customer-development
, market-research
, lean-startup
When attempting to validate a hypothesis using the lean startup canvas, how do you express customer development questions to increase the accuracy & precision of responses that are received, and to insure the maximum amount of learning during a given iteration?
Accuracy and precision of responses stem from a variety of factors that could be further researched under proper survey techniques and methods. I would point to larger areas of improvement such as sample size, randomization, and proper question writing. However, it seems like the most applicable answer to what you’re looking for are principles that illicit truthful and honest responses to varying questions that you will use to pivot or focus your startup. With this in mind I’ll throw out a few suggestions that you can use when writing your questions regardless of the method of collection of this data.
The Focus Shift (which you mention in a comment) – Rather than asking a particular respondent for their opinion, which can make them fear answering honestly due to perceived or real fear of retribution, ask them something like “What are some concerns or bugs you think this type of customer might have with our service?” or asking an individual at a larger company about a service “How do your peers rate our services and what do they complain about?”. These would be in contrast to questions like “What do you think are bugs with our system?” or “Are you satisfied with this feature?”
One other thing I would suggest if you’re surveying repeatedly over time is to try to develop questions or surveys that you use consistently. That way you can measure the change in responses over time and see the trends without having to attribute them to changes in survey questions. This can actually help you get good data even if your responses aren’t as accurate and precise as you would like.
There are lots of other survey suggestions online that you can take a look at if you search for “best survey practices” or something similar.
Hope this helps!
All content is licensed under CC BY-SA 3.0.