Let's continue with the same preceding case: an ETL that processes sales files coming from different branches of a business. This time, we will add one more task: after loading the file, you should generate a report with the result of the process. Each branch may decide how it wants the report. Some may want a report by email, others may want that you generate a simple text file with the results and save it in a given folder, and so on.
Now for each branch, you have two specific jobs: one to get and read the file, another to report the results.
We could continue with the idea of naming the jobs/transformations differently for each branch, for example, report_results_A.kjb, report_results_B.kjb ...