Your definition of a 'NEW' Customer implies
that each those Customer should be identified
as having Sales inside a 52-Week Moving Window.
Though Tableau has Moving Table Calculations --
thanks to a pair of arguments inside each WINDOW_...() function --
they are of little use in this particular case, because the view
should have a granularity (Visual Level Of Details, or VizLOD)
of [Cust No] AND [ISO Week] (those Dims should be on Details).
A workaround exists, but it needs some preparation of the datasource.
Namely, a Range Join to the ISO Week (calendar) table is required.
With the HyPer engine (as of version 10.5), non-equi joins are possible,
so I've prepared a distinct datasource for this view using a Join Calculation
and a supplementary Filter -- to work around the Tableau HyPer engine restriction
on using the single field twice in Join conditions (even if they are Join Calculations).
Essentially, a Range Join is multiplying rows (of the original datasource) --
making possible to aggregate them (group by the ISO Week calendar table column)
to get the same result(s) as the 'Moving Window' calculation(s) would do.
This is an old-school (pre SQL-2003) approach to either Running or Moving calculations.
Please find the attached.