Ziff Davis B2B is a leading provider of research to technology buyers and high-quality leads to IT vendors. As part of the Ziff Davis family, Ziff Davis B2B has access to over 50 million in-market technology buyers every month and supports the companys core mission of enabling technology buyers to make more informed business decisions. Contact Ziff Davis B2B 100 California Street, 4th Fl., San Francisco, CA 94111 Tel: 415.318.7200 | Fax: 415.318.7219 Email: b2bsales@ziffdavis.com www.ziffdavis.com Copyright 2012 Ziff Davis B2B. All rights reserved. ziffdavis.com ziffdavis.com Executive Summary As an SMB, what intelligence should be used for your business? Can specic tools make a real difference to the way that your business is run or even improve your protability? In broad terms, BI means possessing a complex knowledge of every factor that affects the business: your customers, your competitors, your business partners, the economic environment and the way the business is run. Having all of this information at your ngertips permits you to make quality and effective business decisions. It has also been described as the practice of giving assistance to businesses for the purpose of them gaining a better understanding of their methods. Also, specic goals are met through BI such as market share increase and customer service improvements. Much research has been done on the subject of BI and in every case it has been shown that this is the number one priority for every CIO. In addition to the observations of CIOs, it has been shown also that many companies are spending a great deal of money on tools that purport to generate business intelligence from raw data, but which fall short of this goal. In recent years the increase in standards, the addition of automation and technologies that facilitate data collection all mean that an exponential amount of new data has become available, along with an exponential increase in speed and gained storage capacity for all the new information. Add to this scenario the rapidly increasing demand for information, spurred on by the equally rapidly exploding mobile arena, where, if it is on the company servers, and it is not on any member of the companys mobile device, be it Smartphone, iPad or another of the ever increasing tablets that are around today, this is points off for the CIO, or in fact the entire IT department. Business Intelligence provides the capability for an organization to convert all of its facilities into data, which in turn, is translated into useful knowledge that can be applied to further the organizations business by means of imparting this knowledge to appropriate parties at the appropriate time. 1
The only thing that can be added to this is the fact that the information that is required is highly specic and gets old very quickly. This is to say that information must allow timely access for it to be useful. So with that in mind, business intelligence has to be smarter, more mobile, and easily accessible to all the right people. According to Gartner, the capability for appropriate access will lead to the generation of new opportunities through the use of specic business rules. Business intelligence (BI) platforms enable all types of users from IT staff to consultants to business users to build applications that help organizations learn about and understand their business. 2 The way that this is implemented will provide the technology that gives us specic views of business operations: it will show past, present and future predictions of these. It achieves this through Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 2 of 10 ziffdavis.com Page 3 of 10 reporting specic information using business analytics, data and process mining, complex event processing, benchmarking and predictive and prescriptive analytics. In its simplest form, business intelligence sets out to be a decision support system. That is to say that business intelligence is in place to assist executives in making informed and protable decisions. They can do this through the use of certain commonly accepted practices which use certain specic technological tools to do so. The term, business intelligence, can mean both internal and external processes, for the purpose of internally improving the decision-making of an organization, or externally weighing up the competition and using intelligence to do so. The term, business intelligence, can sometimes be used synonymously with the concept of competitive intelligence. In this compendium of business intelligence tools and best practices, we shall examine which tools are available for your business and how you can leverage these to maximize protability. Read on to learn about BI best practices for your business. What Exactly Are BI Tools? In todays complex arena, we see plenty of business intelligence tools. Some claim to do specic tasks, like manipulating rows and columns as in a spreadsheet, and others will take all of this one stage further and collect data from several different sources, manipulate it so that it can be presented in many different formats for different uses under different circumstances. SMBs have a specic use for BI. That is to say the specic uses for BI that SMBs rely upon are benecial to SMBs in a way that is not the same as in an enterprise-level business. Nonetheless, they are equally important. In fact, they are more crucial to the success of an SMB because so much is riding upon such a small amount of data by comparison. Some of the most stringent scenarios for data processing, appear in sales and marketing, where data will dictate how a product is sold, to whom it is marketed, and keep track of results-oriented progress while this is going on. Much of this must occur in real-time or something very close to it. It goes without saying that this kind of data directly affects the bottom line. Therefore it must be accurate, and it must appear in a timely manner in order for companies to make use of it. The most convenient BI tools are those that combine existing technology and accept multiple formats, translating them easily into an acceptable output that will be fed into whatever method is required to get the reports out to the target. For instance, Oracle Business Intelligence Publisher 3 is one such tool that combines multiple inputs from various sources and allows the choice of multiple outputs, depending on the requirements. Another such system is Microsoft SharePoint Server that permits collaborative working and multiple formatted outputs including SOA (service-oriented architectures) and other formatted information which is all actionable. Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business ziffdavis.com BI tools are applications that are designed to work with data. And the production of new data has been facilitated by a rapidly decreasing cost of storage, the increase of easy data acquisition through the use of software packages that use new techniques and shared APIs from open sources. Other technologies that have grown in recent years have improved the collection of data, like banking transactions, transactions online, inventory tracks by RFID and the logging of queries on Web sites, product review comments and even blogs. The data could have been gathered previously, or it may be more immediate. The most basic form of classic BI tools available is in the form of ofce suites. The most important BI tool is one that queries and reports its data. Raw data can be pretty meaningless until it has been presented in a format that makes sense of it. Microsoft Excel produces reports in the form of charts, that produce the visualize analysis of columns and rows of data in many different forms, from scatter plots to pies and bar charts. These can also be represented in 3 dimensional forms, in addition to the standard default 2D representations. Another Microsoft Ofce Suite application is Microsoft Access, which can be used to query and report the data automatically. This software extracts, sorts, summarizes and presents the data in a form that can be customized to highlight salient information. Tables stored in this database can be queried to amplify the existing information to give it more relevant meaning. In fact, whole applications that work automatically to provide information can be created using a database such as Microsoft Access. And all of this data can be migrated at a later date to a commercial weight database such as SQL server. Meantime, there are many different ofce suites available, ranging from open source, no cost or low cost software, which in most cases shares functionality through its le extensions, and is easily portable from one system to another. If an SMB has to be particularly cost-conscious, sometimes these systems are a method of keeping costs down while being able to work with data and documents. Nonetheless, it is the cost of the work which becomes expensive when a non-standard ofce suite is used for reporting. Lets face it, a mature product such as Microsoft Excel will have had many years of development and will therefore have an ease-of-use factor which is far higher than anything that is in an open source suite. Also, however easily facilitated the conversion from one system to the other might be, the effortless recognition of a Microsoft document, could well save the day when minutes count. Surprisingly enough, even for SMBs, the high cost of maintaining ofce software which does only half of the work, is prohibitive. To the people who purchase the software, is commonly attached a mindset that foresees nothing but ease-of-use with ofce software and report production. In fact it is quite often seen as a low-cost item which does a lot for the outlay. The problems occur when this software has to link to another piece of software that might be bespoke for some reason. Even QuickBooks is hard to query and does not output in a format that can be easily translated into the ofce system, despite the promise of easy export formats. Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 4 of 10 ziffdavis.com Maintaining and running many different reporting systems under the same IT umbrella, is costly and needs skilled personnel who are familiar with these software packages, down to a level which is under the hood rather than behind the wheel. So, in the time that it takes to engage the under the hood specialist, and teach them how one particular organizations reports are done, a fair bit of expense has already been racked up. What is really needed is a one-stop source for gathering data and generating reports in a timely manner. It seems almost incredible that this whole business intelligence topic has been discussed for well over 10 years, and so far, very little progress has been made to simplify it. Frequently, the quickest way is the slowest way for SMBs. And what might well happen is that an under the hood specialist is engaged from a company that provides a complete one-stop solution, and who installs the hardware and software to support this, while training the trainers and giving the company the impetus to get started with a new system and carry on once the under the hood specialist has left. After this has happened, quick refresher courses are available and representatives from the company who supplied the system will revisit to ensure that its working the way it was intended to. In addition to querying and reporting software, we will often encounter OLAP software (online analytical processing). This software is a way of preparing multi-dimensional analytical queries and producing reports that encompass relational reporting and data mining, much of which is in real time. So, let us suppose that the sales team wants to see how each division is performing while they are working. OLAP will be presented as a virtual dashboard, that is the rst thing that will be seen when the computer screen is viewed. This is useful for marketing, business process reorientation, or business management reporting. OLAP typically has a fast execution time. In the previous paragraph, dashboards were mentioned. These are typically a real-time user interface often on a single page and show KPIs (key performance indicators), and a graphical representation of the current real-time status, either with historical trends shown in the same area or quickly clickable links. These enable instantaneous, informed decisions to be made. Frequently, these kinds of screens will be used in money management, market trading and sales. Along with OLAPS and digital dashboards, other types of BI tools are data mining, data warehousing and decision engineering tools. Many of these comprise component parts of ERP systems (enterprise resource planning). Data mining is literally mining for data or going out to nd information in the massive heaps of data that have been gathered. The process is far simpler than this since bespoke software has been created that does this easily. Along with data, there is also process mining and this is used very successfully to analyze business processes by electronically analyzing logs of events that have occurred. The purpose of this is to improve business processes by examining interactions of organizations and social structures, through the detailed analysis of captured event logs. Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 5 of 10 ziffdavis.com Lastly, BI tools would not be complete without business performance management tools or local information systems tools (LIS), which, in terms of information technology, are a niche form of information system. These report geospatial and geographic information to offer knowledge management functionality in the storage and retrieval of unstructured data. Most the time, the systems will describe situations occurring in specic geographical areas. These last systems have been specically used in government driven projects mainly in the United Kingdom. Current BI Techniques Here we will examine traditional BI techniques used to manipulate an organizations data that has been collected. Most all of these methods are used in current business practice, and may well also be included in a proprietary all-inclusive system that is capable of much more. Regardless of the software used of which these techniques comprise, these classic techniques typify business intelligence in essence. These are also components of data mining and which all spring from traditional statistical disciplines: TECHNIQUE DESCRIPTION Predictive modeling An attribute of a particular datum can be predicted as having a specic value. Classication Data has been gathered and is ready to be included in a set. Classication determines the class or set to which it belongs. Model visualization Data that has been transformed into knowledge via an information layer, is made easy to understand using visual objects such as histograms, scatter diagrams, bar and pie charts and other plots. Characterization, data mining Data is distributed, dispersed or made exception of. Clustering, analysis of outlier Data sets are partitioned into classes. Items with similar characteristics from the sets are clustered together. Association, correlation, link and causality analysis Data attributes are examined and relationships between them identied. Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 6 of 10 ziffdavis.com TECHNIQUE DESCRIPTION OLAP, MOLAP, ROLAP and HOLAP (online analytical processing) OLAP allows BI practitioners to analyze multi-dimensional data from the point of view of different dimensions. This is probably the best way to examine trends and milestones. EDA (exploratory data analysis) EDA is exploratory by denition, and determines identied patterns literally without depending on assumptions or models. OLAP Technologies Online analytical processing can be useful for examining real-time results and gives a very rapid snapshot of the whole system, and this permits decisions to be made quickly based upon trends, compared with historical statistics. When considering data warehousing concepts, three methodologies for online analytical processing emerge: MOLAP (multidimensional OLAP), ROLAP (relational OLAP) and HOLAP (hybrid OLAP) all of which process using slightly different ways of doing it. MOLAP MOLAP servers directly support the multidimensional view of data through a storage engine that uses a multidimensional array cube. They typically pre-compute large data cubes to speed up query processing. Such an approach has the advantage of excellent indexing properties and fast query response times, but provides relatively poor storage utilization, especially when the data set is sparse. Source: Association for Computing Machinery ROLAP ROLAP technology manipulates the stored data to appear like a traditional OLAP that slices and dices. What makes this work is that every slice and dice is like adding a WHERE clause in SQL. Pros: The advantage of ROLAP is that it is limitless in its capability of handling massive data sizes. Also it can use functions that are already in a relational database. Cons: the disadvantages of ROLAP is that queries need to be optimal at all times because performance will always be impacted by large amounts of data. Source: 1keydata Copyright 2012 Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 7 of 10 ziffdavis.com | Copyright 2012 Ziff Davis, Inc. | v071712 HOLAP HOLAP technology is a combination of MOLAP and ROLAP attempting to provide the best of both worlds, leveraging cube technology to increase performance levels. Source: 1keydata Copyright 2012 Real world BI Tools scenarios You may be looking at this description of all the tools that you can leverage for your business and you would be justiably wondering just how these could be implemented to improve the current situation. Here is one scenario. A company purchases a new BI system which is a complete one-stop, turnkey application that requires nothing but legacy data, which might include a number of internal sources including what has been gathered for years in the order processing database. Also included in this legacy data, would likely be data from external sources, like order data from the companies direct vendors, payroll (particularly if its handled by some external agency like many companies do), and perhaps some other sources of external data such as customer satisfaction data or any form of spreadsheet that shows actual historical results over a number of previous years. Now heres where the magic occurs: the new system seems to be running OK. Middle managers see the system bedding-in well. Because they are busy, (and who isnt?), it takes a little while for them to notice the correlation between customer dissatisfaction and periods when less product is leaving their warehouse. At the same time they notice that customer orders are on the increase, and their warehouse is booking fewer hours. In fact, the warehouse is almost idle. Putting it all together, these managers realize that they need to implement a just-in-time warehousing system, but, way before that, they need to examine the orders that are getting out quickly as opposed to the ones which are causing customer dissatisfaction. The bottom line is that the orders that are getting out quickly are never entering the warehouse because they are being drop-shipped! With business intelligence in place this company has all the tools they need to reengineer the business process where orders either enter the order processing department and get booked for distribution from their warehouse, or alternatively drop-shipped directly to the customer. In todays arena, many enterprise-level companies have adopted this methodology and have been using it for years. Prior to this SMBs have utilized software such as Casewise Modeler 4
(business process reengineering) in order to nd the weak spots. Casewise Modeler actually produces a visual representation of the entire process from start to nish, so it is fairly easy to see which parts of the process could be dened as bottlenecks and which areas are moving too fast for the next department to be able to keep up with their speed. Of course, it is easy Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 8 of 10 ziffdavis.com to get the idea that a little bit of rearrangement from one department to another, possibly taking employees from the department that is moving too fast, and placing them in the slower-moving department, will ultimately correct the bottlenecks and also speed up the whole process, therefore gaining customer satisfaction and increasing protability. Requirements For BI Tools 1. It must provide a framework for data processing. This means that it must be able to gather data, process it in a way that provides meaningful information that can be used to improve business processes. 2. It must provide a suitable output format that ts in with the organizations current multiple endpoints, formats and delivery methods. This means that, once the data has been processed, it should be able to be delivered to end-users, whether they be company employees, or customers, in the format and delivery method of their choice. 3. It must provide the requisite data capability. This means that it should be capable of handling the extent of legacy data that currently exists in the organization and have provision for the addition of a far higher volume of future data. If this is not a current possibility, there should be an easy upgrade available to handle any eventualities. 4. It must be easy to use and not require extensive training or retraining in order to perform basic operations. This means that whatever package it may be, it must be able to be installed without negatively affecting the present data-processing. Once installed, it should be easily implemented with the existing software framework and should be able to start work immediately processing all legacy and current data products or in parallel to permit evaluation before going live. Software should be intuitive and not require enormous manuals and lengthy training classes to be able to used. 5. It must be portable and exible, so that should the organization decide to change their platform or operating system, the software will move with it. This means that if the company servers are currently Windows servers, the data that is owned by the company now should be able to be moved to a Linux or other server system with little or no loss of time or data. BI Suites When considering implementing BI in a green eld situation, a new implementation of software that covers all the ground at once is probably the best solution for SMBs. Whether it is Microsoft SharePoint 5 , Oracle BI Publisher, or another COTS (commercial off the shelf) solution, the environment should be delivered by one system and allows multiple endpoints, formats and delivery methods. Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 9 of 10 ziffdavis.com A consolidated BI suite should provide a data-gathering and reporting tool that permits an organization to meet requirements for the creation of all types of documents. These would include but are not limited to: nancial documents, including checks printed with MICR fonts and labels with embedded barcodes and QR codes. And this could also include XML, EFT or EDI. Conclusion Little doubt exists that SMBs must use BI in order to leverage the maximum from the data that they gather every day in order to survive in a down economy. Nonetheless, some judicial weeding out of systems that will not help, because they are not specically targeting the information that is required, will go a long way to lowering costs and providing quality information which can be used to increase the bottom line. The techniques and systems described in this paper will be able to be assessed in the light of the readers requirements for information intelligence and their use of BI. About the Expert Steve Tufll, is a freelance technical and creative writer based in Valencia, CA. Steve has worked in business computing for some 25 years, and has been writing since 1998 covering many technical and business-related topics. He has written for Toolbox since 2006. You can reach Steve at stevetuf@gmail.com. 1 Business Intelligence (BI) can only be culled from complete information - information that spans across the range of functional silos that make up the enterprise, organizes for clarity and reliability, provides true decision support, and lays the base for competitive superiority and action. It is involved with various applications such like ERP, CRM and SCM applications which organize the vast amount of information for analyzing the business activities. Every decision must be customized to the companys own requirements and resources. (Vivek Jaglan et al. / International Journal of Engineering Science and Technology (IJEST) Improving Performance Of Business Intelligence Through Case Based Reasoning - Introduction) 2 Gartner Inc., Magic Quadrant for Business Intelligence Platforms, Rita L. Sallam, James Richardson, John Hagerty, 6 February 2012 3 http://www.oracle.com/technetwork/middleware/bi-enterprise-edition/overview/index.html 4 http://www.casewise.com/products/modeler 5 http://sharepoint.microsoft.com/en-us/Pages/default.aspx About Ziff Davis Ziff Davis, Inc. is a leading digital media company specializing in the technology market, reaching over 40 million highly engaged in-market buyers and inuencers every month. Ziff Davis sites, which feature trusted and comprehensive evaluations of the newest, hottest products, and the most advanced ad targeting platform. Ziff Davis B2B is a leading provider of online research to enterprise buyers and high-quality leads to IT vendors. More information on Ziff Davis can be found at ziffdavis.com. Ziff Davis | Best Practice | Best Practices for Business Intelligence in a Small or Medium Business Page 10 of 10