Old business intelligence looked to the past to understand how the business performed and to analyze what will improve future business performance. For that historical approach to analysis, old BI relied on data stored in formatted business records usually found in data warehouses or marts. The data assembly requirements included bringing data into the system from various groups and departments within the company, as well as from vendors and channel partners outside the company. However, because this is always done after the fact, there is no need to synchronize the timing of old BI analysis with ongoing operating or financial processes at the company.
Data delivery to old BI users typically came in the form of dashboards, printed reports or data transferred into other applications, such as Microsoft PowerPoint for presentation or Excel for further analysis. Whatever form they took, the timing of the deliveries was not critical to the success of old BI.
The user group for old business intelligence was typically small and high-level, consisting largely of two groups: business analysts and financial professionals, and senior managers and executives. The decisions that resulted from old BI analysis had a relatively long-term.
focus that affected the company’s business plans, and so they were not generally time-critical with respect to day-to-day operations.
Adapting to New Business Needs for Information Access
New business intelligence, in contrast, focuses on those day-to-day operations of the company, and it is used by a wide array of line management personnel who are responsible for making decisions that drive the current business performance of their units. That difference has several implications for how much and what type of data a business intelligence application needs, what types of decisions its users make, how the business applies those decisions and the number and type of users that a BI system must serve. It also has implications for how a business intelligence system presents and analyzes data.
New BI assembles data from the business as it happens, reports on it and analyzes it. While this set of processes uses some historical data, it relies much more on current transactional information. It is not unusual for BI systems to require real-time access to data and to update its data stores several times during a day. And their mission to make operational decisions requires more types of data and more of it than the longer-term analysis of old BI.
More people use new than old business intelligence systems, and most of those people are not at that senior managerial or analytical level. As a result, new business intelligence systems have to scale to accommodate much greater user demand, which may require additional hardware and networking resources, as well as different types of software able to sustain high usage volumes. New BI systems also perform reporting and analysis differently than traditional systems.
Hurdles for New Business Intelligence Systems
Decision-making in both old and new business intelligence environments is collaborative. However, the decision latency that prevails in old BI environments was unacceptable in environments where decisions may affect business operations on the day they are made. That means that the decision-making process supported by new business intelligence has to keep moving at a fast, steady pace. To do that, new BI systems have to assemble a huge volume of data, analyze it and present it in accessible ways to many users.
Not only do the types of data presentation and analysis used in business intelligence have to be geared toward large numbers of people; few of those users have the advanced technical and analytical skills of those using old BI systems. And even if the skill levels are comparable, those users do not have the luxury of time because they need to make operational decisions immediately. They need quickly to access, absorb and act on information and analytical results.
The lack of timely access to information – latency – is one of the most difficult challenges to using business intelligence systems successfully. In one business intelligence study, for example, nearly three-quarters of respondents said that reducing the time it takes to update their data was somewhat important or very important. In addition, nearly one-quarter of the respondents said that adequate business intelligence analysis requires real-time data, and more than one-third said they require daily or more frequent updates.
Collaborative decision-making in a BI environment depends on shared data, but data access often is confined to those who use it regularly. It is critical to share information to be able to improve the quality of decisions. But many organizations do not enable collaboration or understand it to be important; this business intelligence study found only a small percentage of users who rated collaboration as a high priority.
Compounding the collaboration problem are data silos, which effectively hide data from many users. Departmental transaction data that users require for business intelligence frequently resides inaccessibly in other functional or departmental areas of the enterprise. Even when it is available, automated transfer mechanisms are not in place, so users must extract it manually, which can be difficult and time-consuming to do. It can be even more difficult to acquire data for the business intelligence platform from external sources, such as trading partners and industry associations. And some of the data may be unstructured and therefore inaccessible to older reporting tools, further compounding the problem.
Integrating data from all the disparate sources into the business intelligence environment requires intra-company and inter-partner collaboration. To enlist the cooperation of these separate units requires executive support and diligent bridge-building. What’s more, an operational extraction, transformation and loading (ETL) environment of some sort has to be present at the front of the system to automate transport of the data. This extra step can add overhead to the process, of course; as well, it sometimes can create disparity in versions of the data. An effective BI system can provide dynamic access to data across data silos, making it possible to provide real-time information to operational workers.