All Models Are Wrong
Based on the number of responses I received, the idea that spreadsheets can be dangerous resonated. One of the more interesting takes came from Kyle, who described himself as having years of experience “on both sides of the spreadsheet argument.” He pointed out that “great people, great ideas, and great markets can overcome the worst spreadsheets.”
That context makes now a good time to point out that no financial model is ever correct. In fact, I’m pretty sure that it’s a law of the universe that once you forecast numbers in a spreadsheet those numbers will never ever come to be. So, to Kyle’s point, you want to make sure you are working with people, ideas, and markets that are likely to make your projections wrong about the upside, not the downside.
One of the first investments I ever made with my own money was buying stock in Whole Foods when I was in college. While I am loath to ever sell anything, I did eventually sell Whole Foods years later when the valuation got pretty crazy in late 2005.
For context, the chain at that time was growing quickly with tailwinds and had approximately $400M of operating cash flow against $4.7B of sales, for a pretty great (for a grocery store) cash flow margin of almost 9%. It was a good company! Investors thought so, too, because they had bid the value of the company up to over $10B, meaning shares could be bought or sold for 25x cash flow.
Now, I don’t know about you, but I invest to try to earn at least double-digit returns annually. For Whole Foods to deliver that on a cash flow basis on that valuation it would have to more than double in size while also maintaining best-in-class margins at the same time that copycats were emerging and competitors were catching up.
Was I able to make a spreadsheet at that time that made a case for Whole Foods doing just that? Of course I could. But as I stared at the numbers, it became clear that if that scenario were to come to pass, Whole Foods would have to sustain economics that no grocery store had ever achieved.
Reasoning through it I thought that might be difficult since Whole Foods was a grocery store.
Twelve years later Amazon acquired that grocery store for $13.7B. Whole Foods at that time had $16B of sales and $1B of operating cash flow. It had grown a lot and done great, but margins had contracted, not expanded (to look more like a grocery store’s) and its valuation multiples had compressed (to look more like a grocery store’s). In other words, holding Whole Foods would not have been a bad decision because it was a good grocery store but if I’d held from the point I sold to the point Amazon bought, I’d have earned about 3% annually. Not great!
This is an example where, to steal from Kyle’s framework, there was a spreadsheet and arguably great people and great ideas, but not a great market. And that’s the factor that ultimately carried the day.
I was talking to one of our operators recently and he pointed out that despite putting a lot of time, energy, and thought into his budget models, his projections had never been close, and we could laugh about that because reality had always been better. As I’ve reflected on why that is, and this is why I am shamelessly appropriating Kyle’s framework, it’s because the people, ideas, and market at that business have proved to create opportunities that didn’t fit in any spreadsheet, which is a fantastic problem to have.
– By Tim Hanson