be returned.

• E-mail: A live link within the body of the message that takes the customer to a simple and visually appealing evaluation of their experience. Make sure the link works fast and is compatible with a variety of Internet browsers and computer monitors.

Benchmark customer feedback tools with other organizations. There's no virtue in being original. Borrow good ideas and approaches as you see them. Hundreds, if not thousands, of examples exist for each of the tools described above. See what other people are doing and adapt the methods to your own needs.

Step 4: Focus on open-ended questions. If you want to grab the attention of your customers, ask them what they like and don't like. It's that simple. Asking simple, open-ended questions of this sort enables the customer to dictate the content of their feedback. What's important to them is what you'll hear about. This is exactly the sort of feedback you want. Trends in open-ended feedback will inform you on the issues that customers care most about, something that many organizations don't understand.

Open-ended feedback also provides a clear path to improvement. Numerical ratings can help you prioritize issues, but they don't tell you exactly what to do. Open-ended feedback can. When 75 percent of your customers answer the question, "What makes you most frustrated about being our customer?" in the same way, you know exactly what you need to do to improve. There's no ambiguity.

Open-ended feedback doesn't help you make fancy charts. But do you really need more fancy charts to cover the walls of the conference room? No, you need improved customer satisfaction and loyalty. Open-ended feedback will reveal exactly what actions lead to long-term success, which is much more important than fancy charts.
Here are my favorite open-ended questions. Three are about all you need. They will quickly point the way to improvements that matter to your customers.
• Do you have any problems with our products that you haven't told us about?
• Is there anything you think we do particularly well?
• What could we do in the future that would make your job easier?

Step 5: Act on your opportunities. Action is the most critical step of the entire process. It starts with identifying trouble areas. Problems that are revealed through feedback must be addressed immediately. This is the business equivalent of triage: Stop the bleeding and stabilize the patient. Let's hope you won't discover too many issues that require triage, but it's better to learn of these proactively while the customer is still your customer, and not a former customer.
After addressing the trouble areas, the organization must analyze the trends. Open-ended feedback follows the same rules as most traditional numerical data: It tends to clump into categories. Group the feedback into categories and apply Pareto analysis to the results. Your opportunity areas will quickly emerge. Input these opportunities into your corrective/preventive action system and track them to completion. Treat every improvement action as a mini project, with assigned tasks, responsibilities, time frames, resources and reviews. The more sunlight shines on your improvement action, the better it will be. In other words, communicate widely. The final communication about your improvement will be to your customer: "Here's what we've done based on your feedback." These may be the most important words you ever say--and you don't have to use a traditional customer survey to say them.

2 comments:

add said...

Yes. That’s true. There will be such kind of customers where the feedback will be something complex and unwidely. There is a chance of loosing the confidence with this kind of customer survey. Make sure that the survey is done with the proper persons.
http://www.addsurveys.com

jelly1921 said...

I made a translation of this article into Chinese on http://jelly1921.blogspot.com/2009/02/dont-survey-your-customers.html