Generate groups using previous rows

I think you meant ASCending sort.  Can you add an index column to your source data? If yes then you can use that when you add the group column.

let
    Source = Table.FromRows(Json.Document(Binary.Decompress(Binary.FromText("i45WCijKTylNLlEwVNJRCvcHEn4gwtBU38BY38jA0FQpVgeLokjXYCBpZKRvaIBHFdgoY0NijDIw1zewAKkyx6cKZJYlHlVgC4EqIK4iZJShIVRVLAA=", BinaryEncoding.Base64), Compression.Deflate)), let _t = ((type nullable text) meta [Serialized.Text = true]) in type table [Product = _t, Type = _t, Cycle = _t, Date = _t]),
    #"Added Index" = Table.AddIndexColumn(Source, "Index", 0, 1, Int64.Type),
    #"Added Custom" = Table.AddColumn(#"Added Index", "Group", 
each 1+List.Count(
    List.Select(
        Table.Column(
            Table.FirstN(#"Added Index",[Index])
            ,"Cycle")
        ,each _ ="YES")
    )
)
in
    #"Added Custom"

Note that your sample data is inconsistent.

Hi Anonymous 

are you looking for the following solution?

Measure = 
VAR _ActualDate =
    MIN ( 'Table'[Date] )
VAR _PreviousDate =
    CALCULATE (
        MAX ( 'Table'[Date] ),
        FILTER ( ALL ( 'Table' ), 'Table'[Date] < _ActualDate )
    )
VAR _Group =
    CALCULATE (
        COUNTROWS ( 'Table' ),
        FILTER (
            ALLEXCEPT ( 'Table', 'Table'[Product] ),
            'Table'[Cycle] = "Yes"
                && 'Table'[Date] <= MIN ( 'Table'[Date] )
        )
    )
RETURN
    _Group

With kind regards from the town where the legend of the 'Pied Piper of Hamelin' is at home
FrankAT (Proud to be a Datanaut)

Anonymous - Do you have or can you add an Index? Otherwise, is "previous row" defined by the date column? If there are no previous rows is the group 1?