Hi,

Just here in Boulder loving the turn to warmer weather. If February was the new March with lots of record highs, what to think of this weekend in the mid-70s and sunny? It’s a good thing we had such a strong December and January with tons of powder days on the slopes!

Curious to hear your feedback about the topic below.

Best,

David

I created a short video where I describe some unique features of a financial model that I developed for a client that raised over $350 million. You can see it at brodetelecom.com/model-overview.

Turning Modeling from a Craft to a Manufacturing Process

Every so often I run across a new piece of software which starts off its pitch by denigrating the competition: spreadsheets.

I remember getting marketing materials from Sinper in the 1980s telling me to throw away my spreadsheet and use TM/1, the new “relational spreadsheet.” Well, that didn’t happen, even though TM1 grew into an amazing package that was acquired in turn by Cognos and then IBM, where it survives to this day.

When I worked with TM1 most heavily (approximately from 1995 to 2002) they had changed tactics and decided to be a power tool which you could attach to Excel. Later it became the analytic engine driving budgeting and reporting packages. And certainly we didn’t all throw away our spreadsheets, but they influenced the development of pivot tables in the Excel we enjoy today.

For ease of prototyping, Excel is unparalleled. For a quick and dirty analysis that you need to finish before lunch, why would you use a different tool? But some Excel spreadsheets live long lives and are used to guide important decisions. Here I’m thinking about strategic planning models with a multi-year time frame. In more complicated business environments the models grow in complexity as well and have thousands of inputs.

Much has been written about Excel’s problems and many of the techniques I’ve developed, discovered, or co-opted have been in response to those problems. You probably recognize many of these issues:

• How do you separate input data from calculation logic?
• How do you ensure that income statement, balance sheet, and cash flow statements form a coherent whole?
• How do you display months, quarters, and years elegantly?
• How do you ensure you’re grabbing the “right” Total Revenue number and not one that is missing some parts?
• How do you show the evolution of assumptions over time in a model?
• How do you keep track of data for multiple scenarios?
• How do you replicate and maintain sections of the model which are supposed to be identical but for their inputs?

In the end, I believe that I can build world-class financial planning models and help companies raise money and make the right decisions, but there is a limitation here. I’ve come to see modeling as similar to a craft. A master craftsman can make a beautiful cabinet, but it takes time. Most woodworkers cannot produce the same quality of work, and it’s hard to train new masters.

A better analogy might be with custom ski boot inserts. I got a pair eight years ago and it profoundly changed how I could feel my skis and the snow. It was as incredible as the first time I ever clipped into a road bike and felt truly connected to the machine. At one time, such inserts were the work of some tinkerer who looked at your foot and heard your complaints and tweaked until it was just right. One such genius master lived in Boulder and people swore by him. And then came Surefoot. There, a computer scans your foot and sends data to a machine which creates the footbed. The company might tweak it later, but basically they industrialized the process and allowed relatively less skilled workers to create a product that compares well to a master’s work, and with greater consistency.

With very complex strategic forecasts, Excel’s great advantages can become its biggest weaknesses. That ability to customize down to the single cell is great, but when you add $10M to a revenue scenario what-if and forget to remove it, well, that +10M is a needle in a haystack and you don’t even know you need to look for it.

So the dream is to turn modeling from a craft to a manufacturing process.

In next month’s newsletter I’ll discuss a new entrant that promises to do just that.