Founder and Principal

 

 

September 2014 Vol. 3 No. 9

 

 

 

Visit the newsletter archive.

 

 

Goodrich & Associates
[email protected]
www.goodrich-associates.com
781.863.5019

 

We use models very day - more often, in fact, than we generally realize.

As a finance person, I use models to project financial statements for my clients, project weekly cash flow and more. When pricing contracts for customers, businesses model expected profits (or at least they should).

But models are for more than just finance. The set of expectations each of us has for how people and groups of people will react to our actions is the result of a model. Likewise, when you get in your car to drive to work in the morning, you use a model based on previous experience and an estimate of current conditions to gauge what time you need to leave your house.

Like I said, we use models every day.

But what makes a model a good model? The best definition I have come across is by Stephen Hawking in his book, The Grand Design. According to Hawking:

"A model is a good model if it:

1. Is Elegant

2. Contains few arbitrary or adjustable elements

3. Agrees with and explains all existing observations

4. Makes detailed predictions about future observations than can disprove or falsify the model if they are not borne out"

So, what does this mean for the financial models we use?

Here again, Hawking provides some guidance: An elegant model is as simple as possible, but not simpler.

For example, when projecting monthly financial statements, we could forecast cash receipts and accounts receivable balances by projecting forward all customer orders, shipments, invoices and when we expect each customer to pay each invoice. Very complicated.

Or, instead, we could employ the simpler (more elegant) approach of projecting monthly sales and Days Sales Outstanding (DSO) to project accounts receivable balances. This second approach has far fewer elements than a customer-specific model would entail.

Remember, however, "as simple as possible" means the model can predict with sufficient accuracy for the model's intended purpose.

With Hawking's advice in mind, here then are six tips for better financial models:

1. Use the right model for the right kind of projection.
   
   A financial model that projects reasonably detailed monthly financial statements, including an income statement, balance sheet and cash flow, won't tell you if you are going to run out of cash next week. A 13-week cash flow model, on the other hand, would.
   
   When forecasting accounts receivable balances and cash receipts, the DSO model referenced above is great for monthly projections. Projecting near-term liquidity with a 13-week cash flow model, however, requires information regarding when big invoices will be paid. Here the DSO approach by itself, is too simple.
   
   At the other extreme, and when projecting for several years, and/or testing alternative plans, models based on a few inputs and ratios will be directionally correct and perhaps just as accurate 18 months out or so.
   
   Regardless of which model you decide to use in a given situation, wherever these models overlap in what they are trying to forecast, they should agree. So, for the 13-week cash flow model, a good check is to eyeball the DSO in the later weeks. For the model that projects detailed monthly financials, starting 12 months out or so, reasonableness checks with the ratios that might be used for the multiyear model - such as gross margin as a percent of sales, return on sales, and operating leverage - are good checks.
   
   When I worked in corporate departments, my first check on annual budgets was how much these aforementioned ratios change. I then looked for a valid business reason for the change (and often found only wishful thinking).

2. Use the proper level of granularity.
   
   In both making the projection and measuring the projection, it's important to determine the correct level of detail needed. For example, if a business sells to more than one channel or market, these channels may have significantly different payment terms.
   
   In this case, tracking DSO by channel will result in a better forecast and sharper insights - insights that might be masked by lumping the two channels together.

3. See if your projections worked.
   
   Performing a variance analysis after the projection period ends, will not only help spot operating problems and things for the business to adjust, it will also lead to model refinement.
   
   And while smart companies do such variance analysis for projections of operating results, they often neglect to do this for their other important models. For example, did the projection for a capital project return what was expected, and if not, why not? When a model is used to project costs when pricing, go back and spot check the actual cost for that product or service.
   
   Models can be tested without knowing the future. Using historical data, models should predict the past. Some call it back testing. Such back testing will also give us an idea of how accurate we should expect the model to be.

4. Don't confuse problems with the modeling tool with problems with the logic and data in the model itself.
   
   For convenience, many financial models are developed using electronic spreadsheets, such as Excel. While such models are relatively quick to develop and use, they are prone to fat finger errors and more. Computer-based models that rely on hard-coded logic and databases are less prone to these errors, but take more time to develop and cost more.

5. Get diversity in perspective and measurement.
   
   For centuries, most people thought the world was flat and that the universe revolved around the Earth. That is, of course, until men started observing the sky with suitable tools to make accurate measurements, demonstrating that a flat Earth in the center of the universe didn't fit the observations.
   
   For business, this same enhanced perspective often requires bringing in experience from people who have seen many different situations or who are just younger, newer and without bias regarding existing models.
   
   When I was in the containerboard business, for example (many years ago), we had a product known as Mist White, the mottled white paper you see on the outside of copier paper boxes. For years, the company had limited price increases because it didn't want to "kill the golden goose." I built a simple model of the competitors' mills and realized we had an inherent cost advantage. The competition had to charge a certain minimum price, higher than ours, given the market price of the two principal ingredients. The profit impact from the price increase was dramatic, all because of a fresh perspective in modeling prices.

6. Consider human biases.
   
   When comparing a model's predictions with actual results, remember that we humans are biased towards focusing on data that confirms what we think... and glossing over data that conflicts.

Clearly, models play an important role in our everyday lives, particularly in the world of finance. Apply these six tips to ensure that your models are powerful and accurate, to help you achieve the best business results possible.

Click here to share this newsletter with a friend or colleague.
Click here to visit our newsletter archive.

Goodrich & Associates is a management consulting firm. We specialize in helping our business clients solve urgent financial problems. Our Founder and Principal, Charlie Goodrich, holds an MBA in Finance from the University of Chicago and a Bachelor's Degree in Economics from the University of Virginia, and has over 30 years experience in this area.

To ensure that you continue to receive emails from us, please add
[email protected] to your address book today.

Goodrich & Associates respects your privacy.
We do not sell, rent, or share your information with anybody.

Copyright © 2014 Goodrich & Associates LLC. All rights reserved.

For more on Goodrich & Associates and the services we offer, click here.

Newsletter developed by Blue Penguin Development