4 ms·

I've always wondered: these clunky Excel sheets are used at billion-dollar companies. Why don't they invest in building a sane web application (or similar) that

I've always wondered: these clunky Excel sheets are used at billion-dollar companies. Why don't they invest in building a sane web application (or similar) that can do all the validation and calculations in a much better looking interface that is suited for the job / application?

dmannorreys 8y ago
Some do, but the way of working can be very deeply rooted. I work with an in-house made DCIM whose primary competition is excel. Switching way of working and also migrating bad quality excel spreadsheets is hard, especially when the spreadsheets have diverged between sites.
commandlinefan 8y ago
> building a sane web application Every time I build a sane web application, the first feature I'm asked to add is: "can you add an 'export to excel' function?"
froindt 8y ago
>Why don't they invest in building a sane web application (or similar) that can do all the validation and calculations in a much better looking interface that is suited for the job / application? Inertia, training cost, and risk of failure. 1 year ago I made a very extensive spreadsheet to analyze planned production and forecasts at a factory. It took about 6 months after the logic was thoroughly vetted to transition it into an IT supported auto-update to a database as the primary project for 1 IT analyst. No pretty front end to modify production parameters or anything like that. If we transitioned any earlier, it would have been the typical "our user keeps changing the scope and logic of the application we're trying to build for them". This was probably the best way things could have reasonably gone. No long term spreadsheet usage, spreadsheet had good documentation, used tables, names ranges to make the formulas easy to read, etc. Excel can do a decent job at figuring out the required logic. But it is used as an everything tool where task specific applications would be best. And people only know enough to be dangerous
mr_toad 8y ago
The best reason is probably that the person creating the system doesn’t know at the outset exactly how, or even if, what they are building will work. Excel often gets used for prototyping new models. A new model for allocating overheads - try it in Excel and see what the effects will be. A new forecasting method - get a ballpark estimate of the impact on the bottom line in Excel. Once the model is in Excel it isn’t cheap or easy to get it out, so it often just stays that way, accumulating technical debt.

dmannorreys 8y ago

Some do, but the way of working can be very deeply rooted. I work with an in-house made DCIM whose primary competition is excel. Switching way of working and also migrating bad quality excel spreadsheets is hard, especially when the spreadsheets have diverged between sites.

commandlinefan 8y ago

> building a sane web application Every time I build a sane web application, the first feature I'm asked to add is: "can you add an 'export to excel' function?"

froindt 8y ago

>Why don't they invest in building a sane web application (or similar) that can do all the validation and calculations in a much better looking interface that is suited for the job / application? Inertia, training cost, and risk of failure. 1 year ago I made a very extensive spreadsheet to analyze planned production and forecasts at a factory. It took about 6 months after the logic was thoroughly vetted to transition it into an IT supported auto-update to a database as the primary project for 1 IT analyst. No pretty front end to modify production parameters or anything like that. If we transitioned any earlier, it would have been the typical "our user keeps changing the scope and logic of the application we're trying to build for them". This was probably the best way things could have reasonably gone. No long term spreadsheet usage, spreadsheet had good documentation, used tables, names ranges to make the formulas easy to read, etc. Excel can do a decent job at figuring out the required logic. But it is used as an everything tool where task specific applications would be best. And people only know enough to be dangerous

mr_toad 8y ago

The best reason is probably that the person creating the system doesn’t know at the outset exactly how, or even if, what they are building will work. Excel often gets used for prototyping new models. A new model for allocating overheads - try it in Excel and see what the effects will be. A new forecasting method - get a ballpark estimate of the impact on the bottom line in Excel. Once the model is in Excel it isn’t cheap or easy to get it out, so it often just stays that way, accumulating technical debt.