When planning for a data project it is impossible to know all the issues that may arise, so as you investigate and resolve tasks, issues are very likely to appear.
To avoid this difficulty, testing and planning are essential. For example, if you looking to identify duplicate records and then subsequently merge them, you will need to test the data within the duplicates to see how they impact on the merging process. Often the results affect the strategy for the merging process, but this is not known until the duplicate data is made visible.
Suppliers require a flexible approach to managing customer expectations. Data will always give rise to difficulties; the key is to know where these difficulties are likely to occur and then test any results before committing plans.
The other 3 Cornerstones are: