Here at Gridlines, we are always banging on about spreadsheet risk and the risk of errors in Excel.
Usually, we keep it light:
Or perhaps even…..
“Bargain Booze’s spreadsheet hell sinks the business”.
All quite amusing, unless you are the poor modeller in the eye of the storm.
However, you can’t have missed that this week the debate on Excel risk got very serious indeed.
So what went wrong?
Public Health England had set up an automatic process to paste test data into Excel spreadsheets so that it could then be uploaded to a central system. This would then be made available to the NHS Track and Trace team, whilst also being viewable on relevant government computers.
PHE had chosen to use the (.XLS) file format in Excel which has a limit of 65,000 data rows, once the rows became full, any cases which the automation tried to add simply disappeared.
If your modelling error makes it on to the front page of the Daily Mail, you know you really have got problems.
The impact of errors in Excel can be massive. Usually, they result in faceless bankers somewhere losing millions of dollars. With this one, however, there could be a real impact on people’s health.
Is Excel the right tool for the job?
Excel is not the perfect tool for this application. For storing large quantities of data, there are far better, purpose-designed database applications out there.
But……
When you do not have the luxury of time for implementation, if you are looking for a flexible tool which is universally available and can be used very quickly to solve an urgent problem, Excel is the default solution.
The problem here though was not the use of Excel.
It was the misuse of Excel.
Treating Excel applications with respect
If you are using Excel in a critical application you need to be aware of the risks and take steps to manage those risks. Developing a tool in Excel should be treated with the same rigour as any other software development.
As a minimum, you need to ensure that you are using well-qualified, experienced staff that are aware of the functions and importantly, the limitations of the tool.
You need clarity of ownership and accountability and a robust review and sign-off process.
Frustratingly, the UK Government has published some excellent guidance in managing risk in its “aqua book”. The challenge is implementing the guidance and ensuring that in the heat of the moment, at least the key principles are retained.
But we still love Excel
Here at Gridlines, we love Excel and really hope that the tool is not unfairly blamed for these errors. We hope that this is a catalyst for a renewed focus on proper process and procedure in its use both in the UK public sector and beyond.
Our team believe that the powerful combination of training, a standardised approach to model development, and independent model assurance can help to greatly reduce the risk of error.
We are passionate about helping our clients to use Excel well. If you would like to hear more about how we manage modelling risk, please get in touch.
