Decision trees are a valuable tool for business. They set out complex decision processes in a way which is visually clear while being suitable for workflow management. Read on to learn the benefits of using the decision tree framework to map out and manage your processes.
Decision management is a key component of process management. Here at ProcessPlan, we've discussed a number of ways to improve your processes, from how business process management can streamline your company to how standardized processes can boost customer retention. These concepts are integral to automating your workflow and eliminating mistakes. However, there's another important concept we have not talked about as regularly — using the decision tree framework to map out and manage your processes.
Decision trees are a vital component of decision management. Projects both large and small have multiple objectives that are sometimes in competition with one another.
The most obvious way to get through a complex set of choices is to handle each one as it comes. Sayings like "We'll cross that bridge when we come to it" reflect this approach. However, that can lead to a burned-out bridge with no way back. Decision trees enumerate the possibilities in advance and help to avoid costly dead ends. People like to think intuitively rather than systematically. They have theories about what works or doesn't. This isn't necessarily a bad thing. Someone with a lot of experience can be a very good judge. However, intuitive thinking relies on hidden assumptions which may change. A decision tree brings all the assumptions into the open.
Decision trees are a valuable tool for business. They set out complex decision processes in a way which is visually clear while being suitable for workflow management. They let an organization weigh the expected returns from different choices.
An overview of decision trees
A decision tree is a tool for representing a process that involves choices. It’s like a flowchart and graphically maps a sequence of decisions and possible outcomes. It has one starting point and any number of levels of branching. The branches represent choices or the possible results of a choice.
At each point in the process where there is a choice, there are two or more branches going out from a node. They represent alternatives such as "Accept offer" and "Make a counter-offer." "Do nothing" is very often one of the choices. The tree's creator can designate a probability for each result where there are multiple possibilities. Each result can also carry a cost or benefit value in time or money.
A decision tree lets planners see what set of choices is most likely to produce a given outcome and what the expected costs or rewards of a decision are. Software can run through all the possibilities and show which ones offer the best outcomes.
Users can revise their decision trees as necessary, adding or removing branches and changing the numbers when they get new information.
While creating a simple decision tree on paper or a whiteboard isn't difficult, computer software makes the job more manageable for large ones. It's easier to update one on a screen, and there's no problem with running out of work area. Calculating probabilities and outcomes manually is tedious work, but software can do those calculations and find optimal decision paths instantly. Software can do sanity checks, such as making sure that the probabilities at any point add up to 100%.
The value of decision trees
Employees at lower levels or with less experience benefit from the standardization that comes with using decision trees. They won't find themselves out of their depth, having to make choices on the fly. They'll have the choices in advance, or they'll have received instructions on how to make them. At the time they have to make a decision, the people with greater experience may not be available to help them. With a process that's already planned, these employees now know what to do.
Additionally, decision trees give managers an overview of the whole process. They break it down into a series of concrete actions and definite criteria. Hidden assumptions come out into the open, where managers can decide whether they make sense or not. They can accept or reject each decision step on its merits, rather than using gut feelings to make decisions on the fly.
When managers start on a project, necessary numbers may not be immediately available. Laying out an explicit process makes it clear where more information is necessary. When every cost, risk, and benefit is quantified, everyone is clear on what choices to make and what their basis is.
Mapping out a process as a decision tree offers several benefits:
- It presents all the steps in one place, helping to make sure all important considerations are included.
- It turns a decision process into an algorithm, so everything is clear and consistent in the creator's mind and visible to others.
- It presents the same information about the process to all the people on the team, avoiding misunderstandings.
- Constructing the tree forces its author to be objective about the choices to make. There will be fewer spur-of-the-moment decisions.
- The tree represents advance planning. The project head needs to identify all the possibilities before taking any actions.
Setting up a decision tree doesn't mean putting the decision into the hands of an algorithm. Most often, each option will have advantages and disadvantages. Some choices offer a high return but also high risk. The choices are up to the human decision makers.
Process management and decision management
The concepts of decision management and process management are related but distinct. Process management deals with the bigger picture. It aligns business processes with strategic goals. Business Process Management (BPM) software automates aspects of process management. Decision management deals with the choices which will achieve those goals. Process management is the strategy of business, decision management the tactics.
The two need to work together, with a flow from the general to the specific. Policies fall under process management and apply to all processes. They form the basis of business rules, which apply to specific kinds of actions. Business rules can be automated, and they delimit the choices which are available in any given decision tree. If, for example, an expenditure exceeds a given threshold, one rule will apply; below the threshold, a different rule applies. Each rule may allow multiple choices, but they aren't the same set of choices.
Process management and decision management too often are in separate silos. This is unfortunate; the strategic and tactical aspects of business need to be in close communication. Software that integrates business policies, business rules, and decision-making lets people at all levels work from the same information and approach goals in a consistent way.
Examples of usePurchasing decisions
A very simple example is deciding whether or not to buy printer paper. It might go something like this: Is there enough paper to last a week? If not, buy 20 reams. If yes, buy 10 reams. Otherwise don't take any action. The tree could expand to take any number of special cases into account. Some purchasing decisions might have complicated decision trees, but the basic idea doesn't change.
Employees vs. contractors
A company needs to decide how to staff a project. A decision tree breaks the process down into discrete choices and measures the costs.
The first question is whether to hire employees, hire contractors, or let the existing staff do the job. Each one has its own set of options. If the choice is to hire employees, the next step is to decide between permanent and temporary ones. For contractors, the questions include choosing between an established firm and independent freelancers, and between onsite and offsite work. Hiring no one means more overtime or reducing the work on other tasks.
Each of these choices carries its own costs and benefits. Full-time employees carry greater long-term costs, but they provide an experienced team for later projects. Costs are uncertain until the company starts exploring the market; the estimate which a contractor offers will be a major factor in deciding whether to hire it. Using the existing staff is the least expensive but could increase employee dissatisfaction and impose costs on other tasks. A decision tree will break down the possible choices and the factors influencing the choice at each point.
There isn't necessarily a single "best" outcome. Every choice involves tradeoffs, such as long-term and short-term costs. The decision tree lets managers compare these tradeoffs in a quantitative way.
An assisted living facility is considering expanding its range of services. It could add one or several offerings, or it could stick to the status quo. Services might include grocery shopping, transportation, and classes.
Each of the services may be charged when used or part of the basic package. Charging per use lets the facility keep the base price down and getting income in proportion to their use. At the same time, it means usage levels will be lower and less predictable. Combining several services in one package may be advantageous, especially if they share cost factors.
Factors to consider include not only monetary costs, but uses of available space, regulatory and contractual compliance, and public relations. A decision tree can assign costs and returns to all of those values.
Limitations and cautions
A decision tree is only as good as the information given to it. Its probabilities, costs, or benefits are projections based on the project leader's judgment and the available information. If those numbers are seriously wrong, using the tree will produce misleading results. Managers should not hesitate to revise decision tree parameters when they get new information.
By using a decision tree, managers make the process more objective. Spelling all decisions out eliminates hidden biases and assumptions. Explicit costs and options become the basis for each decision. Whatever choice managers make, they can say, "These were the alternatives, and these are the costs and benefits of each one." Without an explicit process, the best they can say is "This choice seems better than the others."
Some people start with a conclusion and then build a decision tree to match. This misses the point. It may allow a presentation that will push people into agreement, but if the decision is a bad one, its consequences will come around before long.
Having access to relevant data allows better numbers. When putting probabilities on the response that a mailing will get, knowing how previous mailings have done will give better numbers than pulling one out of a hat. The cost assigned to a trip should come from people who do the bookings, not from glancing at a travel website.
Some factors are hard to quantify in a decision tree. A decision might branch on "The interview went well" vs. "The interview went badly," but what do those actually mean? The criteria should be put in the most objective terms possible. The interviewer's recommendation to hire, reject, or hold after an interview is a clearer criterion than a vague evaluation.
Bringing everything together
Decision trees are one aspect of decision management, which needs to work in coordination with process management. The people who work on the development of policies and processes need to understand how the decision processes work on projects. Then they can understand when decision management is out of step with policies, or when the policies don't take important decision-making factors into account. ProcessPlan brings process and decision management together.
Process management uses workflows, the set of steps taken to achieve a goal. By itself, a workflow shows the "what" but not the "why." The decision tree which shaped the workflow should always be available, so that managers understand the reasoning behind the actions taken. If unexpected events interrupt a workflow, the decision tree shows what the alternatives are.
The integration of process and decision management has important ramifications for an organization. It allows consistency of policies and actions at all levels. Decisions made without organization-wide direction are likely to be subjective. They're certain to vary with each project head's style and temperament. Business-wide policies and rules provide consistent guidance to each project, so that they all operate on the same basis.
The decision process needs to be in line with the organization's way of doing things, down to small details. Everything is laid out, so that there's as little need as possible to improvise. People will know what to do, and they won't break rules through lack of information.
Good mechanisms for information flow are among the most important assets a business can have. Getting all the relevant information into a decision process and making its steps explicit will result in a much higher success rate for projects. This is where ProcessPlan comes in. Learn more and sign up for a free trial or request a demo today.