Hey Lin, 

yeah i can understand this is confusing, it's been confusing me all day!

Here is an example of what I want:

So the user is able to choose a destination week (example: last week), and a source week, lets say the week before that, using the slicers. The table will then show all combinations of segments that people had in those weeks. For example, if there were only 3 people in my data, and this is what happened:

"Johnny went from Leales to Baja Frequentia" and "Lizzy went from Novatos to En Riesgo", and "Benny also went from Novatos to En Riesgo"

The resulting data should show:

Destination Segment       Source Segment     Number of users

Baja Frequentia                Leales                      1

En Riesgo                         Novatos                  2

That's the idea.

As you can see in the example I have come reasonable far with it. The problem is that the second column has to be a measure, and that measure currently uses a "MAX(destination segment)" which will always give Novatos.

I hope this helps a bit. Sorry for confusing you!

Jaap