By Avraham Mordoch and Eli Schragenheim
How can the new fast development of technology effectively help organizations to achieve more of their goal? The vast majority of the new technologies have a considerable impact on IT (Information Technology departments), which causes huge pressure on the workload of the IT people in most organizations that need to keep themselves frequently updated, causing headache to top management and lack of focus instead of helping them to move the organization forward.
But, the potential of the new power of computerization, including methodologies like Big Data, Artificial Intelligence (AI) and the Internet of Things (IoT), could also be used cleverly to improve the effectiveness of management by leading the organization to a secure and successful growth of its activity.
On one hand the new technologies allow getting more data, which is also more accurate than ever before. That data, when properly analyzed with focus on what is truly meaningful, could serve managers in analyzing the current state of the business, its weaknesses and could lead to new ideas of how to improve the bottom line.
This article offers new ways to use the recent technology to allow management to get beneficial support in evaluating new ideas or prepare for expected changes coming from elsewhere. We have chosen to focus on manufacturing organizations, which face the threat and the potential benefits, of digitization of the manufacturing shop-floor, considered to be the fourth industrial revolution, and thus gained the title of Industry 4.0. The threat is being pushed to enormous expenses without gaining any business benefits. The capabilities of the new technology could assist a dramatic improvement in the way tactical and strategic moves are evaluated. The point, though, is that in order to materialize the benefits some management paradigms have to be challenged and replaced with common-sense paradigms that utilize the new capabilities to support decisions.
Viewing the current types of software systems supporting manufacturing organizations, these systems can be classified into four types:
- MES (Manufacturing Execution Systems). This type of systems is focused on the very short term and aims at providing operators and production management with the most updated state of the flow of raw materials all the way to the finished goods inventory. It allows handling priorities, fast fixation of problems and achieving efficient utilization of the equipment. MES collects data and organize it in a way that can be easily viewed by middle operational managers. Scada systems, for example, are a subset of MES systems
- ERP (Enterprise Resource Planning). We include in this class also the CRM (Customer Relationship Management) systems. This class consists of a suite of integrated applications that the manufacturing organization can use to plan operations and collect, store, manage, and interpret data from different business activities. What integrates all the various part of the ERP class of systems is one database of all the key transactions, of the financial and the material, that have been recorded or are planned to be done in the short to medium-term. The main function of this type of systems is planning the basic operations required to deliver the firm orders, while also record the transactions and institute order and the systemization of all the data related to the main processes in the organization. ERP and CRM systems are mainly data systems with some crude planning functionality. When information is defined as the answer to the question asked (Goldratt, The Haystack Syndrome) ERP and CRM supply answers to the most frequent and simple questions, like what needs to be done in order to deliver a customer order. SAP, Oracle or Dynamics 365 are just a few examples of ERP systems
- BI (Business Intelligent). The objective of the BI programs is to display high level information for top management, providing a picture of the current situation, and possibly pointing to certain observed trends. The power of the BI technologies is to be able to collect relevant data elements from various databases. Internal data is mixed with data that is collected from the Internet and used to create graphs and charts for management to be aware of what’s going on within their organization and how it compares to what’s going on in the market and with their competitors. The Key Performance Indicators (KPIs) are supported by BI making them clear to top management. This gives the background to management to evaluate ‘what to change?’ But, it does not provide the tools to ‘what to change to?’, definitely not to ‘how to cause the change?’. In other words, BI supports decisions by pointing to required areas, but it does not support specific decisions.
- Decision Support Systems (DSS). While the title of DSS was raised already in the 80s of the previous century, the true capability of actually supporting decisions has been achieved only recently. Every management decision is considering a change to the current state. Every significant decision is also exposed to considerable uncertainty. So, the key capability of a DSS is to be able to direct the managers to various alternatives to the considered decisions and present the possible ramifications of these changes. We can divide this level of DSS into two parts:
- Supporting routine decisions done by experts, so less experienced people can take them, or even let the computer make this decision. These types of computerized programs are based on Artificial Intelligence new technologies and create a variety of expert systems that support such decisions.
- Supporting more significant tactical and strategic decisions by providing the decision makers with the holistic analysis of the potential financial and other ramifications of the decisions. These are systems that support business organizational decision-making including decisions that consider unstructured or semi-structured potential opportunities that are exposed to significant uncertain situations. The assessment should consider the short term as well as the long term. A DSS must “understand” the cause and effect relationships between the different functions of the organization. These systems should allow a direct interaction between the human decision-maker(s) and the computerized algorithm. The objective, given the amount of uncertainty and lack of full precise information, is to present the decision-makers with a full picture of what MIGHT happen, for good and for worse, as the result of the decision(s).
The above four classifications are not clear-cut and there are systems that cross the lines between them.
On top of that there is often an interaction, even a loop, between the above types of systems. The ERP consumes data accumulated by the MES and accordingly creates work orders that feed MES system. The ERP database has a major role for the BI system showing the current state and the ERP and BI data are input into the DSS programs.
Interestingly enough the effort a manufacturing organization needs to make to implement these systems is especially significant when implementing an MES system, since there is a need to overcome cultural objections, including the antagonism that one finds in organizations with no established culture to report what has been done. When this initial infrastructure is laid down, it is a bit easier to implement and properly use the ERP system and by far easier to continues to climb the ladder and implement Expert Systems and the higher level of DSSs. So, the effort is reduced going up the ladder through the four types of systems, but the benefits from the implementations is increased and there are very significant benefits when the top management, the C-level managers, are using a DSS for solving the crucial dilemmas they may have.
We have to take into account that manufacturing organizations are both complex and exposed to significant uncertainties. Still the C-level managers have to make tough decisions like:
- Should the company offer packages of its existing products for a reduced price?
- Should the company accept small orders for customized products for a not-too-high markup?
- Should the company expand the product-mix with additional product family (or families)?
- Should the company save considerable cost by shrinking its resources, as well as stopping the production of products with very low demand?
- Should the company go on massive advertizing campaign?
- Should the company participate in a big tender, quoting a moderate price, knowing that winning might affect the good delivery performance of the regular orders?
- Should the company invest in opening a new export market?
- Should the company invest in a new production-line when the market seems to go up, but some people believe this upward trend is going to stop?
These decisions lie outside the comfort zone of the decision makers, because of the obvious risk and having much less past experience with such situations. The decisions are risky not just from the perspective of the organization, but also from the perspective of the personal risk of the decision maker, who ties himself/herself to the success or failure of the initiative.
The above risks force conservative decisions whenever the needed decision is beyond the known comfort zone. Lack of proper support for a holistic analysis blocks many organizations from achieving the true potential of the organization.
There are two big obstacles for any DSS to tackle the above decisions and many others: One such obstacle is expressing the intuition of the people close to the relevant area to play its role in the analysis. Even when the situation is beyond the comfort zone of the decision-maker, it is still valuable as the people involved always know something that is more than nothing. While lack of good precise relevant data is a constant issue, analyzing what MIGHT happen is a valid possibility, which yields a focused picture of the actual risk.
The second obstacle is being able to evaluate the proposed decision when it is added to everything else the organization is doing or committed to do. This requires deep understanding of the rules behind the flow of materials, products, orders and financial transactions, including the various dependencies in Operations and in Sales. This leads to the massive calculations, checking the state of capacity, materials and cash.
The DSS program needs to “simulate” the top-level dilemmas (like the examples above) and come up with the predicted financial results. It has to make it easy to run a variety of ‘what-if’ scenarios and compare the results. In the end, it has to display the predicted results for, at least, two different scenarios: one that is based on reasonable conservative assessments and the other on reasonable optimistic one. The range of the end results means the reasonable result should fall anywhere in between the extreme sides of the range. The decisions should never be done automatically by the system – it needs constant intervention by the decision maker looking for better alternatives and use human judgment to make the final decision.
Generally speaking there could be two main ways to accomplish an effective support for decisions:
- Being able to carry a mass of calculations, based on good cause-and-effect rules, which describe the materials and capacity requirements for every product sold as well as the impact on the revenues and cost. This way is described in detail by the book Throughput Economics, written by Eli Schragenheim, Henry Camp and Rocco Surace.
- Using a powerful computerized simulator that closely follows the flow rules, and records revenues, truly variable expenses and the cost of capacity as an integral part of the simulation. The uncertainty has to be input into the simulator’s critical parameters to provide the possible range of the results.
The mass calculations way is more visible to the decision-makers, as the calculations are all straightforward and the added-value of the computer is the ability to carry such mass calculations. This means the decision makers fully understand the assumptions that are at the core of the calculations.
Using computerized simulation better fits complex situations, either within the production-floor, or with complicated dependencies within the sales. For instance, simulating different flow rules, like batch sizing and different prioritization, are much more effective than mere calculations that have to rely on assumptions regarding the effectiveness of the flow rules. On the other hand, the user has to inquire deeply to validate that the internal parameters of such simulation are in line with reality in order to trust the results.
Both ways have to start with a good representation of the current state as a reference according to which all the changes are compared to. For a simulation it means creating a ‘digital twin’ that seems to come up with the current performance of the organization.
A computerized system that produces reliable reference or a digital twin, and is able to introduce variety of changes and compare the results to the reference, while also depicts the potential impact of uncertainty and lack of accurate data, deserves to be called a decision-support-system (DSS). Such a system will reduce significantly the risk in taking top-level decisions and will also reduce procrastination that is usually found whenever ‘hard decisions’ are evaluated. This would help significantly to put the company ahead of the competition.
The first few true DSSs to appear in the market will enjoy a “Blue Ocean Strategy” compared to the “Red Ocean” which is typically the current situation in systems supporting manufacturing organizations.