Building a Financial Model Around How the Business Works
Marquee believes that financial models are the most important decision-making tools in modern finance. But to serve their critical purpose, models need to have strong explanatory power. The model needs to “tell the story” of the business rationale for the projections, and therefore the conclusion, generated by the model. This is a fine balance. In our consulting practice, we help our clients develop a “balanced” level of detail for assumptions – to achieve explanatory power without the model becoming too convoluted or hard to manage.
For the model to achieve explanatory power, we believe that each line on the financial statements needs to be constructed in a way that reflects how the business actually works. Of course, there’s always a “quick and dirty” way of doing things to obtain a set of projections but doesn’t help to explain WHY the projections are what they are.
Use our checklist to ensure you are thinking about things constructively:
Use a growth rate.
Issue: Anyone can pick a growth rate. But as soon as you tell someone what growth rate you assume, the next question is “why that rate?”
Price x Volume.
Rationale: Every business’ sales are driven by these two levers. Modeling this way allows you to explain and sensitize what’s driving the growth.
Bottom line: Using price x volume will require a bit more work but will give the model more value as a decision-making tool. Sometimes, it can be challenging to figure out the right inputs for price and volume, but there’s always a balanced metric that can be found. Some examples include:
|Revenue per sq ft
|Square footage of retail space
|Revenue per seat mile
|Seat miles flown
|Rate per hour
Use a margin.
Issue: People think in terms of Gross Margins and EBITDA margins all the time, but we need to be able to explain why they may change from one period to another.
Separate fixed vs. variable costs.
Rationale: Fixed costs change through time on a total dollar basis and should therefore be modeled as such. But variable costs change largely because of sales volume, so it makes sense to model these costs first on per unit basis then multiply by sales volume.
Bottom line: We need to be able to explain why we think margins will change over time. Is it because the company is expected to be able to spread more revenues over fixed costs and therefore benefit from “operating leverage”? Or are inputs costs changing to offset this impact? Note that sometimes, costs don’t neatly fall in one category or the other. Some costs are essentially semi-variable – i.e. they are fixed for a period of time and then might have a step function as the company grows. You can use a lookup function to model this type of relationship.
Use a percentage of sales or Capex.
Issue: The rationale here is that as a company grows, so should its asset base and therefore its depreciation expense. But what if the Capex profile will be volatile or different from the past?
Separate existing assets vs. new assets acquired.
Rationale: By separating the current base of existing depreciable assets and assuming those have a certain remaining life, and then treating each amount of future Capex as a separate calculation based on timing, this better approximates the overall expense.
Bottom line: Building a depreciation “waterfall” doesn’t take long and it’s more helpful in explaining expected depreciation. It’s important to make sure your formulas are dynamic enough to ensure new Capex doesn’t begin to depreciate before the period of purchase, and that you don’t over depreciate any assets. Also, keep in mind that you may need multiple waterfalls from different types of depreciable assets (i.e. buildings vs. computer software) as the useful lives and depreciation methodologies can vary.
Use a blended principal amount and rate.
Issue: This of course will be quick to build, but a company may have a big refinancing coming up on an individual piece of debt that needs to be modeled. Or the credit ratios could be very sensitive to rates so we need to be specific.
Build a schedule for each piece of debt.
Rationale: Every piece of debt has its own features (rate, amortization, term etc.). Modeling each piece separately lets us understand how he capital structure evolves over time.
Bottom line: Sometimes, building numerous debt schedules may not be practical. But where time and information is available, doing so gives the model more flexibility and greater visibility into financing decisions into the future. In addition, sometimes it may be helpful or necessary to model accrued interest vs. cash interest payments for each piece of debt – this is especially important when doing monthly budgeting.
Use a simple tax rate.
Issue: A simple tax rate will often miss the impact significant differences between accounting and government tax calculations.
Model timing differences and tax attributes.
Rationale: Tax rules in a company’s main operating regions may cause cash taxes to deviate significantly from taxes for accounting purposes, so it’s important to measure this impact.
Bottom line: While it’s important to understand income tax expense for accounting purposes (to project earnings and EPS), it shouldn’t end there. Most financial models need to track cash flow for valuation and / or credit purposes – so to be accurate we need a sound estimate of cash taxes. We need to be careful not to go into too much detail here as tax can be extremely complex. We recommend modeling those timing differences (i.e. depreciation for capital intensive companies or stock-based compensation for higher growth companies) and tax attributes (i.e. significant tax losses or credits) that have a material impact on cash flow.
Use a percentage of sales or assume no change.
Issue: There may be value in assuming a change in the efficiency with which a company manages its working capital. This is hard to explain with simple percentages.
Model based on days.
Rationale: In real life, we always think about working capital in terms of days (i.e. how many days to collect A/R). The math is very simple and it’s a more meaningful way of analyzing WC.
Bottom line: When you model based on days, it becomes very easy to translate efficiency changes into the model assumptions. Going back a few periods to calculate historical days trends is often helpful in understanding how a company has been able to manage their working capital accounts. If you’re working on a monthly or quarterly model, make sure to adjust for the correct number of period days in your calculations.
When we build models at Marquee, we spend a lot of time up front in the planning and information gathering phases before we even touch Excel. We always keep the above concepts in mind when trying to achieve that important balance between detail and usability. It may take more time, but it’s worth it in that it inspires debate and increases the explanatory power and value of the model.
Feel reach to reach out to us if you want to discuss applying these ideas in your own models.