Spreadsheet testing -Evaluating potential deficiencies 2260



  • Our ercently listed company has recently begun testing spreadsheet controls with limited experience in the group.
    I believe as a group we have the technical know-how to perform the testing. What we lack is the experience in knowing how to then assess the findings and determine what is a deficiency and what is not.
    My concern is largely around knowing how to address the ‘issues’ we have found thus far.
    I have listed them below, and would greatly appreciate guidance on this before making a fool of myself by sharing these with the rest of the team
    SPREADSHEET A
    -no documentation
    -noticed a tab with subtotals that do not include the entire row (in the month under review the excluded cells included nil values so this happened not to have any financial impact)
    -this spreadsheet essentially aggregates data from various divisions for review of aging at the company level. One divisions data is hard coded instead of updatd by formulas…My concern is the increased risk of error from manual updates, butis this really an issue
    -one check digit had an error result in it…nobody really knew why…My concern is that management is either not testing or not using check digits as they should or this would have been resolved

    • it appears the formulas arent being tested, however it is a fairly straightforward calculation anyway…is routine testing of formulas mandatory nonetheless?
      SPREADSHEET B
      This spreadsheet was deemed key, however it is just a listing of the output of macrs and other calcuations in another spreadsheet
      -should we change the ‘key’ spreadsheet tothe other undeerlying spreadsheet or would it be more appropriate to just test the original spreadsheet for data integrity since JEs are made based on that ‘output’ and proceed to test the underlying spreadsheet as I would any other critical spreadsheet
      -with the underlying spreadsheet it includes multiple macros. Is it sufficient to discuss and obtaine vidence of the review , testign andc hange controls around these macros or would we be expected to go as far as reviewing the code line by line and essentially testing it ourselves. We dont want to beat it to death, but we would like to meet recommended minimum standards.
      -the original ‘output spreadsheet’ includes some manual entries that are populated if for some reason the program malfunctions and does not bring some items across drom the source database. Is this an issue in itself? Or not?
      SPREADSHEET C
      Again this spreadsheet contains data that has been exported from an external database. How far do I take the testing to trace back to source? Is it sufficient to just pick one line item and make sure it was pulled correctly from the source or should I sample based on the no of items in the document?[/list]

Log in to reply