Start With the End in Mind

Starting with the end in mind is one of the simplest ideas that is frequently ignored. I have seen so many analysts begin designing a financial model without having a clear understanding of the purpose of the model. A simple example may make this clearer. 

When I was a young financial analyst in the mergers and acquisitions department of the largest company in its industry. One of my primary jobs was to prepare valuation models. Sometimes, we were in the preliminary phases of an acquisition, researching the target and the competitive landscape. In those scenarios, my goal was to establish an acquisition purchase price that was a good deal for our company. 

Many times, however, the deal had already been agreed upon by the CEOs of our company and the target company. They had roughed out a price based on a few simple rules of thumb like revenue or cash flow multiples. If that was the case, my goal was simply to build a financial model that resulted in a valuation in line with what was already agreed upon. To be clear, if the model and resultant valuation showed the agreed upon price was a bad deal, our CEO would use that information to renegotiate the deal, but usually we just built the model to hit a target outcome. 

Understand your end goal before you start building any financial model or you may spin your wheels and end up rebuilding it.

Types of Financial Models

There are a wide variety of financial models you could use. I’ll focus on just three, and when you should consider using each.

Judgement-Based Models

Judgement-based models, sometimes referred to as human models because they rely on human judgement or expert judgement models, are frequently used in situations where resources make it impractical to use other models and where experts are able to provide the requisite inputs for the model. In this type of model, you as the modeler would ask the expert for the value to input in your model. It’s simple and quick, but you better have a good expert or your results may not be very useful.

Statistical Models

Statistical models are fueled by historical data and statistical analysis. Regression models are an example of statistical models and spreadsheets have functionality that allows you to do regression analysis as part of your financial modeling. 

Risk management professionals frequently use this type of model, as do weather forecasters. When used properly, these models can be very accurate and reliable, but you always need to keep in mind that there will always be unforeseen circumstances that cause variability.

Driver-Based Modeling

Driver-based modeling is characterized by using formulas that rely on a thorough understanding of the relationship between the independent and dependent variables used to model outcomes. Driver-based models make what-if analysis a breeze when they’re designed properly.

Designing Models: A Few How-Tos

Here are a few of my favorite tips to keep in mind when you’re designing your financial models. Organization and documentation are key factors in effective model design because if something is complicated enough to require a model, it’s likely the model will get complicated enough to get unwieldy if you aren’t careful. 

Organization

Organize your assumptions in one place and make sure you explain them clearly. Assume someone will have to use your model without the knowledge you had when building it, because that’s likely to happen. You will eventually get promoted, change jobs, or leave your position for a variety of other reasons, so plan for this eventuality by documenting your model well.

Number Every Line, Even Blank Ones

One way I have made things easier on myself and others when designing models is to simply number every line item in the model, including blank rows. Yes, spreadsheet programs have rows and columns identified in the program on screen, but when was the last time you had to refer to a report without those to reference? It’s much easier to direct someone to line eight instead of trying to estimate how far down the page to focus someone on some obscure expense line item, like COS-Other?

Range Names Are Your Friend

When using spreadsheets, use sheet and range names and make them descriptive. A simple example will say all that need be said about this. Which of these formulas is easier to read and understand?

=Assumptions!L6*IF(Allocations=0,Allocations!$L$7,Allocations!$M$7)
=Sheet3!L6*IF(B4=0,Sheet5!$L$7,Sheet5!$M$7)

You can tell just from the names used that this calculation involves allocations calculated based on certain assumptions on another tab. That is useful information when you’re reviewing a spreadsheet and you don’t want to spend a lot of time just understanding the basic flow of the model.

Technologies Role in Financial Modeling

Many people are still using spreadsheets for financial modeling. Consider when it may make sense to transition your models to a different tool. Large-scale modeling in spreadsheets may become problematic for a variety of reasons. There are lots of tools available that have been purpose-built for financial modeling that are improvements on the standard spreadsheet tools most modelers use and many offer free trials, so do your homework and choose the right tool.

By knowing your goal and understanding the appropriate type of model to build you’ll be off to a great start. Follow some of the basic how-tos I shared and you’re on your way to improving your model building.

Why Use Driver-Based Planning

Driver-based planning is characterized by using formulas that rely on a thorough understanding of the relationship between the independent and dependent variables used to model outcomes. A simple example of driver-based modeling is forecasting sales by using a rate multiplied by units formula. A big strength of driver-based models is that they allow for quick and easy what-if analysis. All you need to do to see the results of different scenarios is to change a variable and the results flow through your model to let you see the outcome. These models provide great flexibility, but they require a deep understanding on the part of the modeler of how the variables affect each other. These models can also require more time up-front to develop than judgment-based or statistical models.

When creating a driver-based model it is important to produce one that is not overly detailed or complex, but one that is accurate and actionable. Avoid complexity by adding variables if they do not provide analytical benefits. Starting with the chart of accounts is not a good idea. The model should focus on key performance drivers. 

Identifying Key Drivers (80/20)

Every planning professional I’ve ever met is familiar with the Pareto Principle, also known as the 80/20 rule. It simply says that roughly 20% of inputs account for 80% of outputs. The interesting thing is, most people I’ve met routinely ignore this principle. 

A little analysis will tell you where the 20% are in your business. This will then be your focus when you’re designing your driver-based model. You may not need to use driver-based planning for every item you want to forecast. Let your 80/20 analysis steer your modeling. Your planning should integrate the operational elements of the business, not just financial data from the general ledger. Many of the drivers you will want to evaluate include non-financial, operational, variables. Here are some examples of financial and operational drivers:

Examples of Financial Drivers:

• Price
• Volume
• Margin
• COGS

Examples of operational drivers:

• Call Volume 
• Employee turnover
• Service Level: 
• % of orders fulfilled

One of the best ways to gain a deep understanding of the operational drivers is to get out of your office and get into the operations of your business. See how the business operates. Observe first-hand what makes the needle move in the business. When I had to develop the operating budget for the reservations department at Royal Caribbean Cruises, I spent some time observing reservation agents taking calls and booking cruises. This gave me a good understanding of what variables affected their call volume, average talk time, and a variety of other variables. This helped me immeasurably when I built a driver-based model to develop the reservations department’s budget.

Steps to Building a Driver-Based Model

Remember Pareto

The first thing I would do when building a driver-based model is the 80/20 analysis. First, figure out what you are going to model because it is not going to be the entire chart of accounts. 

Discovery

Once I know the output I’m looking for, it’s time to go into discovery mode. That’s where the operational experience I mentioned comes in. Go talk to the people in the departments involved in the things that drive the outcome you’re planning for. If you’re forecasting sales (in dollars), go talk to the vice president of sales and some managers in the sales department, maybe even some salespeople who are the ones racking up the sales. Gain an understanding of their sales cycle. What are the things that affect prices, discounts given, receivables terms, chargebacks/returns, sales volume, the timing of various promotions, etc.? These are all things that can affect your projection, so you better understand them. 

Put the Variables to Work

Now it’s time to build your model. Be organized and develop your formulas based on what you learned in the discovery process. Let what you learned to drive the level of detail in which you build your model. For example, if sales prices don’t vary, it’s sufficient to set up a price list and use that to drive all the formulas that involve sales price. If, on the other hand, your business runs lots of promotions or routinely gives discounts, you’ll want to build your formulas to incorporate these variables so you are not constantly reworking the model when things change. That’s the whole purpose for using this method, flexibility.

Using Technology to Enhance Driver-based Planning

There is a wide variety of technology tools that incorporate functionality that makes driver-based planning easier. Many of the newest planning software packages allow you to use formulas that are very similar to spreadsheet formulas, making the learning curve short and many planners still choose to stick with spreadsheets. If you go this route, you need to be aware of spreadsheets inherent limitations and proceed accordingly. Broken links, formula errors, data integrity and security issues abound in spreadsheets, so if you are doing serious planning, it is worth evaluating some of the other tools that are available.

Summary

My preference for driver-based planning stems from years of using them and understanding their strengths. It takes an upfront investment of time and energy to learn your business well enough to build a well-designed driver-based model, so it may not be the right choice for your purpose. If you go this route, though, it pays big dividends over and over again. 

University of Pennsylvania researchers report that the human retina processes data at approximately 10 million bits per second. That’s awfully fast compared to a typical adult reading rate of about 400 to 600 words per minute. There are many advantages to visualizing data, aside from speed.

Why data visualizations?

When we view visual imagery instead of only numbers, our minds are able to make sense of information more quickly than if we were just looking at numbers by themselves. I’m not talking about simply perceiving the existence of the information, but actually understanding its meaning, in context.

This is important for reasons beyond expediency. With all the data most business professionals deal with, the ability to communicate a lot of information quickly and effectively is very valuable.

That old saying, “a picture is worth a thousand words” hits the nail on the head when it comes to data visualization.

Visual imagery helps us identify relationships more easily. Whether it’s the parts of a whole, like in a puzzle or a connection that is easy to see, but hard to discern using logic, visuals sometimes make relationships clearer. 

Seeing two people holding hands tells us something about the closeness of their relationship that would be hard to discern so quickly and clearly any other way. 

Another benefit of using visuals is that it allows us to find information that we might not otherwise have found because of the way we process visual information. Aberdeen Group research tells us that, managers who used data visualization tools were 28 percent more likely to find timely information than those who didn’t use visualization tools.

Overview of Gestalt principles of visual perception

If you read about Gestalt principles, you’ll probably run across this quote from Kurt Koffka a lot: “The whole is other than the sum of the parts.” It’s an overarching idea that encapsulates Gestalt theory and it’s a good thought to keep in mind when pondering visualizations of any kind. If your final product doesn’t improve on simply presenting data elements on their own, you’ve probably gone wrong somewhere.

There are six core principles of visual perception in gestalt psychology that apply to data visualization. They are proximity, similarity, closure, continuation, figure/ground, and symmetry/order. 

The concept of proximity says that when items are near each other our brains perceive they are part of a group. Similarity is the idea that when items look much the same, our minds perceive them to be of the same type. We naturally assume that shapes that look the same are related. The principle of closure describes how our eyes tend to add missing pieces of familiar shapes or patterns. If two sections are taken out of a circle, for example, people still perceive the whole circle. Continuation, also called continuity, is what makes people perceive objects as continuing to move in a certain direction when they see them moving. Depending on how people look at a picture, they see either the foreground or background as more prominent. That is the idea of figure/ground. Our minds tend to lean toward visually symmetrical things, so the principle of symmetry tells us to avoid giving someone a visual that is out of balance, or missing, or wrong whenever possible.

How Context affects visualizations

Context is information that helps viewers of your visualization better understand what they’re looking at. In the example below, the text “Record High” helps the person viewing the graph know that the data item shown is the highest ever recorded. Without this small bit of text for context it would be difficult to show this visually without showing the entire history of data for this item. Even then, the text might do the job better.

Adding context is often helpful, so consider when it can add to your presentation.

Here are three simple rules for adding context:

1. Brevity - Use as few words/letters as possible. Keep it to one line when possible. For example, use TTM instead of trailing twelve months if you can.
2. Simplicity - Use simple words or acronyms. For example, use vs. utilize.
3. Clarity - Be clear and specific. Don’t say good or bad if you can provide an objective and specific piece of information that speaks for itself.

Data visualization functionality is increasingly making its way into may technology tools. There are also purpose-built data visualization software applications you can use to create effective visualizations. Whether you opt for a separate piece of software or use the functionality already incorporated in tools you currently have, using the concepts discussed here will help you create more effective visualizations.

Most business people are familiar with the Pareto Principle, also known as the 80/20 rule. Simply put, it says that 20% of inputs generate 80% of outputs. It's a very simple concept, but many people struggle to use it. When it comes to communication, the practical application of the 80/20 rule is to listen 80% of the time.

If we are going to listen most of the time, we ought to do it well and that means using active listening. It is the most effective form of listening and it’s also the most challenging to use. Once you understand how to listen actively, you’ll start to understand why more people don’t do it all the time. It takes effort, especially when you’re first mastering the skills, but it has a huge payoff.

Let’s look at active listening as a process, step-by-step:

1. Mindset/Focus – The first step to active listening is to consciously focus on being an active listener. The act of listening actively cannot be contrived or the person to whom you are listening will know it and it will break your rapport and keep them from speaking openly and honestly. The importance of this cannot be overstated, as without rapport and trust the communication cannot be as effective as it otherwise could be.
2. Pay Attention – Listen to understand. Don’t get off task thinking about what you will say next.
3. Defer Judgement – Separate the person from the idea.
4. Feed it Back – Feeding back what you heard is different from just repeating what the speaker said. When you feed it back you put what the speaker said into your own words to confirm that you correctly understood, not only the words they said, but their intent, emotion and any other things that you sense are embedded in the words you heard. It is not uncommon that people use lingo, acronyms or words that are highly influenced by their culture. When you feed back what you heard, you are ensuring that whatever might have been unclear is brought out and clarified.

Active listening is often an iterative process. When you feed back what you heard the person may make clarifications, add on to what they just said or flat out correct a misunderstanding.

Going back through the process until you and the person you’re communicating with are satisfied that both sides are on the same page ensures you really understand and you have used active listening effectively.

One of the most common barriers to active listening is listening with an intent to reply instead of listening to understand. This is most people’s natural inclination for a variety of reasons and it makes you prone to being distracted trying to formulate what you will say next. Avoid this at all costs. 

A great way to get better at identifying and using active listening skills is to model people who are exceptional active listeners. This is easy thanks to the vast amount of video clips of great active listeners at your fingertips online. A quick search for a few names will give you lots of great examples. Some great examples of active listeners include Oprah Winfrey, Charlie Rose and James Lipton, to name just a few. 

Some common traits they share are their skillful use of open-ended questions, their excellent body language, and the way they give feedback. When you watch them, you will notice they ask a question and shut up. There are long stretches of people they interview talking uninterrupted. This skill is harder than it looks, as most people feel compelled to interject when someone talks for a while. Great active listeners revel in what they learn when they actively listen. They get information and stories out of people few others can. 

The principles of active listening can be applied by FP&A professionals using analytics by looking for the minority of analytic results that drive the small number of the most important decisions. This is what the 80/20 rule is all about and modern analytics tools make it even easier to spot the critical 20%. 

Conscientious application of the 80/20 rule can pay big dividends in your communication, so think more about how you can implement it in your day to day FP&A activities.

The article was first published in Unit 4 Prevero Blog

FP&A professionals work in an environment that requires communicating with a wide variety of people, from the CFO and possibly the organization's board of directors, to department heads and maybe even managers and entry-level professionals all across the organization. This requires us to be adaptable to communicate effectively with all these different audiences. There are some things we can do to ensure we are consistently and reliably on point by tailoring our communication to each audience.

Above all else, effectively tailoring communication is rooted in asking good questions and listening carefully to fully understand people's responses. Here are some of the questions you should consider asking:

• What is your audience's experience level?
• How familiar are you with your audience and how familiar with you are they?
• What is your audience's interest level?
• How much topic knowledge does your audience have?
• How diverse is your audience?

There are many more questions we could ask, but this is a good start and will give you some great information about your audience. While questions are the key to understanding our audience, they are only as good as our listening skills. If we don't listen effectively, even the best questions will bring us limited benefits, so let's talk about listening.

EFFECTIVE LISTENING TAKES SKILLS

We spend about 70% of our waking hours communicating and roughly 45% of that time is in listening mode, so let's make sure we're being effective with this time. Active listening is the type of listening we should focus on. There are specific steps to the active listening process, and once you learn and practice them you will become a much better communicator. This simple visual illustrates the active listening process.

• ​​​​​​​Mindset/Focus -The first step to active listening is to consciously focus on being an active listener. 
• Pay Attention - Listen to understand. Don’t get off task thinking about what you will say next.
• Defer Judgement - Separate the person from the idea.
• Feed it Back – Feeding back what you heard is different from just repeating what the speaker said. When you feed it back you put what the speaker said into your own words to confirm that you correctly understood, not only the words they said, but their intent, emotion and any other things that you sense are embedded in the words you heard.

Active listening is often an iterative process. When you feed back what you heard the person may make clarifications, add on to what they just said, or correct a misunderstanding.  Going back through the process until you and they are satisfied that both sides are on the same page ensures you really understand and you have used active listening effectively.

Now that you know how to listen actively, let's talk about what to do with the information you glean from asking great questions and being a great active listener. You’ve probably heard about The Golden Rule. Simply put it says, “Do unto others as you would like to have done unto you.” Most people think this is a good rule of thumb and intuitively, it seems like an idea with merit. The problem with this principle is people are different and not everyone wants their communication the way you want yours. Even gloves don’t fit two different people the same, they don't "fit like a glove" universally. 

THE PLATINUM RULE

A better rule to use in your communication is the Platinum Rule, which says "do unto others as they would like done unto them." In other words, give them communication the way they want to receive it. If they prefer lots of details, give it to them. If they prefer brevity, like most CFOs and boards of directors do, give it to them. Adapt your communication to your audience based on what you learn from asking great questions and being a good listener.

​​​​By practicing these few simple things, you will ask good questions, use active listening ​​​​to really understand your audience and tailor your communication to them for maximum impact.​​​​​​​

The article was first published in Unit 4 Prevero Blog