THE STORY
A man walks into a home improvement store to purchase a ladder. He goes up to the salesman and asks for a recommendation on which ladder to buy.
The salesman replies, “That depends on what you plan on using the ladder for. We have tall ones, short ones, durable ones, inexpensive ones, and flexible ones. Tell me how you plan on using the ladder and I’ll recommend the right one.”
The man says, “I don’t know how it will be used. Heck, just get me a blue ladder. I like the color blue.”
THE ANALOGY
It’s hard to buy the right tool when you do not have a clue as to how it will be used. The same is true with financial models. You can design all sorts of different computer models for your strategic scenarios—simple ones, complex ones, flexible ones, and so on. Like the ladders in the story, each type of model has its place—and each type can be inappropriate under certain circumstances.
Models are like ladders; they are both tools to get a job done. The better you understand the job, the better choice of tool you will make. Choosing a ladder because you like its color makes as much sense as choosing a modeling technique because it is your favorite. Instead, get the tool based on the job to be done.
THE PRINCIPLE
The principle here is that the best type of model is the one that best answers the question at hand. Simple models have their place. They are quick and easy to build and easy to understand. However, they may not be adaptable to a wide variety of scenarios and they may not be sophisticated enough to provide a meaningful answer.
By contrast, a sophisticated and complex model can allow you to understand a situation more deeply. They can also be adaptable to more alternatives. Unfortunately, they can also be a time-consuming nightmare to build, debug, and input data. In addition, you may not have enough information to know how all the pieces in the model should interact.
Remember, a computer model is just a tool, not the end result. The end result is a strategic decision. Depending on what that decision is, different models may be more or less appropriate.
Keeping that in mind, here are my rules for designing models.
1) Start With the End
The first thing you need to do is ask what decision will be made based on the modeling. Knowing that (the end) will let you know where to start. I’ve seen cases where someone rushes off to build a model before fully understanding how the model will be used. They come back with something inappropriate. That is a waste of time for the modeler and the ones for whom the model was made.
If you are trying to decide between two different ways to operate your business (quality vs. low price; in-house vs. outsourced; mass vs. niche; automated vs. flexible; etc.) perhaps the best model is just a single look at each option in its mature state. There would be more complexity around the factors that are different in the scenarios and less complexity in the areas where they are the same.
If you are trying to decide whether to do an acquisition, then you probably want a model which spans several years—long enough to capture the value of the deal. You need enough detail to compute cash flow. Since a lot of the value of an acquisition is gained or lost during the transition process, you would want to model that as well.
If you are trying to choose between short-term tactics (like a pricing plan or an advertising plan), the model can probably be simplified to only looking at the areas of the business impacted by the tactic.
If you are in a crisis mode where a decision has to be made immediately, stick to the key issues and crank something out quickly.
2) Never Asssume You Will Get it Right In One Take
I worked with a guy who had an interesting take on model building. Once he got all the formulas right, he would run the model once and then freeze the results. In other words, he would erase all the formulas and links from the model and replace them with the actual numbers which came out of the first running of the model.
Then, he would present his results. Invariably, someone would want to adjust some of the assumptions in the model or try another scenario. This guy would then throw a fit because he had erased all of the formulas. He couldn’t run the model again because he had frozen each cell in the spreadsheet with the number from the prior scenario.
This is an extreme case, but the principle applies broadly. Assume that there will be future adjustments to the model. Build it with enough flexibility so that it can be adapted to the future changes.
3) The Questions You Ask Are More Important than the Model You Build
Models are built to provide more clarity around a business decision. Fuzzy notions are hard to quantify and even harder to properly evaluate. In the process of building a model, one has an opportunity to help your audience become less fuzzy by asking a lot of questions.
A computer spreadsheet model has a lot of cells which need to be filled. By working with your audience and asking the right questions, you can force them to become clearer about how each of those cells inter-relate. They may not have thought it all through. By asking the right questions, you can make them think about things that need to be thought through in order to fill in all the cells.
The value of getting them to think these things through may be a lot more valuable than the actual number which comes out of the model. For example, they may be looking at changing the price of a particular product/service. To make sure they fully understand the ramifications of the price change, you can ask questions like:
- How would that price change impact the price perception (and cannibalization) on the rest of the product portfolio?
- How will it impact your quality image?
- What happens if competition matches your price?
- How much price elasticity is there in the marketplace?
- If a lower price raises demand, what items are fixed and what items are variable in meeting that demand?
Just by asking those penetrating questions, you can create better decision-making, regardless of the model. I remember someone from McDonalds telling me about their test of a new product called McShrimp Cocktail. The original model looked pretty good until someone asked the question, “How much shrimp would it take to roll this thing out chain-wide?” Once it was determined that 100% of the known shrimp in the world would not be enough to cover annual sales projections, the project was scrapped. Simple questions can be very powerful. Use the tool of the model as an excuse to get in front of people to ask these questions.
4) Once You Have Enough Information to Make the Right Decision, Stop
The goal here is not the perfect model, but the right decision. Sometimes the choice is so obvious that it doesn’t take much of a model to show it. The gap between option A and option B at times can be so large that you don’t need to waste a lot of time fine-tuning the model. If no amount of fine-tuning could ever make option B better than A, then stop the fine tuning.
As you build and refine the model and the assumptions, continually ask yourself this question, “What is the likelihood that further refinement would lead me to a different conclusion?” At the point where you see little to no chance that further refinement would change your decision, then make the decision now and stop wasting time refining the model.
Sometimes the difference between option A and option B can be very slight. In those cases, it can be well worth the time to further refine your modeling in order to better understand which is the right decision. Focus on the key areas which are the least certain and the most influential.
SUMMARY
Financial modeling is just a tool. Its value comes from its ability to help you make better decisions. Depending on the decision, you made need a different model. So start by understanding exactly what that decision needs to be. Then bring clarity around that decision by asking the right questions. Finally, once you have enough clarity, stop fine-tuning the model.
FINAL THOUGHTS
Some strategic implications of a decision are hard to quantify, like the impact on corporate culture or the value of strategic flexibility. Just because they are hard to quantify does not mean they should be ignored. Often, the soft issues make or break a decision. A financial model is just one tool in the toolkit. Combine it with softer tools which take these other aspects into account.
No comments:
Post a Comment