To give an example, we consider that binary search requires less work than a linear search. But there are costs and usecase considerations involved. Insertion of new recod need to use binary search to keep the data sorted. Also if the number of lookups is far less than number of records, the overall cost is more than appending and linear search. That's what I mean by by moving the complexity around.
A problem scenario doesn't have absolute characteristics. It's relative to your way of looking at it, and your definition of a problem.