When I mention the word research, what comes to mind? For most people, research has a lot to do with a team of strict-looking intelligent individuals going over piles and piles of documentation, going to various places to obtain data, and undergoing numerous processes and activities to analyze the data they were able to gather.

What Is Operations Research And Its Best Practices

That just goes to show how little we know about research. Chances are high that, even if you are knee-deep in the operations side of things in your job, day in and day out, you still don’t fully know or understand the concept of operations research, or what others also refer to as operational research. This is our opportunity to correct that.


Decision-making is one of the most vital processes in management since decisions are made in order to achieve something. In the context of business management, managerial and organizational goals are what most managers seek to achieve. And that’s what they are driving towards with every decision they make.

But decision-making is not something to be taken lightly. Considering what’s being aimed for, and what’s at stake in the grand scheme of things, you can’t just randomly pick one among a few options. (Well, technically, you can do that, but considering the risks involved, especially if you end up making a wrong choice, you’d want to spend more time thinking about it and weighing your options before deciding.)

With that being said, describing decision-making as a process is accurate. You have to undergo several steps or phases before you can confidently arrive at a decision. The same thing applies with problem-solving. There are also several steps to pass through before you can get to a viable solution to a problem. There are certainly more than two ways to go about the decision-making and problem-solving process, and operations research is one of them.

Have you ever heard of the phrases “management science” and “industrial engineering”? They are terms that also apply to operations research. If you hear about management using business analytics, marketing analysis, logistics planning and even the more broad-sounding decision support, it is basically the application of operations research.

“Operations research”, or simply OR, is described as an analytical method of problem-solving and decision-making used in managing businesses or organizations. It involves the application of advanced quantitative techniques in order to arrive at a decision or solution to a problem, so we’re talking about using mathematical and numerical techniques here.

What sets OR apart from other types of decision-making processes is how mathematical analysis plays a central role. The identified problem is broken down into its most basic components, and mathematical methods and techniques are employed to solve them.

The application of OR is widespread. In fact, all businesses can never be completely free from having to apply OR in their own business environments, regardless of the nature of their business operations or the size and scale. Retail businesses and service providers, which have pretty much straightforward business processes, will still need OR in their decision making processes.

With that said, the application of OR is more necessary on the larger and more complex operations, such as companies involved in highly technical industries such as information technology, biotechnology, engineering, military operations, and telecommunications.

The fact that all businesses have to perform functions and processes such as financial planning, manpower and resource allocation, and risk management means that they all require the practice of OR.

Here are some examples of OR applied or used in a real-world business settings:

  • Forecasting and planning, such as in the determination of production capacity, manpower and resources allocation, and establishing the economic order (and reorder) quantity.
  • Scheduling, such as sequencing in a supply and procurement chain, or processing orders in manufacturing assembly line.
  • Marketing, such as in customer profiling and implementation of sale promotions and other campaigns.
  • Facility planning or layouting, such as when designing an online processing system or the floor plan of a manufacturing building.


If you’re wondering why there is a need for OR, the most obvious answer is in order to facilitate business decision-making. After all, OR is a huge part of planning.

The decision-making we are referring to in this specific context has to do with optimization. Therefore, we can say that OR is very important because it enables businesses to “do things best under the given circumstances”.

But that’s too broad of an answer, and does not really explain in detail why you should use OR in decision-making and problem-solving and in managing your business organization in general. Let’s drill in.

  • OR simplifies the business environment. Now you might be wondering, how is that possible? Wouldn’t it be even more complicated if we threw in some mathematical elements in the mix? Well, yes and no, but that mainly depends on how you go about the OR process. In a business environment, numbers and figures often provide the most reliable information. Quantification gives more room for objectivity, so business decisions can be made objectively, since there numbers say so.
  • OR maximizes the usefulness of data. Depending on the size of the business operations, there are a lot of data that have to be dealt with on a daily basis. Larger operations are faced with millions of bits of information, and going through each and every piece of data can be tedious, time-consuming and, therefore, counter-productive. Through the use of OR techniques and analytical methods, there is a way to handle all those volumes and volumes of data in significantly less time. Obviously, this will lead to being able to make better decisions, faster.
  • OR aids in the optimization of resources. Resources are scarce, so businesses have to find ways to make the best use of the resources that are currently available to them, while ensuring that they are of high quality or, at least, with quality that meets the expectations of the end users.
  • OR ensures effective and efficient delivery of products or services to the end users. By applying OR in decision-making, the process becomes more systematic, so that you are able to provide the high quality products or services to the customers or end users when and where they are needed. Having high-quality products will be of no use if you are unable to deliver them to the end users when you’re supposed to.


Follow a systematic and logical approach. For that, we suggest the seven steps of OR, which we will get into more detail below.

Other literature broadly described OR to involve only three steps. First, you identify the potential answers or solutions to the problem, and they will then be analyzed and narrowed down into the most feasible or viable options. The third step involves further analysis, this time using more specific analytical tools.

When we talk of OR, however, there is the “Operations Research Approach”, which is composed of seven sequential steps. Let us walk through it together.

The Seven-Step Operations Research Approach

This approach is represented in this diagram. We’ll be taking a look at the activities involved in each step.

Step 1. Orientation

OR is not a one-man activity. It takes a team, with members equipped with various skills and specializations assigned with various tasks and functions, depending on their strengths or the areas they excel at. Therefore, there are two things that must be done in this step.

  • Form the team that will conduct the OR study. Take into account the multifunctional nature of OR when choosing the members of the team. You want them to be qualified to conduct OR, so you don’t have to start pulling in just any random person from the other departments midway through the study for the simple reason that the existing members turn out to be unable to do the job. It is advised that the areas or divisions that are directly or even indirectly affected by, or related to, the OR be represented in the team. If the OR is on product design, you’d want to include the engineering, assembly and quality control divisions to be represented, along with someone from finance and marketing, specifically those that are involved in customer and market research. Install a team leader who will be able to steer the team in the right direction, and one with the ability to manage both the work and the members of the OR team.
  • Ensure that all members of the team fully understand the issues at hand, specifically on the matter regarding the OR. What are they supposed to study, and what should they pay attention to? For what reason are they conducting this specific activity, and how will it benefit the organization? These are only a few of the primary questions that you must address before the members of the team so that they won’t be “flying blind”.

Bringing them into the loop will also motivate them to do the best they could in the OR study. It is also important that you are able to instill an appreciation within the team for the objectives of the activity and for what have been done so far (if there are).

Step 2. Problem Definition

Most processes – even the scientific method – puts this on the first step, and it is considered by most to be the most difficult part of the entire process, since it will set the tone for the rest of the activities or tasks that will follow. If you don’t know what the problem is, then you will simply be spinning your wheels and going nowhere.

If, on the other hand, you were able to identify a problem, but it’s not the actual or real problem, then you will also end up wasting a lot of time and resources, and you might even end up making the wrong decisions.

In defining the problem, you have to clearly identify its scope and the results that you desire or expect to have at the end. This time, you will be more specific. Instead of saying that you want to improve the company’s product design system, you will have a more targeted objective, such as “to lower the unit production cost of the product”.

Once you’ve identified the specifics, delve deeper into it.

  • Identify the specific factors that will affect your objective, clearly distinguishing those that are within your control from those that are not, and determine all possible alternative courses of action that may be taken. Say that you want to lower the unit production cost of the product, so the factors may include the flexibility of product design, factors of production used (e.g. direct materials, direct labor, overhead).
  • Identify the constraints on the courses of action. There are bound to be limits that all decision-makers in business have to operate within. It is possible that the nature of the product and even government regulations and legislation do not provide enough room for flexibility in product design. Availability of resources – especially the alternative resources should you decide to change some of the inputs into the product – is also another constraint.
Step 3. Data Collection

In this step, there are two things you should take note of before you can go about successful data collection. Of course, this is under the assumption that you already know what type of data you should collect.

  • Sources of data. There are many identifiable sources of data, depending on the data type you need. Generally, we look to existing standards, such as current and historical trends and set values. Another source is the system or process that is being studied, particularly on how it works in actuality.
  • Methods and tools for data collection. Observation remains to be one of the most commonly used methods of data collection and, thanks to automation and computerization, combined with the flexibilities brought on by the internet, data collection is greatly facilitated. What used to take businesses years to collect data and process it into valuable information is now doable in just a matter of hours, days even.
Step 4. Model Formulation

Modeling is what sets OR apart from other decision-making processes. Where other approaches would directly look into the system and analyze it, OR goes about it by formulating a model, or a representation of the system, and using that model for its analysis.

Modeling allows the researchers to simplify the system while maintaining its accuracy and faithfulness to the original. Besides, it is much easier – and less costly – to analyze the model instead of the actual system.

The team conducting OR may create different types of models, and there are four general types of models that are often formulated and employed.

  • Analog models. These are models with physical properties that are significantly smaller than the actual system being studied, and having similar characteristics with the latter. These similarities make the model and the original analogous, even if they are not identical.
  • Simulation models. This involves the approach where the behavior of individual elements within the system is mimicked or mirrored. In other words, a model of a real-life situation is created, and that’s where techniques such as sampling and experimentation, if need be, are conducted. This method is usually favored as it allows testing for future improvement. Through simulation, you can analyze even complex systems by coming up with estimates of statistical measures. Values are inputted and, with every replication, you can observe the response of the system. In this day and age, when technology plays a very important role in almost all businesses, computer simulation is often applied. This allows you to look for areas of improvement, specifically in an automated business environment.
  • Mathematical models. OR is considered one of the many branches of mathematics, so do not be surprised when you find yourself having to apply many mathematical methods in your OR. Without going into the most intricate details, let us list down the various logical methods employed in OR, which were also cited by Springer. The preference for usage of mathematical models is how they effectively map out all the variables and describe their relationships with each other.
  • Physical models. As the name implies, this is a tangible model, which is basically a copy of the original system, but scaled down appropriately. Unlike the analogic models, which are simply made to be analogous to the original system, these scaled down versions are smaller replicas of the original. Among the four model categories, this is the hardest to pull off, especially in the case of complex systems.
Step 5. Model Solution

This is where you will attempt to solve the problem; in other words, it’s the analysis stage. Needless to say, this is the part where the OR team will spend the most amount of time and resources, employing a variety of analytical methods and techniques on the models formulated in the previous step.

Briefly, the most commonly used techniques are:

  • Simulation techniques, for the analysis of simulation models. These techniques often come part and parcel with several statistical techniques. That’s right. If you though that resorting to simulation will save you from dealing with numbers, you can’t fully get away from it, since statistical computations will still hound you.
  • Mathematical analysis techniques, which dominantly utilizes statistical methods, such as regression analysis, variance analysis, queuing, and statistical inference.
  • Optimization techniques, where you will try to determine the best values or the optimum levels that will affect decision-making. That involves the application of various mathematical programs and statistical methods. Mathematical programming techniques often used include linear and non-linear programming, integer programming and network theory.

At the end of this step, you should have obtained a solution, after considering the results of the analytical tasks you used.

Step 6. Validation and Output Analysis

Does the process end once you’ve identified the solution? No, it doesn’t. You still have to make sure that the model you used in your analysis is, indeed, an accurate representation of the system. This is the validation part.

And that’s not all. If you thought you’re done with the analysis bit, there’s still more analysis to be done. In this case, you’ll be going over various “what if” scenarios, where you will consider the possible outcomes if the solutions obtained are implemented.

Step 7. Implementation and Monitoring

Finally you settled on the best solution or recommendation and made a decision. It is time to implement that decision.

Of course, you need to still have control over the implementation, which is why there should be a team in place to be in charge of the implementation. It is highly recommended that you place some members of the OR team in the implementing team.

Monitoring is a must, since you want to ensure that the solution decided upon is the one actually being implemented. This is also a way to remain on your toes, since unforeseen circumstances might lead to some aspects of the solution needing some tweaking along the way.

Use only the relevant data.

Out of a million pieces of data, you’re probably going to need only a fraction of it. Wading through all that data may be all right with you, but SHOULD YOU? Think of the resources you will be wasting if you do that. It will also take a lot of time, which you can devote to other core functions, instead of poring over data that won’t have an impact – even if indirect – to the matter at hand.

One of the reasons that you use advanced analytical methods is so that you can maximize data and handle as much of it as you can at one time. But that does not mean that you should analyze 100% of the data, even if 50% of it are not relevant to the problem you are solving or the decision you are trying to arrive at.

But do not focus solely on the quantity of data; you also have to count quality. Having too much data is not the only problem; having poor quality data is also just as problematic. In fact, researchers prefer having a small amount of high quality data, instead of having too much data of poor quality or no relevance at all.

Maintain close collaboration between managers and the researchers.

A certain degree of independence is encouraged when it comes to the people directly conducting OR, or the researchers. This is so that they can maintain a level of objectivity in their analysis.

But that does not mean that they should be completely removed from the management. Management support is vital if you want your OR to be successful. After all, at the end of the day, it is the management that will make the decision, and will see to its implementation. By striking a partnership with the researchers, the process will be smoother.

In fact, it is recommended that researchers work alongside the managers, or those who are directly involved in the process being analyzed.

Establish policies or a framework for the conduct of OR.

One way to give OR a strong presence in the organization is to institutionalize it. How can you go about that?

  • Create a policy framework, providing details about OR – even if they are couched in general terms – which will then serve as a guide for staff who will later on conduct research. This is also one way to impress on the members of the organization the important role of OR.
  • Create a reference document containing the policies or the framework, and disseminate it to the members of the organization. It won’t make any sense if you have a framework, and it is well-documented, but it remains inside the office of select few members of top management.

Make Operation Research an integral part of your business processes.

In other words, do not treat it as just a minor function that you can just assign to whoever has free time. Research activities, in general, take time and certain level of commitment on the part of the researchers, so treating it as a throwaway task is not a good idea.

  • Assign staff members to focus on OR. Some businesses, refusing to spend on OR, make research as an additional task for its managers. This may be workable, but if the managers are already overworked, chances are high that they won’t devote as much time and attention to the research side of things. Oh, and one other thing: do not forget to assign someone to manage the research activities carried out. Pick someone to lead the team, so as to maintain some supervision and cohesiveness in the unit.
  • Specify a dedicated time to conduct research activities. Again, it’s not a good idea to request your staff to “do their research whenever they have free time or during breaks”, or even demand that they render overtime specifically for research activities. Doing that will only make OR seem like an afterthought, instead of the important business process that it actually is. Maybe you can schedule at least one day per week for staff to do their research. This way, they will also be able to maintain focus when they’re on the job.
  • Make room in your budget for OR. Yes, you need funding to conduct research successfully. Conducting OR means you will have to spend on salaries and compensation of the researchers, and other expenses incidental to the conduct of the research activities.
  • Pick the right people. You have to make sure that the researchers understand what they are supposed to do, and they have the skills required to carry out the research successfully. You may have to conduct some on-the-job trainings, if necessary.
  • Equip your people with the right tools. Arm them with the things they need in conducting a successful OR. If you think they will benefit more by undergoing training and workshops on OR, then send them to those activities. And we’re also talking about providing the staff with the hardware they’d need to carry out the many OR techniques that they have to use.

At the end of the day, the best practice that a company can apply in operations research is to fully commit itself to actually doing it. OR is an indispensable process in managing a business, and you’d do well to keep that in mind, if you plan on taking your business to greater heights.

Comments are closed.