I'm Brendan Hyland. I help regulated facilities transform their software, spreadsheets, workflows and documents from time-consuming, deviation-invoking, regulatory burdens, to the competitive advantage they were meant to be. Join me every week as we take a few minutes to explore, design, test and improve the critical systems we use in our facilities.
Is that validated spreadsheet a finished product, or just a well-tested prototype?
|
I’ve seen this pattern repeatedly: Instrument data is saved as an Excel or CSV. Data is then copied to a bare-bones spreadsheet with several columns of calculations or transformations. Results from the spreadsheet’s calculations are copied for use further down the data analysis pipeline. And I, the auditor, get handed a signed and dated pdf of the worksheet. Ok, they are rarely this bad - I created an extreme example here to prove a point - but I've seen several cases that were not far off! The printout has column headers like “Data 1” and “Calc A”. No way to tell for sure where data came from, or what was pasted versus calculated. It has some cells in the data column that are blank, and Excel errors in the calculation column wherever the data was entered as “ND”. The Study Director assures me that the spreadsheet was validated, and the input data passed a full QC. The calculations were proven to work. Validation deliverables are on file. SOPs exist. What’s the problem? Here’s the thing. The spreadsheet got validated when it worked. But nobody asked if it was finished. Its job was more than just to calculate the results - it was also supposed to do things like maintain data integrity through input and output, standardize analysis, facilitate review and audit, act as a traceable record. Without meeting all those requirements, it’s a well-tested prototype that went straight into production. A few hours of finishing work would fix most of this:
Not validation work. Finishing work. These aren’t complex fixes - they’re just often overlooked because spreadsheet validation focuses on calculations, not completeness. I wrote about these fixes (and three others) in a guide I just released. If this ‘prototype in production’ scenario sounds familiar, you might find it useful. You can download the guide for free at: https://daily.haiqu.ca/csrr Until next time, thanks for reading! – Brendan p.s. Enjoy this message? Read more at the Hyland Quality Systems website. |
The HaiQu Newsletter
I'm Brendan Hyland. I help regulated facilities transform their software, spreadsheets, workflows and documents from time-consuming, deviation-invoking, regulatory burdens, to the competitive advantage they were meant to be. Join me every week as we take a few minutes to explore, design, test and improve the critical systems we use in our facilities.
Calling quality people everywhere! I've just released something I think you'll find useful. Over the past 20 years working in regulated environments, I've seen the same spreadsheet problems show up again and again: A quick Excel tool gets built to solve an immediate need. It works. Gets "validated" with a few hand calculations. Then gets reused and modified for different datasets or slightly different purposes. Eventually - sometimes months or years later - someone discovers an error. Or an...
I'm thrilled to announce that I'll be presenting at the CCSQA/NERCSQA Joint Annual Meeting in October! The conference is 2 days, Thursday & Friday, 16-17 October 2025, in Laval, Quebec, Canada. On-site and virtual attendance available. Here's the announcement link: https://sqa.org/CCSQA/CCSQA/Events/Upcoming_Events.aspx My session is on Friday morning, and titled "The power of a specification: Freeing your creative self to go beyond compliance." Here's the abstract: As busy quality...
It’s the first step of the problem solving framework that I was taught back in Engineering school. Not ‘Plan’. Not “Define”. “I want to and I can”. That particular framework - the McMaster Six Step - never gained the popularity of the ones now used today, but in the end they all contain the same basic elements - research, planning & design, implementation, evaluation and iteration - just stated in different ways. However I’ve never really seen this particular element called out explicitly...