A new study is shedding light on the dangers of using spreadsheets for decision making, finding 94% of them contain errors.

The study’s authors set out to determine what risks businesses and organizations faced as a result of the widespread use of spreadsheets across industries.

Since their inception, spreadsheets have been widely used in many scientific and mission-critical applications such as linear programming and regression, control systems, statistical radiobiological evaluation, drug-alcohol interaction, patient specimen collection, neurosciencea nd nuclear fuel production. Many decision support systems (DSS) in the commercial and social sectors are built on spreadsheets to generate useful information for strategic decision making.

Unfortunately, businesses are relying on information that is riddles with errors.

Along with the high popularity of spreadsheet applications, it was found that about 94% of the spreadsheets in use contained faults.

The authors go on to highlight the potential issues associated with making decisions based on faulty data.

Faulty spreadsheets could result in business risks including: (a) loss in revenue, profit, cash, assets, and tax, (b) mispricing and poor decision-making, and (c) financial failure. If faulty spreadsheets are used in critical areas such as clinical medicine and nuclear operations, catastrophic consequences may result. Hence, spreadsheet quality assurance (QA) is a serious issue that cannot be ignored.

The study found that one of the biggest causes of errors in spreadsheets is the low barrier-to-entry for their use and programming.

A major reason for a high number of faulty spreadsheets is the accelerating trend in end-user computing (or end-user programming) over the last few decades. Spreadsheet development, now a prominent example of end-user computing, has shifted from being often done by well-trained IT professionals to something millions of non-technical departmental end users or end-user programmers are now responsible to do. As most end-user programmers are not well trained in software development and testing, it is not surprising that many spreadsheets they developed are poorly coded and inadequately tested. Consequently, these spreadsheets are likely to contain faults that are not properly detected and removed before release for daily operational use

Could This Portend Issues With Low-Code/No-Code Programming?

In many ways, spreadsheets are the original low-code development model, giving non-technical users an easy way to manipulate data and create what is essentially very basic programs. Unfortunately, as the study discovered, without a more thorough understanding of programming concepts and techniques, fundamental issues quickly become apparent, such as spreadsheets that “are poorly coded and inadequately tested,” and ultimately contain errors.

While some low/no-code platforms provide better guardrails for new users than spreadsheets, the potential for inexperienced users to create poorly designed applications is greater than ever. What’s more, increasing cybersecurity place greater emphasis than ever before on good development practices, something that many of those using low/no-code tools lack.

The above study illustrates the need for companies and organizations to devote resources to proper training of their employees, even for things that may at first glance seem designed for use with very little training.

The study was conducted by Professor Pak-Lok Poon, Man Fai Lau, Yuen Tak Yu, and Sau-Fun Tang, from School of Engineering and Technology, Central Queensland University; Department of Computing Technologies, Swinburne University of Technology; Department of Computer Science, City University of Hong Kong; and The Royal Victorian Eye and Ear Hospital, respectively.