Support & FAQs

Adaptive Questions® use the power of crowds to evaluate answers submitted by other respondents on the fly. Each respondent evaluates 10 answers that have been submitted by others. About every third respondent will specify an answer which becomes available for future respondents to review and evaluate. In a survey of 300 respondents, a single Adaptive Question® will gather and prioritize about 100 answers.

Your objectives dictate what type of methodology is appropriate. If you are in a situation where you want to evaluate internal business solutions to those offered by customers, employees or other groups, Adaptive is a good choice.

Yes.

• Adaptive Questions® are good substitutes for open-ended questions because respondents read the comments and tell you what issues are most actionable.
• Adaptive Questions® are also good substitutes for a series of rating scales where your objective is to prioritize next steps, product development answers or features.

• When your existing survey is too long to get all the answers you need
• When you are not certain what to do after reading the results from your current survey
• Just about any time that you don’t know the issues
• When you need a fast result
• When you need to innovate for product development, line extensions, messaging or any other situation that calls for brainstorming or innovating

Yes. While many NPS® users struggle to get clear direction to drive their scores, our clients frequently tell us that they regularly get specific and actionable insights from Adaptive Surveys®.

While Adaptive Surveys® are new to a lot of researchers the technique has been used by leading companies for several years including: Intuit (TurboTax, Quicken, QuickBooks, ProTax and App Center), Callaway Golf, A&E TV, My M&Ms®, Disney Interactive, Fidelity, Capital One, Ogilvy, Virgin Media and others.

You can get useful results from a surprisingly low number of responses. Adaptive Survey® Technology automatically adjusts the results for sample size using the exact confidence interval calculation described by C. J. Clopper and E. S. Pearson in 1934.

For other types of questions, sample size is calculated the same as any other project depending on the accuracy of the results you require for your specific project. Many people look for a sample size of about 380 in order to see results that are accurate to plus or minus 5 percentage points. Of course you can increase accuracy by getting more responses. If you are interested in segments that are a small percentage of the total, you might consider increasing the sample size or oversampling the segment.

Yes. You might want to use comments from a previous research project or from your social network such as a blog or Facebook page.

Respondents review answers and comments entered by other respondents which might include distractions such as misspelled words or bad grammar. That makes them more difficult to read and evaluate. For that reason, we recommend that you keep surveys short – one Adaptive Question® and up to five segmentation questions ideally. Our experience indicates that best-practice Adaptive Surveys® include no more than three Adaptive Questions®.

Branding is a concern for some companies because respondents don’t necessarily follow established guidelines. You can opt to monitor comments as they come in to correct or reject comments that don’t meet your criteria. We only recommend monitoring where absolutely necessary; answers waiting in a queue are not tested until the moderator approves them.

Sometimes offensive language can be an issue. Respondents usually don’t agree with these types of comments and they fall out of the testing quickly. In addition, we automatically filter most bad words and only show them if you specifically approve them. Monitoring allows you to correct or reject answers that may offend others.

They look like a regular question and respondents never know what is happening behind the scenes. Adaptive Questions® are a little more forgiving since respondents tend to comment about things that are on their mind because of the wording of most Adaptive Questions®. The seed answers that you provide initially are important and should be composed in the proper context just to get everyone started in the right direction. Think about all the rating scale questions you might normally ask and turn those into seed answers instead.

1. Off-topic answers usually fail to get agreement because there are always lots of people who read questions carefully
2. You can remove the answer from the system on the fly. If something still gets through, you can easily remove it from consideration in your analysis.

It should be similar to an open-ended question in order to elicit a thoughtful response. We suggest that you ask for solutions, not problems. Use questions like, “What should we do the next time to make the experience better for you?” instead of “What caused you to give us such a low score?”

The objective is to prioritize answers about a particular subject. Your interest is in the answers that are most likely to be actionable; which answers you should consider first. While an absolute scale works in some cases, it is difficult to read in other cases.

We expect you to have an internal conversation as a result of any research, “We are already working on answer #1.” “Answer #2 is impossible using our current technology. “ “Answer #5 is something we can implement tomorrow with almost no expense.”

There are two schools of thought on this. Some people feel that it is useful to see where positive and negative respondents agree – so put all of them in the same Adaptive Question ™. Other people don’t want positive respondents to be influenced by the negative comments they might see and prefer to split them.

We apply a confidence interval to the results which takes the low base into account. As a result we are 95% confident that top answers are really the best and that the lower answers are really the worst.

No problem. Run Adaptive Surveys® on our site or integrate with your own site using API.