Elements of Decision Making
Despite the fast evolution of technology, the elements of business remain the same. The challenge has consistently been to leverage innovative technologies to respond faster & better to changing market conditions, provide competitive insights into how to differentiate your business model and to improve productivity of personnel and assets. The higher levels of the organization are tasked with designing the processes for planning resource levels, scheduling these resources and executing the schedules. Again, this is very basic stuff. Each of these processes have the same fundamental process collect input data, make a decision based on that data and then output the results. This all sounds easy but the truth is with all the different ways that businesses can construct these processes and the various technologies that can be deployed, the task is daunting. It is like someone trying to find the shortest way out of a thick forest of trees.
Software vendors should not be looked on as trailblazers because they have a stake in their own technology. If you were to believe the claims then it is no wonder the marketplace is being oversold on the hype. However, let us be fair to the vendors, it is because the customers have such a poor understanding of new technologies that there is so much hype. Worse, companies need to articulate a conceptual design before they start to select and deploy technologies.
The intent of this article is to provide a perspective on the various emerging concepts and technologies to help you understand which might help your organization at any point in time.
One of the more promising technologies is Enterprise Information Portals (EIP). First, the related technologies will be described to put EIP into perspective.
ERP – Enterprise Resource Planning Systems
Legacy business systems were built around individual business functions. The whole genre of MRP/MRP II and ERP systems incrementally integrated more and more of the information handled by the various business functions. This is valuable because in reality the functions are integrated. The intent of ERP is to automate standard business transactional functions as much as possible within the corporation.
E.g., processing a customer order is a whole sequence of events from credit checks, checking availability, issuing goods releases, arranging transportation, issuing invoices and receiving payments. Integration improves the process by eliminating double entry and automating parts of the process that were manual before.
Data Warehousing / Mining
Data has a shelf life. The older it gets the less valuable it is. However, given no other information, old data is often used to determine trends and help forecast the future. ERP systems generate and capture in excess of 10 times more data than legacy systems. Matter of fact, so much data is generated that the old data needs to be purged from the core system into a database, so that the performance of the transactions is not impaired.
Data warehousing provides a structure to storing information from internal and external sources with a time relevancy. Data mining is the process of querying the stored data - filtering, aggregating and sorting it into concise reports. These tools provide the most basic management reports: a picture of history, the present and the future. The old process was to collect (often manually) data from many different sources and then construct the reports. The concept of a Data warehouse is to automatically collect unstructured data into a structure database and provide the capability to generate standard and customized reports easier and faster.
Decision Support
One of the most popular decision support tools is the spreadsheet. The concepts and the tools just discussed are fairly well defined and understood due to their limited functional uses. When “decision support” is referenced the connotation is much broader. A simple management report, a graph or chart are basic decision support tools. A spreadsheet provides enhanced capability because scenarios can be what-iffed to provide limited insights. Manually changing a few numbers can be analyzed for their impact on the results. E.g. Changing forecasts can be analyzed for the impact on production plans.
The basic characteristics of decision support are: that the data must be structured, the format provides a picture of a situation and the tool can be used to arrive at a possible solution or an acceptable answer.
More sophisticated programs have been developed for decades to analyze data and provide conclusions. Traditionally, the design of these programs was very narrow for specific business problems, situations and market conditions. A fair amount of reprogramming was required as changes occurred.
Supply Chain Systems
A newer class of decision support tools has evolved that address a wide variety of standard supply chain issues. Advanced Planning Systems characterize this class of systems and include - Forecasting, Finite Scheduling, TOC, distribution planning, capital planning, etc. What these systems have in common is that they take vast amounts of structured data to calculate results based on an objective function.
In essence, advanced mathematics and/or logic are used to (partially) automate decision-making. Simple math can give you “a” forecast. Sophisticated math and logic can give you the best forecast. Simple heuristics can give you “a” production schedule. Sophisticated mathematics can give you a feasible and optimized schedule.
This is a true buyer beware environment because individual business situations vary widely and solving to feasibility and optimality are very, very difficult problems. The redeeming feature is that the rewards from selection of the right tools and their proper use far outweigh their complexity and cost.
The ways in which data is analyzed closely mirrors specific business decisions. The tools are used to analyze problems at relatively low levels, specific to an individual plant or at very high levels including the entire supply/value chain. The objective might be to optimize customer service, maximize profits, minimize costs or maximize asset utilization.
The other class of supply chain systems automates external transactions to the corporation - between suppliers, customers and third party providers. EDI was one of the most deliberate attempts to provide a structured protocol and technology for these communications. The Internet standard, XML is growing in popularity and has become an alternate method of communications.
Supply chain systems technology will continue to evolve in several directions:
· Protocols will become more standardized and easier to deploy.
· More external data will be combined with internal data to provide insights on competitive situations and wider visibility of overall supply chain costs & profits (including service providers, suppliers and customers).
Enterprise Information Portals
The new buzz is about creative uses for the Internet. e-Commerce, B to B, B to C and portals are some of the terms characterizing different types of activity that utilize the Internet. Powerful computing capability combined with a low cost, an extensive communication system (Internet) opens up the potential to achieve enormous capability. The concept of an EIP embodies this capability.
Imagine bringing together information from multiple internal and external data warehouses, from specific industry databases, governmental and non-profit organizations automatically to your desktop based upon the business questions that you need to answer today. The central portal question is how to efficiently provide the right information, at the right time to the right people.
The right information is a combination of raw data and processed data. Raw data can be aggregated, filtered, sorted, structured and converted to common units of measure. Processed data means derived information from sophisticated mathematics and/or logic. Further, the information must be succinct and relevant to the task-at-hand.
The right time depends on when information is needed. Some information is scheduled to be reported at routine times. Additional information is queried on an ad hoc basis. Portals will automate scheduled tasks. Instant messaging in combination with wireless Internet access will be used to send the information to you no matter where you are based on the schedule or predetermined alert situations. Computerized process control systems have had this type of architecture for more than a decade and would be a good model of how to effectively design and implement a real-time portal system.
Planning Practitioners make ad hoc queries based on the dynamics of the marketplace and supply chain. Users need intuitive tools to help them formulate abstract business situations into information requirements. Built-in intelligence into portals will make problem formulation easier and ultimately speed decision-making. The better portals will build more intelligence behind the user interface to speed up and automate all tasks and particularly the ad hoc ones.
Ad hoc queries and calculations are not necessarily real-time. A good example of this is a stock screener like http://moneycentral.msn.com/investor/finder/customstocks.asp Statistics, charts, comparisons and projections are combined to provide a better picture about stock value. The user is completely in control of what data is brought to their desktop. The users are becoming more demanding that the data be more current and more insightful so that they can pick stocks faster. Timeliness is directly proportional to the value of the input information and the decisions. Portal designers will have to anticipate the questions and provide options that can be easily accessed and combined with other parts of the process.
The right person implies that different people have different needs. It also implies that sometimes people work independently and other times collaboratively. In the stock-screening example, every investor can set up their own criteria for finding stocks that meet multiple customized objectives. Once the criteria are set up then the underlying data and the results will dynamically change. This is an example of independent work where there is “one-to-one” relationship between the information and the person.
When there are primary and secondary people involved in either the input, decision-making and the execution parts of a process, the portal needs to integrate functions that facilitate collaboration. Eventually application sharing, file transferring, chat (text & voice), screen sharing and e-mail will be seamlessly integrated to portals, as devices to enhance collaboration. This is an example of a “one-to-many” relationship.
“One-to-many” relationships are a magnitude more complex than “one-to-one” relationships. A process must be designed to layout the sequence of events. Timing of tasks, that various people perform, need to be defined, monitored and controlled. Access to various parts of the system is limited to ensure data integrity and assign responsibility. Layers of security may also need to be established.
Portal Implementation Process
Despite the unlimited number of potential applications and uses that there may be for this capability, we can speculate what it takes to realize the benefits.
Step 1 – Determine the business question(s) to be addressed.
Step 2 – Map the process(es) necessary to answer the question(s).
Step 3 – Determine the data required to answer the question.
Step 4 – Determine the source of the input data.
Step 5 – Determine the technology to acquire the data from each source.
Step 6 – Condition the data from its unstructured sources to a structured environment.
Step 7 – Develop and/or integrate the decision support or supply chain systems as needed.
Step 8 – Determine the format for how the results are to be presented to the user on their desktop or mobile device.
Step 9 – Automate the data feed at the frequency required for the situation surrounding the business question.
Step 10 – Determine if and how to integrate with other business applications that could use this same information.
Step 11 – Determine how often and when this information should be updated in the other business applications to synchronize data.
Step 12 – Dynamically change the results viewed by the end-user based on the changing situations internal and external to their function or business.
Portal Evolution
Steps 1-6 are required for the first generation of portals. Steps 7-9 pertain to second generation EIP or Intelligent Portals. Steps 10-12 are needed for third generation EIP tools or Dynamic Portals. The evolution of EIP will occur by the maturing of its attributes:
· Increasing breadth of data sources.
· Creative ways of presenting the data that provide a better picture.
Summary
It should be noted that all the technologies mentioned in this article are complementary. There is no implication made as to which technology should be deployed first or second, one over the other or which is more valuable. This depends on the individual situation of businesses at point in time.
Deploying advanced technologies is like trying to grow the most productive crop for a season. A number of things need to come together to yield the best results – soil conditions, fertilizer, pest control, weather and type of crop selected. There is not one that is particularly more important than the others. The right combination & proportions of the necessary elements provide the best results. Just as the farmer needs to make choices along the way, so do the people that select and deploy technologies at different times. Those with the best understanding, plans, selection of technologies and execution will be the most successful.
Editors of KMM. “Through the Looking Glass”, Knowledge Management Magazine. February 2001, pp. 28-35.
Hammond, Chad. “The intelligent enterprise”, Info World Magazine. February 5, 2001, Volume 23, Issue 6, pp.45-46.