Patrick O’Beirne

Managing Director at Systems Modelling LTD

Patrick O’Beirne is the Managing Director at Systems Modelling Limited ( His current focus is Excel/VBA development for data analysis and spreadsheet auditing services as well as training courses like – “Spreadsheet Safe”(tm). He is chair of the European Spreadsheet Risks Interest Group (EuSpRIG) and a Fellow of the Irish Computer Society.

He has advised two certification bodies on the content of a syllabus for a new certification in good spreadsheet practices.

He is the author of “Spreadsheet Check and Control” (Systems Publishing, 2005, ISBN 190540400X) and the developer of the XLTest spreadsheet auditing Excel addin.

IT Under The Radar; Are Your Company’s SpreadSheets Tested As They Should?

Software testers conventionally work on line-of-business software produced by professional developers. Yet there is a largely untapped opportunity to apply the special skills of software testing to add value to the business by finding the low-hanging fruit of defects in spreadsheet-based information systems. Excel spreadsheet formulas are programming logic, and often they contain Visual Basic for Applications code which automates their operation further. They are created by people who are not software professionals, and so typically do not follow software engineering practices (e.g. review, testing, audit trail, change control, security). Yet the users trust them enough to apply them to business-critical decisions and reporting.

This session provides both anecdotal horror stories and formal metrics to illustrate the real extent of spreadsheet errors, such as the one that cost Tibco shareholders $100M, or the Value-at-Risk model that masked a JP Morgan $2Bn loss.

The session shares how professional spreadsheet auditors and testers find defects. It covers how to assess risks and complexities in the context of spreadsheet criticality and a process of high level and detailed reviews. It gives tips on how to recognize the “bad smells”, the symptoms of the most common defects, before the user has become aware of the defect materializing in a failure.