The development for businesses in the age of big data is the availability of business intelligence software to help you get the most value out of the data that your organization is collecting. For a long time, business users who wanted analytical data would need the help of a data specialist, either in the business or in IT, who had both technical and business knowledge. The technical knowledge was needed in order to query the database effectively, and business knowledge was needed to refine the results or modify the search to produce the data that was needed – which was not always precisely what was asked for.
Retrieving information could be slow, so a dedicated specialist was tapped to allow users to continue doing their regular jobs as they waited for the data. Between the slowness of querying a large dataset and the fact that all queries had to go through a specialist, bottlenecks were created at the data retrieval point – bottlenecks that became unacceptable the more organizations relied on big data analytics to make decisions about business strategy.
As a result, software has evolved to provide business intelligence based on big data – and smaller companies are finding this software helpful in analyzing their small data as well. BI software provides analytic capabilities that provide insights into one’s business, making strategic decisions more fact-based and reducing overall risk.
Many business intelligence (BI) solutions exist, but choosing among them can be difficult. Some software focuses on providing robust back-end capabilities, and good integration of existing systems; some focus on providing intuitive interfaces that empower users to dig deep into data, some focus on security features, and others focus on reporting capabilities. How is an organization to choose among them?
After identifying your organization’s needs, one way to evaluate BI software is by examining its feature set. In truth, good BI software should provide all of the above: architectural features that ease integration, development and growth; security features; and reporting features. The following is a guide to a good robust feature set that can help you identify the BI software that will help you get the most value out of your data.
- Open architecture: Whenever you have a choice between proprietary and open architecture, it’s wise to choose open architecture.This not only ensures better integration with existing platforms, but is also a sound way of preparing for the unknown future. The open platform approach means you are not tied to a single path, a single vendor, or a single architecture. Open architecture and open integration will make it possible to integrate data from databases, email accounts, social media, and other sources, which is a key to modern business intelligence.
- Broad database support: Staying flexible also means being interoperable with a wide variety of databases. This usually involves an additional architectural layer that allows your applications to interact with the database or databases in use. Not only should your BI solution support any database you may have in use or wish to put into use, your BI applications may be pulling data from multiple databases and types of databases.
- Data mart and data warehouse architecture: Incorporating these structures enable the BI tool to process data from multiple sources. These structures incorporate data in a single location where it can be consumed by the BI system for reporting and analytics.
- Mobile support: In this era of BYOD and a recognition that business value can be gained with small, mobile devices like tablets and cell phones, a good BI solution should be available on a wide variety of devices, providing a complete native experience and set of features, like reports, no matter what the device.
- Cloud-ready systems: Another technology that you either need to enable now, or will probably be using in the near future is cloud computing. Especially useful for big data analytics and applications that must be available 24/7/365, capabilities of the cloud are expanding as security is becoming more solid. Even if your organization is not yet considering the cloud, being cloud-ready guarantees that you won’t be prevented from moving to the cloud in the future for some or all of your data needs.
- Multiple application data importing: Often a report will rely on data drawn from multiple applications and/or systems. Look for BI software that allows this to be done simply, so you can get the most value from your data by the simple expedient of having all your data available to your BI solution.
- Development Platform: Your BI solution should be treated as a work in progress. BI systems have a reputation for lack of flexibility, but it doesn’t have to be that way. Architecting it as a development platform means that your BI solution can take advantage of evolving technologies and environments.
Your business intelligence represents some of your most valuable assets, so you want to be vigilant in protecting them. There are many aspects of security to consider.
- Application and system access and privileges: User access can be granted on a per-role or per-user basis, and institutional wisdom has taught us it’s best to provide the least amount of access necessary for a given user to do his or her job. This can mean different menu items available within applications based on user role, or access to different applications or servers.
- Data access by user: Access might also apply to what data or databases a user has access to. For instance, a publisher with multiple tracks might want to limit access by one imprint to the assets of another. This can be managed at the user, role, or domain level authentication.
- Single sign-on (SSO): Single sign-on is a session-based user authentication process that allows your users to manage only one password to gain access to all applications and systems. SSO also makes users more efficient, as signing in to multiple applications can add a surprising amount of time to a simple job. Simpler password management means less chance of human error in exposing passwords, but it is necessary when relying on SSO that passwords are strong, and rules enforced for frequent changing.
- Activity logging and auditing: A basic security process is logging network and application activity, and periodic review of these logs for unusual traffic. Besides the security benefits, the information returned in this way also helps developers measure and understand application and system usage, identifying bottlenecks and pain points.
The software should have a variety of reporting capabilities:
Traditionally, reports were generated within a large organization only by a few individuals, people who were savvy about both data and business analysis. This often meant that the reporting area became a bottleneck, creating unnecessary delays in receiving reports. Self-service reporting means that users and analysts don’t need technical skills – for instance, writing SQL queries – in order to get powerful, customizable reporting features. Good BI software will come with this kind of capability built in.
Dashboards offering a real-time view of business data are invaluable for a high-level summary, which can guide a user to more detailed reports. Users should be able to customize the dashboard to display the data most useful to them.
While it’s important to be able to build reports on the fly, out-of-the-box reporting is important as well, so users can quickly select a type of report to run, or to use as a base for customization. Typical reports that should be available include:
- Ranking reports: These can rank data objects, such as customers or products, across multiple dimensions and using multiple criteria. This allows you to see what is working best and worst within your organization, develop targeted marketing strategies, and build effective product roadmaps.
- Pivot table / OLAP: Pivot tables are used to extract, organize, and summarize data – specifically, they can present data in different ways, hence the “pivot”, which refers to the ability to pivot views. Online analytical processing, or OLAP, is a database technology used for querying and reporting, and is often used to enhance pivot table reporting. OLAP organizes its data hierarchically, and uses data cubes instead of tables. OLAP pivot tables provide an enormous amount of flexibility in reporting and viewing data, and enable an enormous amount of insight for spotting marketing trends, successful ad campaigns, association information, and much more.
- What-if reports: A What-if analysis does exactly what it sounds like it would do: it allows you to pose a hypothetical situation and see how changes might affect aspects of your business. You can use big data analytics to assess how customer behavior might change if you made a certain change to your product, its delivery or pricing models, and so on, providing a better foundation for decision-making and strategic thinking.
- Geospatial / mapping reports: Being able to analyze your business geographically to gain location-oriented insights can be invaluable. What products are most popular in which locations, or how markets vary demographically and socially, are tools for business intelligence that can seriously increase your organization’s competitive advantage.
- Ad-hoc reports: This is an important out-of-the-box functionality that you need: the ability for end users to create simple but custom reports as needed, on the fly, without setup and configuration concerns, query languages, or customization time. Users simply choose what elements they want in their report, and specify a delivery format.
In-memory reports and analytics
In-memory analysis and reporting enables business users to access reports extremely quickly, using data that is queried frequently. As the cost of RAM has plummeted, analytics software can store information that is queried frequently in RAM or page files, allowing rapid access. Combined with OLAP pivot tables, users can ask business questions, drill down through the data, filter it, and view it in a variety of ways – all at high speed, making big analytics available to business users without special skills, in a timely way that supports the pace of business. In-memory analytics can be run, even for complex queries, without damaging system performance for other users.
Speed of access using in-memory reporting has to be balanced with the need to use real-time data for the most accurate analytics possible. Some BI applications work with days-old data, or even week-old data, because the in-memory data or the data mart used for reporting is not updated frequently enough. Your BI software should be able to update reporting data frequently enough for your business intelligence needs.
Interactive reports provide a way to create and build on reports. A user can, for instance, generate a high-level report, and drill down through the layers in one or multiple areas, building on the report that they already generated. Data can be viewed across time and across locations, as well as across applications and systems. Conditional formatting can draw the analyst’s attention to important details. Powerful interactive reporting helps business analysts and data scientists synthesize conclusions from data, giving them business insights that are not available with less flexible reports.
Business intelligence isn’t only about marketing; it’s also about optimizing internal operations. Operational reporting gives the analyst a look at the day-to-day activities of your organization. These reports are often run automatically on a daily or weekly basis, usually at low-traffic times for the system. They can provide information on system usage, sales reports, expenditures, traffic analysis, and so on. Understanding a business’s operations is a key part of business intelligence that can be used to optimize the functioning of the organization. This is something to look for specifically, as much BI software doesn’t take into account the need to evaluate internal operations to get the most insight into one’s business.
Another form of reporting that can provide a great deal of business intelligence is intelligent alerts. These can be configured to report to a specific individual when a predefined event happens or a threshold is passed. A salesperson might configure an intelligent alert to send an SMS message to his or her phone whenever a customer cancels an order, or the Chief Security Officer might get an alert when off-hours traffic is abnormal.
White labeling, rebranding, or customizable skins are all terms that mean you can re-brand the BI software to look and feel like your existing business software. Often the customization is more than skin-deep, and allows true integration with your systems – but even if it is only for appearance, organizations today recognize the value of branding. Giving users a consistent interface improves user acceptance, making it easier for your internal users to feel “ownership” of the system, and builds on your organization’s software standards to provide a familiar interface, which helps users feel comfortable while diving deep into the system.
ETL stands for the basic database functions of Extraction, Transformation, and Loading. Clearly, these are functions you need when using data for BI analytics, but not all BI software has these capabilities out of the box – and yet, when working with big data analytics, you will extract data from different sources, transform it to a common format, and load it into a system or data repository where it will be available to end users and applications.
If you have a powerful data governance policy, you probably have metadata to apply to your data objects – that is, data that tells you something about the data. Metadata can report on a wide variety of data attributes, such as source, whether or not it has been validated, where it is used, what format it’s in, or pretty much anything your organization finds useful. Having metadata available can help non-technical users understand the data, as well as give analytics algorithms “hooks” to your data. However, once you’ve developed a metadata strategy, you must make sure your BI software supports metadata the way you use it.
Having a powerful BI tool will help you to get the most value from your data by providing the right toolset to discover, compare, analyze, and report on your data, with the flexibility to look at it from many different angles and in all dimensions. An open architecture to insure integration with your systems and allow for future growth, security to keep your business assets including business insights gained from analysis, and good reporting tools to enhance your analytics are the main things to look for when selecting a BI solution that will serve your business for a long time to come.