For over 100 years there has been debate among historians, and a few eccentric mathematicians, over the origin of the quote “there are three types of lies: lies, damned lies, and statistics.”
If you’re English, you could go with the 1st Earl of Balfour in 1892 or Benjamin Disraeli, a leading politician at that time. If you’re American, you’re probably thinking Mark Twain came up with it (he didn’t).
Either way, it’s time to update this clever phrase – given today’s technologies.
Spreadsheets, created on personal computers, now generate most of the financial reports and other more challenging computations used worldwide, including the statistics our forebears were so suspicious of.
Complex analysis doesn’t necessarily lead to better conclusions. Monster spreadsheets, with errors and miscalculations buried deep within them, can cause even the most sophisticated parties, like J.P. Morgan and Harvard economists, to make big mistakes, as highlighted by Paul Krugman in his New York Times opinion piece “Excel Depression.”
I believe there’s wisdom in our earlier methods of problem-solving.
Back of the Envelope: Before Spreadsheets
Prior to the 1990’s when laptops weren’t ubiquitous, you estimated first since a deeper analysis would involve laborious, long-hand computations using pencil and paper, maybe assisted by an abacus, slide rule or, eventually, hand-held calculator. Why waste time getting into the details if the ballpark result didn’t make sense?
The back of an envelope was handy because, before email, snail mail dominated, and there were a lot more envelopes lying around. Pencils were the go-to because keyboards hadn’t evolved from pianos and typewriters.
The simple beauty of estimating is that it forces you to prioritize the most important elements within your analysis. To quickly get a sense for what the answer should be. In doing so, you ponder the more essential factors first, before complicating the process with those less critical to the outcome.
In other words, you are less likely to miss the forest for the trees.
Personal computers, with their new software programs, arrived in the early 1980’s. Pre-Windows keystroke commands of early spreadsheet programs were intimidating, but we adapted. In the mid- 1980’s, as a young investment banking Associate at an earlier Bank of America in San Francisco, I started using Lotus 1-2-3, the first broadly adopted spreadsheet for PCs. It has long since been superseded, for most of us in the finance world, by Microsoft Excel.
Early spreadsheet use wasn’t so different from the back-of-the-envelope approach. We started with a blank spreadsheet and far fewer formulas than current software editions supply. This meant that the calculations needed within those spreadsheets had to be individually constructed by their builders.
This was still a great advance over pencil and paper, and the thought process behind building those spreadsheets was essentially the same as before. Because you had to physically tie together all cell relationships, you maintained the integrity of the evaluation process. Constructing your spreadsheet from the ground up also meant you were less likely to miss a step or take something for granted.
The current generation of bankers and business analysts start with spreadsheets when tackling their assignments. We need to be careful …
Errors and the Rise of Templates, Formulas, and Pivot Tables
Errors – Spreadsheet dependency and related input errors have become so commonplace that they’re now a research topic for universities including Dartmouth’s MBA program. A spreadsheet containing errors isn’t what you want to depend on …
Templates – Eventually we started storing our spreadsheets as “templates.” Why start from scratch when you can take an earlier work and adapt it? For investment bankers, this is often an “LBO model” for evaluating leveraged recapitalizations. Templates provide a quicker way to build an impressive spreadsheet, but they can be too convenient.
Formulas – Today’s spreadsheet software provides formulas most of us have never heard of, much less have a clue of what they do. Look on your computer. I have an Engineering degree and 35 years of investment banking experience, and I only recognize a few relating to finance. Which ones are your younger staff using? Are they in spreadsheets critical to your business? If so, you’re flying blind and do not know it.
Pivot Tables – Pivot tables let you manipulate large amounts of data, but they can obscure nuances across the data underpinning the analysis. Out of sight, out of mind.
Nonsensical results – Reference the wrong cell and the answer … isn’t the answer. The more complex the spreadsheet, the more off-base the results can be, even though the “spreadsheet says so.”
Hockey sticks – Spreadsheets make it easy to predict the future. Most of the financial projections I get from start-ups show explosive sales growth – a “hockey stick” when graphed. After 28 years and over $500,000,000 of venture-stage finance, I’ve yet to see a single case where sales projections were met. Consistent, more modest sales growth? Certainly. Exponential? Not yet …
Too many numbers – Numbers in a spreadsheet all look like … just numbers. Spreadsheets can generate a sea of numbers, but not all numbers deserve equal weighting relative to the decisions they support.
Wandering in the dark – Sometimes a young banker brings me some nuclear reactor of a spreadsheet, and it’s clear that something “doesn’t add up.” They’ve built a much more sophisticated spreadsheet than I ever could (or will), but it’s clear they don’t understand the problem we’re trying to solve. It makes for a teaching moment …
No … I use them daily, though I’ll probably never master pivot tables. Mine may be simple and ugly, but they do what I need them to.
Spreadsheets certainly have their place, if you first take time to consider what they should be telling you. Then you can go build your nuclear reactor.
If you are interested in direct private investments, I invite you to look here.
as published on Forbes.com
Lead photo by Belinda Fewings on Unsplash