Customer Surveys 101
Often, customer surveys are poorly executed. Which is too bad, because organizations that conduct surveys show that they understand the importance of asking the opinion of their customers.
Sometimes customers are asked to answer many delicate questions without a clear purpose for these questions. Or they are confronted with a question list longer than their arm, where you have to have a PhD to even be able to understand the questions, let alone answer them!
When done right, surveys can be a great tool to enhance customer satisfaction. Make sure you plan your way through a survey, or your customers are at peril. Don't underestimate how annoying a bad survey can be.
I will outline how you can perform successful surveys, from setting a goal for it through analyzing the results and follow-up on it.
1. Set the goal of the customer survey
Defining a goal is essential to determine if the customer survey has been successful. When choosing the words of the goal, make sure that the results are measurable. So if your goal of the survey is to measure customer satisfaction, select specific areas.
For example, you could survey on usability of your forms. In that case, you can select a number of improvements coming from the survey, and track this list until all improvements are incorporated.
2. Select the responsible persons in your organization
As for any successful undertaking, it must be clear who is responsible for what. If responsibility is unclear, you cannot count that activities are executed when they're supposed to.
Don't think that activities are picked up by the most "logical" person, you may find that your logic can differ widely from other peoples logic. As you can't look inside everyone's mind, assign responsibilities for each step clearly!
3. Determine the medium
Selecting a proper medium can seriously impact the results of your customer surveys, so you have to give it some thought. Here are some possibilities:
- Internet page
- regular mail
Also decide if you want to accompany the customer survey with an accompanying letter, or something like that. Be creative! (But don't forget that each method has its price tag)
4. Select the customers to survey
Choosing the customers that will be surveyed is directly related to the selected goal. Make sure that you get a good spread over your customer base and a large enough selection. It's impossible for me to give lessons in statistics here, but if you want a good introduction on statistics, I recommend "Fundamentals of Applied Statistics and Surveys" by David B. Orr (ISBN 0412988216).
5. What's in it for the customer?
Why should the customer participate in your customer surveys? There must be some kind of perceived value. The promise of improved service may be such an incentive. Or a gift certificate. Or a discount voucher. And remember, if rewards are higher, the customer may be more forthcoming with answering longer surveys or privacy related questions.
6. Write the customer survey
6.1. Types of questions you can use
When writing customer service surveys, you can choose from the following types of questions (I've added an example for each one):
- Closed question (1 possible selection, radio button type)
Please state your marital status:
O Rather not say*
*) Note: for privacy related questions I often add a "rather not say" choice, unless of course it's essential for the survey.
- Open question (text field type)
Describe one function you would like to see in our software:
- Identification (Identifying the customer)
Please state your full name:
- Multiple Choice (more possible selections, checkbox type)
Please select the newsletters you want to receive:
O Customer Servings
O Help desk software tests
O Associate Programs
- Valuation scale (very bad--very good)
What do you think of the newsletter?
Very bad <-----> Very good
- Importance scale (Not important--very important)
For me, status is...
Not important <-> Very important
**) Often I see a 5 scale, but this opens up a "safety choice", where the customer can choose safe middle ground. With a 6 scale, the customer must choose (in those cases) either a slight positive or slight negative angle. Knowing the nuances is often as important as learning the extreme choices.
6.2. Sequence of the questions
Putting the questions in a good order is also important for the success of the customer survey.
The questions should invite the customer into the survey. So, always start with the easy questions, as starting with difficult questions can easily scare her off.
Put identification questions at the end of the customer survey.
6.3. What to avoid
When writing surveys, some mistakes are made that result in a greatly reduced return ratio. Please learn from these mistakes, and don't:
- ... create extremely long surveys. If answering the survey is not in line with the reward for the customer, she will not bother to fill it in. A lengthy survey takes a lot of time, so there must be a great "reward".
- ... ask privacy related questions if not absolutely necessary. Customers are protective of their privacy, and rightfully so. Most customers frown on having to answer many privacy related questions. If you really want to know (f.i. salary level) make sure you have an opt-out choice in the question. If you don't, reassure the customer about how you handle these questions.
- ... create a survey that's difficult to answer. A customer willing to answer your survey, will eventually abandon it if it's incomprehensible.
- ... (DON'T!) forget that a bad customer survey will make your business look bad :-(
6.4. Closing up
- Don't forget to thank the customer for answering the survey.
- Inform the customer what you will do with their answers (and don't forget to follow-up if you promised that!)
- If you offered an incentive, remind them of what and when they will get it.
7. Testing the customer survey
This step is often overlooked, yet can save you a bundle! Since conducting a survey can cost a lot, it is always a good idea to test the survey on a few persons before you put it to the world at large... This can show you any big mistakes you made, like these:
- unclear questions
- similar or same questions in one customer survey
- impertinent questions you had not viewed that way
- a forgotten choice (in case of multiple choice)
- survey not "flowing"
At the very least do an internal test, where the customer survey is executed by some colleagues. You can also choose to do an external test. If the cost of executing the survey is high, an external test is an absolute must! In an external test, you make a small sub selection of your selected customer base (step 4), and conduct the survey with them. It goes without saying that the results of both internal and external tests must be evaluated before taking the next step.
8. Send out the survey
At this point you have selected a medium and a customer test base. You have created and tested the customer survey. Now it is time to send it out. (I realize that this sounds like a mailing action, but it can actually be that you send out your team of interviewers to do their thing, or activate an outbound call center! This depends on your chosen medium)
Timing IS important! For instance, if you chose to call customers, call them on a time they are in. Sometimes a weekend is better, sometimes a weekday. Think about what time is best for your customers!
9. Handle the response
Hopefully the answers on your survey are now pouring in. Especially if you are doing an offline survey, the results must be collected in a format that can be analyzed. Often, this means that the results must be entered in a database or a statistic package.
Online surveys are much easier handled, as the results are usually collected in a usable form by the survey application. But it may be that you have to do some conversion or formatting anyway.
If you promised an incentive, now is the time to send it to the customer, accompanied by a thank you note.
Measure the response! If response is low, you can choose alternative action to minimize money wasted.
One more thing: sometimes the customer uses the survey to bring a serious complaint to your attention. If this is the case, take immediate action on that complaint as if the customer had just called it in.
10. Analyze the information
Analyzing the results of the survey usually involves many statistical calculations and pattern recognition. A well-known package for this is SPSS (www.spss.com). I would like to tell you more, but analyzing surveys goes way beyond the scope of this article.
11. Plan your action
Now you know. You know what your customers want. Don't make the mistake of ignoring their wishes. Remember that customers "vote with their feet", so inaction is no longer an option. Imagine the frustration of taking the time to answer your customer survey, and then noticing absolutely no change. A customer service survey may increase customer satisfaction, but only if things change!
So make a plan that builds on what your customers perceive as strengths, and corrects any found faults. And when you have finished your plan...
12. Just do it!
Action speaks louder than words,
but not nearly as often.
You must now show your customer that you have heard them.
Top of Customer Surveys