“To Excel” or “Not to Excel” – That Is the Process Engineering Question

Why “Not to Excel”

There is an over-reliance on spreadsheets for aggregating critical plant information. Understandable. Yes, Excel has its place in your software landscape, but there are some important limitations which means using it alone is not enough. Especially when dealing with industrial manufacturing data.

Here are some reasons spreadsheets alone are not enough when it comes to Big Data Analytics.

It’s not ideal for operational and time-series data and time-series analytics.

Looking for patterns in immensely huge data files is just incredibly difficult and time-consuming.  Additionally, you have a variety of data sources all containing critical information that remains unintegrated and messy. It becomes cumbersome causing major headaches.

Massive file sizes can create massive headaches.

A single spreadsheet has a limited amount of data it can easily handle. A spreadsheet with about 100 or even 200 megabytes of data, which is not a whole lot, can bring you trouble. Spreadsheets can become very slow and clunky, and once you finally have your huge datasets, not to mention the calculations on top, things can get worse. This can lead to data loss from crashes or workstation or laptop hang-ups which can cause big delays.

Ain’t nobody got time for that.

Next is the human error factor.

One of the biggest reasons spreadsheets shouldn’t be relied on solely for important operational decisions is the aspect of human error. A few years back, a Harvard study discovered that 88% of spreadsheets contain human errors. That means that only one in five spreadsheets will give you truly correct results.

If you want to incredulously laugh at how simple, man-made spreadsheet errors have cost companies serious money and resources, check these Excel spreadsheet horror stories.

There’s the lack of version control too. 

Version Control is another limitation. Even with the advent of cloud tooling or network storage which makes this less of an issue, colleagues will send a certain version over email. You then have to rely on somebody else to put it back on the network drive, so that sharing a version is not ideal.

Another frustration is that engineers often create different versions of the same file, so data gets lost. A huge bummer and a huge potential problem of inaccurate information.

It’s also not suited for real-time analytics.

Spreadsheets are also not particularly suited for real-time analytics. If you want to reduce the time between a process event and taking action on that real time aspect, Excel doesn’t do a good enough job. A common example is when you have to shoot out an alert. To do this, you have to have Excel open to see if the device is going to work or not. It’s clumsy and impractical.

And Excel’s limited sharing capabilities means lack of context.

If you use Excel for data analytics, you’ve learned sharing your work is not easy. But for fun let’s say you have a great share drive and file collaboration isn’t an issue. What is an issue is not having context behind what your colleagues were doing, why they were doing it, and what their findings may have been. Results may remain hidden and inaccessible to all of your team.

If, for example, a process expert opens a new spreadsheet to pull in data, access to all the analyses that maybe you or someone else completed before will not be available – the context of previous analyses will not be there. So, in essence, your process experts are working in the dark, repeating work, and wasting your time and theirs.  

Ain’t nobody got time for that, either.