High % of NULL's - yet they want it!

I have a customer that when I do the profiling of the data source, some of the columns are between 90% to 99% NULL and they insist there could be value in bringing them into DW. Specifically to look at the 1% to 10% outliers. This sort of goes against my belief of what a DW should be for. A peer of mine working at another institution pointed out that if they think they need it and are really pressing for it, they are likely to get it one way or another and I should just give them what they want.

Just curious what others have had to consider in this scenario?

What other purpose is a DW for, other than storing data for users to analyze? The fact that a column is more one value than another doesn't qualify or dis-qualify it. If the business sees value in it, thats usually good enough for me.