Everyone involved with the development or delivery of software or technology solutions is aware of the importance of testing. Testing (or at least validating) what you create using Watson will allow for stakeholders to appreciate and comprehend the insights identified and shared with less concern for the correctness of the base information that the insights may be based upon.
Testing is a huge subject and beyond the scope of this book, so I will simply state the obvious: insights, no matter how exciting, are dangerous if based upon incorrect or invalid data or data taken out of context. Therefore, it is strongly advised to validate your data before making decisions based upon Watson outcomes.
There are an endless number of approaches to solution testing but with Watson solutions, you'll benefit most from preparation and preprocessing. In Chapter 2, Identifying Use Cases, we went over the importance of properly defining a use case, what to ask your data and putting your data into proper context. This is what we call preparation. In Chapter 3, Designing Solutions with Watson Analytics we covered data considerations and how to use the Watson Explore functionalities to handle data considerations or preprocess your data (we also touched on extending Watson in Chapter 6, Customizing and Extending by using external tools to explore and preprocess your data).
If you intend to share your Watson insights you might want to establish approval criteria to determine whether your share is ready for other eyes. These criteria might include the following: