Building a business is all about making a series of good decisions. Here’s a strategy to help you think through the tough ones

By: David Rudofsky

One of the most difficult aspects of being at the top of any business organization is making the hard decisions. How can you choose between two business opportunities that both sound attractive? Should you spend scarce resources to develop a potential product, or are you unlikely to earn back your investment? These are the kinds of decisions that don’t lend themselves to, say, simply drawing a line down the middle of a page and listing the pros and cons of a pending decision on either side of the line.

The good news is that there are more sophisticated techniques that will help you put some science into the art of decision making. You don’t always need to rely on “entrepreneurial instinct” or your “gut” when trying to decide what to do.

Academic disciplines such as operations research and management science have examined and developed different decision-making methods. We’re going to take a look at one of the better-known methods, the decision tree method.

Basically a decision tree uses a graph or a model of decisions, their consequences (cost, time, etc.) and their possible outcomes. It’s a useful method because it’s fairly straightforward and easy to do, and gives the user a graphic representation of the risks and rewards of following various options. However, to use the decision tree method requires business owners to 1) be clear about the various alternative paths that are open to them, 2) estimate what are the possible outcomes that can result from each of the alternative paths they choose, including a probability for each, and 3) estimate a dollar value for each possible outcome.

The real payoff of a decision tree comes from calculating the Expected Value for each of the possible decisions, and then comparing the possibilities to see which is the most lucrative. As a simple example, if a friend tells you he will give you one dollar if a coin flip results in “heads,” and five dollars if it results in “tails,” your expected value is three dollars, i.e.,

$3 = ($1 x 50%) + ($5 x 50%)

It’s important to remember that decisionmaking techniques, including this one, are simply exercises to help you analyze a situation and make a decision. They shouldn’t make the decision for you. Probably the best way to show how this would be used by business owners is to look at a few situations where the decision tree method could be employed.

Software Upgrade Decision

Imagine a small business owner who has to upgrade the enterprise software used to run his manufacturing business. He does not have a full-time in-house information technology department to use as a resource, and he is aware from speaking to other users who have attempted the upgrade that it is not an easy task.

The new version of the enterprise software has added functionality that will allow the business owner to more efficiently order raw material and schedule production jobs, for an annual savings of $160,000. Hiring an outside IT consultant to implement the software upgrade would cost the business owner $50,000, and the consultant guarantees that he will successfully complete the upgrade, given that he has had a 100% track record with other clients on this particular upgrade. But the business owner finds this galling, as he feels it would let the software manufacturer off the hook for the after-sales support it should be providing. However, the IT consultant could complete this project in three months, which would result in nine months’ worth of year one savings, or $120,000.

If he does not hire the outside IT consultant, the business owner would rely on his engineering manager, who has some systems experience, to implement the upgrade. The engineering manager can clear time on his schedule by delegating other projects to his assistant and an intern, but he is not at all sure of the ultimate likelihood of success if he were to manage the upgrade. If he is successful in implementing the upgrade, it would take him six months — twice as long as the outside IT consultant is promising — given his need to first attend some training sessions on the software provided by the manufacturer. As a result, there would be only six months of savings from the project in year one, or $80,000.

If the engineering manager is asked to try the upgrade and is unsuccessful, that outcome would not be known until the end of the six months, at which time the business owner would turn to the outside IT consultant to do the upgrade, which would require an additional three months. In this scenario, there would be only three months’ savings from the software upgrade in year one, or $40,000.

In constructing the decision tree, the owner knows the cost of the consultant and the savings that will be realized by implementing the new upgrade. But he has to estimate the likelihood that his engineering manager will be successful in doing the upgrade. By estimating various probabilities of success, the right course will become clearer. For example the decision tree shown above, (Diagram #1) assumes that the engineering manager has only a 50% chance of success. In this instance, the expected value (EV) of going with the outside IT consultant is $70,000, while the expected value of having the engineering manager attempt the upgrade is only $35,000 (the $40,000 that would come in a successful upgrade minus the $5,000 loss involved in an unsuccessful implementation). Further, even if the engineering manager is successful, the upside is only an additional $10,000, as the lost three months of savings offsets most of the feespaid to the outside IT consultant.

Since the business owner doesn’t know for sure the engineering manager’s probability of success, it’s useful to tinker with that probability figure. Instead of using a 50% chance of success, as with the example shown in diagram #1, the decision tree in diagram #2 (below) shows that unless the owner is more than 90% confident that the engineering manager would be able to successfully implement the enterprise software upgrade, it makes better economic sense to hire the IT consultant straight away. (As a side benefit to the business, which has not been valued, things will probably run more smoothly in the engineering department when it is actually being managed by the engineering manager, and not by his assistant and the intern.)

New Product Introduction

Here’s another example of how a small business might use a decision tree. Small Businesses don’t typically have market research departments, but they might benefit from the use of some market research prior to introducing new products or services, just as larger, more sophisticated marketing-driven companies do.

Pre-introductory market research might reveal the need to make changes in the new product that will lead to an improved financial outcome, such as more accurate pricing, changes in the product or service itself, or an improved communications strategy. Alternately, advance market research might reveal that the new product or service is too weak, leading to a decision that the launch be canceled, if the product or service offering cannot be sufficiently improved to be commercially successful. In either instance, the value of such market research could well exceed its cost.

But, as always, there is a cost to doing that research. Should the business owner lay out the money to do the market testing?

In the hypothetical example shown below, the $25,000 cost of conducting pre-launch market research is paid back by a factor of four to one by increasing the probability of a successful launch from 70% to 80%, decreasing the likelihood of a failure from 30% to 10%, and by increasing the economic value of a successful product launch from $500,000 to $525,000, increasing the overall expected value of the project by $100,000, from $305,000 to $405,000.

Building a decision tree such as this one to justify some incremental spending on market research is a conceptual exercise, not an exact science, so you should not get overly focused on lack of assumptions for the probabilities of success and failure. If you can 1) scope out some basic market research and get a cost estimate for it from a vendor and 2) estimate the net present value of both a successful and failed product launch, then you can use the decision tree template to see how much an improvement in the probability of success the market research has to create to be worth pursuing.

Human Biases

In making your decisions, these methods are meant to help you decide, not dictate, the outcome. But one of the things they can help you do is overcome some decision-making biases that most people seem to share.

Researchers have shown that while humans excel at decision making based on knowledge of past events, they are poor at making decisions where the consequences have to be calculated based on a forward-looking evaluation of consequences — exactly the type of situation a decision tree addresses.

Humans have also been proven to have a “concreteness bias” — a tendency to overvalue and gravitate to what is proven and known, and to avoid what is unproven and new. By our very nature, we want to use our existing employees to solve problems and avoid the risk, trouble and expense of hiring outside consultants. That’s why it’s valuable to see how an objective analysis would show that there is a strong economic rationale for doing so.

Finally, many of us want to either delay or totally avoid making decisions that have unpleasant emotional consequences, such as firing or reprimanding an employee, closing a failing business or taking decisive actions to shore up a business with a weak balance sheet and an impending cash squeeze. Researchers have also determined that when the decision making finally occurs for emotionally laden matters, decision makers work “harder, yet not smarter” in making these decisions, as they try to avoid explicit trade-offs between emotionally painful alternatives.

Decision trees are a wonderful decision- making tool but are not used often enough by small business owners. Business owners who can adopt them in their decision making have an opportunity to make value-added decisions that are analytically sound and avoid a series of proven biases that may be holding them back from making the best decisions for their businesses.

This article was published originally in The New York Enterprise Report.