Hello! I came across this issue today while trying to design a Pareto Chart in DV Desktop. Fortunately could figure out the snag and come up with a workaround. 

I assume that the measure "Cumulative %" has been calculated using a running aggregate function (likely RSUM). As you know, your Pareto distribution needs to be sorted by "Rev" in descendent ordering (giving this Loc order : E,D,B,A,C,G,I,F,H on the chart). However DV Desktop likely sorts the Data Set by "Loc" in ascendent ordering prior to perform RSUM aggregation. So the "Cumulative %" measure follows the wrong sorting order in the Pareto perspective. Notice on the Table visualization that you have the right "Cumulative %" column ordering when "Loc" column is sorted in descendent ordering. If you sorts the table on column "Rev", values in "Cumulative %" column get messed around. That's exactly what happens on the Pareto chart visualization as well.

As a workaround, you should generate a new Data Set natively holding the "Cumulative  %" column using a Data Flow by taking care to perform the right sorting in the appropriate step. Then use this new data set to design your Pareto chart. Alternatively, you could create a new calculated attribute by prepending numbers to the "Loc" column values so as to force this column ascending ordering to reflect the "Rev" column descending ordering as Pareto requires. Then you should use this new attribute column on the X axis of your Pareto chart visualization. The later solution is illustrated on the attached image. It's a bit ugly but it has the merit of getting the job done easily for very short Data Sets.