SINGAPORE (03/21/2000) - Companies in Singapore are experiencing challenges in the form of changing markets, increased competition, faster production cycles and a tougher business climate. It is imperative that these organizations provide the right people with the right information at the right time. It is here that business intelligence (BI) products play a key role.
Business intelligence is about gathering, managing, analyzing, and distributing vast amounts of data in order to gain insights that can be used to drive business decisions. Other terms often associated with such activities include data warehousing, data mining, decision support systems or knowledge management.
In order to get to the bottom of all the hype and find out what business intelligence really means to the local community, Computerworld joined hands with research organization Input to organize a roundtable discussion. We have brought together key vendors and users in this arena to thrash out the issues surrounding this hot technology.
According to Input, the key benefits of BI include:
-- better planning/forecasting
-- more informed decision-making
-- more accurate information
-- improved business control
-- able to see trends clearly
-- better information management
-- increased efficiency
-- better use of resources
Only 26 percent of companies in South Asia have installed BI systems, Input said.
Gerald Wee, Editor, Computerworld (Moderator)George Kane, Regional Manager, Southeast Asia, Brio TechnologyPeter Kokinakos, Marketing Director, Asia-Pacific, CognosLow Keong Keong, Senior Manager, Information Technology, Singapore TechnologiesPat Merican, IT Manager, IkanoJason Oakes, Director, Technical Services, International Distribution Division, Business ObjectsReid Rasmussen, Managing Director, NCG InputSrikanth Ramachandran, Manager, Software and E-business Marketing, IBM SingaporeWong Cheong Phong, Vice President, IT Asia-Pacific Drive & Media Operations, Seagate TechnologyYew Bok Chin, Director, Computer and Info Systems, Temasek PolytechnicGerald Wee: The definition of business intelligence (BI) has often been linked to data warehousing and data mining.?
Jason Oakes (JO): People get caught up with the various terminologies that exist within the marketplace. Even within each term, like OLAP, there are various ways of doing it and various delivery mechanisms. BI is probably larger than the sum of all its parts - data warehousing, CRM, ERP, data mining, static cubes. It is all part of the same theme - making effective use of information that exists within an organization - no matter where it comes from. The motivations are very different, but it is all just making better use of the information that exists.
George Kane (GK): The average business user thinks BI is Dow Jones or Bloomberg. So, BI as a term is meaningful only to people within the industry, and it needs to be more specific, like reporting and analysis.
Wong Cheong Phong (WCP): I share that point of view. BI needs to be addressed from the user standpoint. The technology and tools are all there, but you must know what you want from it. For example, we are running the company based on a lot of performance metrics, and how you measure it. These need to be displayed and sent to management who must run the business better based on that information. To me, that is BI. We don't worry about the technology as there are many products available, but we want to be able to define what we want from it.
Peter Kokinakos (PK): Business intelligence has different meanings. It is not the technology, but as a user, what do you want from a job when you wake up in the morning. If I'm the CEO, I want to know what my performance metrics are. If I'm the audit clerk, I want to know what the status of my inventory is on the shelf. So, BI means different things to different people in context of what job they're trying to do.
So, BI is all about exploiting information. Over the last 30 years, you've had hardware and software vendors dominating the space of IT infrastructure because they provide the repository for this information to live in. And teams of people have been pumping information into the computers, and yet no one is pulling it out. What we've seen is the liberation of the user who needs to get something back now, and that's what it is all about, be it through ERP, CRM, or whatever term is coined. We want to give users information in order to do a better job.
Srikanth Ramachandran (SR): I tend to agree. One study has it that only 7 percent of information is being analyzed. So, if you take a step back and you want to define business intelligence, I would say that the very first step is to define what you need to run your business in the knowledge economy. And then you look at where you find this information - is it from Bloomberg, from your own ERP systems, external sources, vendors that you need to interface with - and then you ask what you do with the information, and how it drives the business. So, it is the whole process of defining your need right up to analyzing it and making decisions based on that.
Yew Bok Chin (YBC): The word "intelligence" is quite scary. It can be something to do with hidden knowledge - when people talk about CRM, data mining, user profile analysis, you are trying to find some knowledge that is yet undiscovered, and hopefully it is useful for business and work.
Low Keong Keong (LKK): Data warehousing has been around for some time. BI is a new buzzword that is packaged by vendors to implement it - you've to buy a data warehouse, get OLAP - but what you actually need is hidden knowledge that is not at the hands of business users. So, how are you going to bring this knowledge to them and that's business intelligence.
PK: I think it is more than just the hidden knowledge. That is the promise of data mining - trawling through reams of knowledge to find the nugget of gold that is going to change the corporation and how they work. The fundamental problem is that we are in an information economy and that is the currency that people are dealing with. You need to be able to distribute it to every one in the organization.
You talk about EIS which was great in the days when companies were organized by military structures. You had a very narrow and steep reporting structure.
Everyone at the bottom was preparing reports for the next layer up, and at the person at the top would make a decision, and that would be implemented down.
What we're seeing today is that the flattening of that structure, so I can't depend on someone above me to tell me what to do. I need that information myself. EIS should no longer be executive information systems, but everyone's information system. It is not just hidden information which we are trying to deal with, but the amount of information a business needs to run. It is the lifeline of the business now. That's the essence of what BI should be about.
GW: What is the most pressing BI need today?
PK: Recently, with the push into the e-world, the Web is basically done a lot to progress the distribution of information. That has made the distribution much more streamlined, and awakened of the BI market. The real issue is what's the difference. If I'm a bank and I implemented an ERP system that is the same as the other bank, where's the differentiator? Our processes are the same, and the information that we use is the same. The difference is which information to use and how we use that information. That is what people are struggling with now.
SR: The distribution has been simplified. If you walked into a supermarket two years back and asked them who their customers were, they would only tell you male or female. They didn't have any other information beyond that because they would never associate a transaction with who was buying it. You just go through a counter and pay for it. Now with the Internet, they can tell you the profiles of customers by age groups, what they buy, when they buy, and how they buy.
Internet has not only simplified distribution, but also gathered a lot more.
YBC: From past experience, we've been creating a lot of databases, islands of automation. Each application houses a database. By integrating the information together, you can view it from different perspectives, and see how it can be used.
With the proliferation of the Internet, there are tons of information outside, and the connectivity creates new opportunities. We also need to pull information from external sources, and to digest it, and see how else we can mine the information, discover new opportunities, and react.
LKK: Data mining has been around for some time, but the problem is that it is not packaged to the needs of the recipients, and they need to do some cleansing. The pressing need is how to get CEOs and top management to use BI.
Now, if they want certain information, they just call in the CFO. So, we how can we take whatever the CFO knows, and put it on the desktop so the CEO can retrieve it instantaneously. Timing is crucial for BI. If the CEO wants it now, tomorrow would be too late.
WCP: Our experience is that we have many companies and factories, all measured differently, so how are we going to manage it all. The difficulties we are facing is defining the set of parameters, how we measure our performance, and how we operate from there. It is actually driven from the top rather than from IT. The management now has to change their mindset and the way they operate.
Rather than asking for information on an ad hoc basis, they should define the information they need and use the business intelligence online and make the decision quickly. From the IT perspective, that is always the challenge because when you try to ask management, you never get the same answer.
Pat Merican (PM): The problem is more the mindset of the users, who will ask for more and more information. You can have all the information you want when you want it, but the decision process is the bottleneck.
Sure, you know how many people are buying this table or that curtain, and for what reasons, for which household, and for what family group, and which income level. We know most of this already. The point is, what next? We need to know what is the fundamental information that is required to really make a decision.
It boils down to one page.
The second problem we encounter is that the supplier of the technology, the integrator of the technology, is the worst criminal out there. They will sell the product at whatever discounts are agreeable, but six months from now, we will regret our decision to choose them. So, we must go through a rigorous process to choose as an integrator. If you survive eight months of visits, you could get the deal.
PK: The tools or information shouldn't get in the way of the decision making process. What we've seen is that users go through a maturing period - as they understand more about the information they get, they start asking harder and harder questions about what they're doing. Yesterday, we ask how many tables we sell, then today, we ask who bought it. It was a good question yesterday, but today, it is only part of the big picture. As I get used to something, that becomes second nature. The more you give users, the more they want. So, you've got to get them to self-service that, and in doing so, it will mature their knowledge and understanding of the business.
GK: Decision-making is an iterative process. When you get back the data, that prompts a different way of looking at things, then another question comes to mind, and ultimately, you arrive at a decision point. What these tools do is help you along that decision-making process, which then has to be a human process. The tools won't make the decision for you.
JO: There is a problem with expectations of what these tools are supposed to do. They are not going to find the answer for you because you need to know what you're looking for. There is a growing area of consulting within the industry in helping the users of the software understand the questions they want to ask.
You can give someone the best product on the market, but it won't be worth a pinch of salt if you don't know what you want to do with it.
The answer is not one number on a page, or a million-row report, but in understanding what questions you need to ask, and the relevance and implication of the answer that comes out of it. A software tool is not going to help you run your industry for you. You have to know that yourself, so the role of the BI tool is to reduce the vast quantities of data that exist into a smaller amount of useable information based on what you know about your business and the things you're looking for.
That essentially boils down to making decisions, and the decision making process is inherently human. The things we provide are just cogs in the wheel.
To a certain extent, people are using technology as a crutch to fill the spaces where they should be thinking harder.
GK: That leads to a big question - whether the business community understands what the benefits are of what we're talking about. They don't understand something like doing an OLAP analysis on an inventory turnover problem. This is obviously true at the SME (small and medium-sized enterprise) level, but even in big corporations, you have to explain how it can help minimize inventory and maximize cash flow, or whatever business need there is.
There is a lack of appreciation of what BI is. Just ask a guy: "What are the five things you need to do with your business to maximize your profit or whatever your goal is?" I guarantee these tools could help him get there, but he wouldn't think about it when he asks for a BI tool.
Reid Rasmussen: The key aspect of BI is the intelligence part. It is not about business information, it is about business intelligence. And one group that has always impressed me are brokers because all they care about is getting information in the form of one of three words - buy, sell, hold. That's all they care about, not the 1,000 page report. Coming up with a one page report is usually more difficult than coming up with a 1,000 page report. The key thing is not so much being able to get information, but the right information, at the right time. In today's e-world, that is one of the biggest contributions that business intelligence vendors are going to be able to bring to the table.
GW: What is the key issue in BI?
PM: Users have to now go back and figure out the business metrics, the one-page decision point. Users have to start to reflect and ask if the information they are using is relevant today. If not, just move on to another decision point.
The IT department has to take on a different role, not just knowing the technology, but also consulting, to know how to deploy tools in the appropriate way. Then we need to close the loop and see how well the decision is really flowing down the line. If we don't then, we may be using wrong reports for a long time. So, users have to learn to get their act together.
BI is not just internal to a company. A lot of work is done by business partners and there is a need to access their information, to integrate our systems with the our suppliers. Business-to-business integration is very important as you have to find a way to collate information to one view.
WCP: I want to go back to the mindset and cultural aspects. I'm seeing the role of IT as very different from many years ago. There is involvement with the operations, knowing what is going on, and how the business is run, so that we can develop that one-page. Also, the person that runs the company has to know what is available out there. I want to know who the vendors sell their products to - IT or top management and educating them on the products. Because if they're selling to IT, they're one step behind because IT has to go back to management and sell a product which they do not know as well the vendors.
PK: Typically, we will address the end users because they have the pain, and if you relieve it, you get the sale. The problem in a lot of companies is that the end users can't buy the product and just install it in the machine unless IT puts its stamp of approval on it. Sometimes, end users like the interface and simplicity, but IT is more interested in features because part of their job is to make sure it will not just address the user's requirement today, but for the next 400 years. So, they come up with a checklist of items, and that is where you get stuck in the IT loop, where they are trying to buy a piledriver, when a hammer will do. So, there is a disconnect, and the user request to just knock a nail in becomes a much bigger issue.
WCP: Projects are really partnerships with the user. The user usually makes the call rather than IT, but IT provides the check so that users don't get carried away. The role has changed. Many years ago, I would make a decision, now every project is funded by a team which includes a business person planning the project with IT involved.
LKK: Most of the time, BI is sold to the wrong party, and even when it is sold to the user, it is still the wrong party, because those are middle management people. They try to figure out what the top person wants, and develop accordingly. But if the top person gives a different query, they have to scramble and try to find the answer.
GW: What are the tangible benefits of BI?
WCP: We have a set of metrics that we compress into one page. Management has come up with a set of metrics like inventory turn, accuracy, fulfillment, and so on, so we use that information to look at what is wrong with the business and how we can fix it. That is the benefit that we're seeing. And bonuses are tied to the metrics. Not being able to meet the target metrics means no bonus, so you actually move to efficiency.
LKK: We have many initiatives to start BI and certainly we see the benefits of it. If BI can bring the same benefits that a spreadsheet brings to individuals, enabling the user to get data whichever way he wants, that's a benefit.
PK: One of the things we see in our sites is that when you share information, you actually start breeding a best practices approach. For example, when you send out sales statistics for consolidation, if you're in the bottom half, all of a sudden, you're looking at guys in the top half, and finding out what and where they're selling and what is the mix of services to products, and what is different. This forces me to get a best practices model in place. This is one of the benefits that we see that may be intangible, but it is real.
RR: Having interviewed a lot of users about BI, the two key things that we found, particularly on the senior IT non-executives is speed - being able to get it quickly, and accuracy -- being able to get information that they knew was correct. Given enough time, they could find whatever information they need.
The key is: being able to get it quickly so that we can react to the problem.
JO: Talk to the business guy what he wants out of the BI solution, and he will say that he wants information that helps him drive his business. Talk to the IT guy, and he will say speed - sub-second performance. Obviously, speed is important. We are all living very fast, but a lot people seem to forget that it used to take us weeks and months to get this, and now if it doesn't happen in the blink of an eyelid, we will all get very upset.
People will commit millions of dollars on big hardware to do transaction-based systems - to keep the telephone system, the banking system, all of these going, and then they'll buy a 286 box with a copy of Windows 3.1 to run their BI suite off. It is damaging your own credibility and ability to meet your expectation.
If BI is so critical, you should learn what it is that is going to help you meet your targets.
Instead of squeezing 10G (bytes) of data on a Microsoft SQL Server and expecting it to turn around quickly, think about your expectations, and understand what it is going to take. Maybe, it will take a whole other server, but if it is that important, people have to pay the right level of attention to it.
I've walked into shops where people have complained about performance, and you ask them if they put indexes on their database, they complain that indexes take up too much disk space. Yet, this was a bank running marketing programs in a highly competitive environment which is worth millions of dollars. But if you're not willing to buy another disk drive to put your indexes on your database on, then decision-making isn't so important after all.
GW: What checklist would you use to make sure a user is ready for BI? How would a vendor approach a sale?
PK: Number one is a business sponsor to marry up with the IT department.
Someone from the business community that needs something as opposed to a project for a project's sake.
Also, a lot of information has never been aired before, and you don't know how dirty it is until you get it into the open. The challenge is to get all this great intelligence about your business, but you don't know what you have and how bad a shape their systems are in until you start scratching the surface.
So, understanding how clean or dirty your systems is a part of it. These are the two fundamental starting points, and from then on it is business needs and users' needs.
I would say that one product would not be able to service a large organization totally. Different users have different requirements, and you may have to get a data mining product from one company, and a different type of analysis product from another company, and a reporting tool from a third company. Some products may be generically good, but when you push the 10,000 user sites, it typically is not one product.
SR: To some extent, it also depends on the users' buying patterns. Some buy servers, and storage, then they add an interaction tool, and a database. There are others who buy an end-to-end solution. You define a business requirement and then do system integration. So, the checklist there tends to be different.
You use two different approaches here. It is not a simple product sale.
YBC: The needs of the organization is important because we can buy something which will become a white elephant. I would like to emphasize two points. One is the link or integration of databases. Second and most important is ease of use. The tool must be very easy and comprehensive for a user, so the user can take it and juggle and massage the data and come up with something to trigger and stimulate creative thinking.
LKK: The checklist I would expect is something that determines the readiness of an organization to do BI. Issues like whether a transaction system is in place, whether the culture is ready to support it. If there is such a checklist that can help us assess how ready we are, where are the gaps we have to close, that would be useful.
RR: There may be some differences in definitions of BI because to me, BI is at the core of the business. Every business needs to have good reliable information driving it, so it needs to have something in place where they get the information they need for decisions. It is not so much whether we are ready for BI, but what kind of BI are we ready for. Do we need something that is focused on giving key information to a couple of decision makers, or something that is pervasive across the organization. We have to get people thinking along the lines that every organization needs business intelligence, and which is the best solution to address that need.
LKK: There is a difference here. An organization needs BI, but readiness is the ability to implement the tools. Are they ready for it or is there some action they need to clean up their data and standardize the systems. Are the systems ready? Yes, they need BI, but they may not be at the stage where they can exploit the use of it.
SR: Back to the supermarket model, we need to see which types of customers buy which types of products. But if your transaction system is not collecting information about the customer, you're not going to get there, so you have to fix certain things before you get that level of information that your business demands.
JO: There is definitely a big role for systems integrators to provide value in helping you understand where your infrastructure and culture is at, and whether you've got the right level of information, and how accessible it is, and what you need to do to get it to a point where it can be easily retrieved, massaged, understood.
GW: What are the potential pitfalls in implementing BI?
PM: Transaction systems consumes a lot of compute power or storage. When you use BI, even when the right decision points and pages are made, you realize that the server really uses a lot of resources. So, the first thing that affects benefits is how you size the system. It can be like running a batch job in the end.
YBC: One of the worrying things is people's mindset. With BI, you are asking people to think differently, and how to look at data for decision-making. It has to do with mindset as well as the availability of information in technology.
JO: You might find sections of the company not willing to be open to the realities of the information age. You can find pockets of resistance to the change of the status quo.
WCP: One important point in implementing BI is who initiates the project - IT, higher management, or middle management. That makes a difference. Many times, it has to be top-down. The guy responsible for the vision has to know how he wants to run their business and what he wants to know, so if they initiate the project, the likelihood of success will be much higher.
SR: From the consumer viewpoint of BI, there are certain privacy issues when you talk about BI. There was a major controversy of a pregnant women registering herself at an Internet site. She had an abortion and six months later, the company contacted her as to whether she wanted to buy children's products. So, intrusion of privacy becomes a major issue. When you get into specific customers, the business has to ensure that there are enough safeguard policies in place so that you don't intrude on privacy.
LKK: The vendors have done a great job marketing BI and have created a lot of excitement. Unfortunately, that is not the whole picture. From day one, we have to let management know that it is just the tip, and there are a lot of things that need to be done before you get there. The pitfall is to ignore those other things.
GW: Any final comments?
JO: Regarding the benefits of BI, words like accountability, metrics, control, and monitoring came out. In Asia, we seem to still be focusing inwardly to be making sure how many widgets are being shipped, and how many tables are sold, which is important as it is at the core of business efficiency. But we are focusing inwards rather than thinking about the opportunity outwards.
In addition to ROI, people have started to talk about return on opportunity, what you would get if you went after something that didn't exist yesterday, and the opportunity cost of not doing it. We are finding that users of BI software in U.S. are finding ways to leverage the data warehouse that they've built and the information they've gathered from suppliers and customers, and provide it to their customers and suppliers over the Web, and building new business models to get new levels of customer service, loyalty, and satisfaction, as well as new revenue streams.
So, turning outwards can present an organization with opportunities to increase market share, revenues, and so on. Let's not lose sight of that.