In September of 2004, I had a survey epiphany that came from the most unusual experience.
When I went to pick up my son from his first day of pre-school, I noticed that he had a star sticker on his shirt. “Hi sweetie, “ I said “How did you get that star?” “Miss Kathy gave it to me.” He answered. Ooops – wrong question. “What did you do to get that star?” I asked? “I was able to say my ABCs” He finally said. See what I mean.

My son is a lot like your respondent – he answered only the question I asked which was “HOW” he got the star. What I wanted to know was “What did you do to get the star” These are two different questions with two very different answers. The first answer was information (ok), but the second answer was information that we could do something with. Saying your ABCs was a behavior we could identify and repeat. That’s good data.

Let’s get straight to the point. Why do a survey? The real reason any of us do a survey is because we want to make “the right” decision. So what’s a “right” decision? It’s when you choose to do something in a certain way – and your customers decide to choose you because of it.

So how do we go about pulling together a survey that asks the right questions and gets our customers to choose us instead of any competing alternative?

  • Figure out what decisions you want to make and what data you need to make the right decision. This is where most people make their first mistake. They start with what they want to know instead of why they want to know it. When you start with what you want to know – you’ll get create too many useless questions. That’s expensive, will piss off your respondents and won’t help you figure out what to do. Here is a suggestion. Write down the decision you want to make “Should I launch product X?” Now, what data do you need to make this decision? What needs to happen for you to decide to move forward?
  • I need to have 100,000 customers for product X at $12.99 per unit to be profitable.
  • I need to complete that sale in the first 2 months.
  • I would love it if I could sell the thing online and just drop ship from the manufacturer. If I can’t do that, it isn’t worth doing.
  • Define your objectives. That is easy now because you’ve done that work in step #1 where you defined what data you needed. Also, don’t get caught up in stating your objectives in some fancy corporate language. State your objectives in the same language that you would use to make your decision. The data will either support or deny your decision. Your objectives in this example are :
    • See if there are at least 100K people who will buy this
    • Determine if more than 100K people are willing to buy within 2 months of launch
      • What information will they need to say “yes” to my offer?
      • How do they want to receive that information?
    • Will they accept online ordering only?
  • Identify your ideal respondent. Who do you currently think is your ideal customer for this? Where are they? How can you reach them? What’s the best way to collect their input? If your ideal customer for this is a group of existing customers, swell. You have their information. But what if your idea customer is a group you’ve not sold to before? Should you buy a list? Maybe. You can recruit a panel of customers with the characteristics of your ideal customer. This is a terrific idea because it’s a passive and less intrusive way to start selling and educating a new group of customers while gathering their preferences.
  • Qualify, Group and Test. Next you want to make sure that the respondents you’ve chosen are good respondents and fit the profile that you want to survey. You have several options for this step. You can use a quick online survey to quickly collect, qualify and profile your ultimate respondents. Identifying your target audience and setting the profile criteria as a quick 3-5 question 1 minute web survey can really pay off later and increase your response rate when it really counts. So, if your target customer is a man between the ages of 25 – 35 with a BS degree that makes between $70 and $125 per year and eats Beef Jerky 3 times a week. Then you’re going to need a large sample and then you’ll need to whittle that down to the several thousand potential respondents that fit that profile.
  • Develop your OPEN Ended questions. Suddenly this doesn’t seem so hard now, does it. The questions almost write themselves at this point. Let’s assume we’ve recruited a respondent base that fits our profile above (the guys who make lots of money and eat beef jerky 3 times a week) Now we want to know more about what they do, why they do it, what’s important to them when they are buying what we’re selling and what we can do to get them to choose us.
    • Who buys products like product X
    • What triggers you to buy product x
    • Why do you buy product x?
  • Test your questions over the phone. I know, phone surveys are expensive – but not as expensive as decisions make with faulty data. So don’t skimp on this one. Take the touchy-feely open ended questions you developed and talk to about 30 sample respondents. You want to do this over the phone with a high-end interviewer so that you can catch what’s really important to these people. You want to catch the emotional triggers that will ultimately drive them to try your product x.
  • Develop your web survey. Now you’re ready to hit the web with a short survey that has taken the core of what you’ve learned from your open ended interviews and distilled them into the fewest number of critical decision-influencing questions you need.
    • Qualification profile questions. These are the demographic categories that you’ll want to use (income, gender, geography, etc.)
    • Would you buy product x?
    • What would you pay for product x (define desired profile for price)
    • Would you buy product x online?
    • If not, what do you prefer?
  • Test . To test your survey take it yourself and ask several other internal people to take it as well with one rule – answer the questions with how you think your respondent will answer them. Now pull the results from that test and PRETEND they are real. What decision will you make and why? Do you have all the information you need to make the decision? Do you know what to do or what to tell other people to do? For example if you asked how satisfied people were with p
    roduct x and the answer was 3.5 – what will you DO? See, you don’t know – so that’s not a good question. Go back and re-write it so that you know what to do.
  • Launch and Follow-up. Now you’re ready to launch and see what happens. Remember, responses come it quick for online surveys, so stay close to the computer in case there are issues. Don’t forget to re-send your survey in a few days to increase your response rate. And offer your respondents something cool for their trouble. An easy and valuable thank-you gift is a download of some sort that your respondents will like and can pass on to potential new customers.

Cool Analytical Tools on Web Surveys

You can do the example I gave using standard web survey software and analysis tools. But this scenario that I used is a prime candidate for “Conjoint Analysis.”

With QuestionPro’s conjoint tool, you’ll create a set of “Living” data that you can use to create product offerings at a variety of price levels for several different profiles of customers. This analysis will actually tell you how many people will buy a specific permutation of product x at a particular price. OR the analysis will tell you what to charge a profile group if you’ve already defined the product.
This analysis used to cost a fortune and require a staff of statisticians – but now even those of us with limited statistical ability (yours truly) can use this tool with ease.
That wasn’t so bad…

See, that wasn’t so bad. It takes a little more work up front, but I promise you the results from this survey process will yield much better results and make you look like a hero.

Guest Posts : Please contact marketing@questionpro.com if you would like your post/company to be featured on the QuestionPro blog.

We’ll cover the following topics:

  • Online Research
  • Innovation
  • Crowdsourcing
  • Marketing Strategies
  • Anything else you can convince us to write about ;)