Your <some text to be ignored> didn't appear in your sample doc. If you provide some real examples of the text to be ignored we can offer some other methods of dealing with them. I still think that it can be a ConvertToTable method but we need to identify the pattern of correcting that extra paragraph of data.
Perhaps this code will sort out the columns correctly. If that works it should be easy to remove the extra columns either before or after making it a table.
Code:
Sub MacroTest()
With ActiveDocument.Range.Find
.ClearFormatting
.Replacement.ClearFormatting
.Text = "(Cake description:•)^l([!^13]{1,})^13([!^13]{1,})^13Color"
.Replacement.Text = "\1^p\2^p^p\3^pColor"
.Forward = True
.Wrap = wdFindContinue
.Format = False
.MatchCase = False
.MatchWholeWord = False
.MatchAllWordForms = False
.MatchSoundsLike = False
.MatchWildcards = True
.Execute Replace:=wdReplaceAll
.Text = "^l"
.Replacement.Text = "^p"
.Execute Replace:=wdReplaceAll
End With
Selection.WholeStory
Selection.ConvertToTable Separator:=wdSeparateByParagraphs, NumColumns:=10, NumRows:=3, AutoFitBehavior:=wdAutoFitFixed
End Sub