This useful article published by CompSci, How XBRL is like the NFL, can help you avoid common recurring errors being made by filers learn about the nuances of working with the XBRL medium which have been pointed out in SEC staff interpretations. Here is a summary of the types of errors this blog post covers:
- Negative values
- Concept fact limitations
- Extension concepts
To that useful list I would add common errors I see in my observations of SEC XBRL financial filings:
- Roll forward computations need to work correctly. Clearly things need to add up, but this can be tricky to verify because XBRL calculations cannot be used to verify this. A roll forward is a reconciliation between two periods, you probably learned about these in school: BASE (beginning balance, additions, subtractions, ending balance). XBRL calculations cannot be used to prove this computation as they do not work across contexts or across dimensions. XBRL Formula can be used to prove this type of computation.
- Dimensional aggregations need to work correctly. An example of a dimensional aggregation is breaking down revenues by business segment or by geographic area. Each segment is a reported using a different dimension. Again, XBRL calculations do not work across dimensions. Again, XBRL Formula can be used to prove this type of computation.
- Financial integrity needs to be intact. For example, the line item for inventory reported on the balance sheet should tie to the breakdown of inventory within the disclosures. This is easy if things are modeled correctly and total inventory in both places use the same reported fact.
Eventually software will watch over the creators of SEC XBRL financial filings, helping to make sure these sorts of errors don't exist. Today's software is still maturing and does not commonly provide this level of functionality.
For more insights, see my document Modeling Business Information Using XBRL.