Requirements for BI
For an organization to become as smart and agile as possible, the process of transforming data into information and knowledge should take as little time as possible and should run with minimal interruptions and errors. An Intelligent Organization will therefore also have to take into account a number of generic requirements that we impose on both consuming information and the Business Intelligence tools. We present you here the top 5 requirements for Business Intelligence.
1. Reliability of data and calculations
The reliability of information largely depends on the quality of the underlying data and more specifically on the accuracy of the calculations that are stored in the Business Intelligence system. If those are accurate, we may assume that the information is correct too. The phenomenon of internal duplicates is a good example of such problems with data quality: details for one and the same customer, for instance, have been registered in the source system more than once with slightly different spelling.
When we start using Business Intelligence such duplicates will come to surface – it might take a while before the actual scope of the problem becomes visible – and can then be removed from the source system.
2. Preventing errors
We should also take into account the errors people tend to make while interpreting and internalising data and information. Business Intelligence tools can ease this pain in several ways (Isenberg, 1984; Van Beek, 1999) for instance through showing the same information from different angles or perspectives, or by designing the system in such a way that searching for ‘proof to the contrary’ is encouraged. Furthermore, Business Intelligence often provides an integral picture of business operations and indicators are interrelated.
By enabling the system to include those interconnected key performance indicators in a single report, users can immediately check whether these relations are correct, for example through combining market share figures with their own sales figures and with information on sales figures of competitors. Through imposing strict requirements on the quality, actuality, tempo, coherence and visualization of data, Business Intelligence contributes to preventing errors in other areas as well.
Above all this, people should be aware of the potential errors they can make and therefore should proactively revise, reassess and check information in order to create (more) reliable information and better knowledge.
3. Actuality and refresh rate of information
Stakeholders can be ‘on top of’ information if its refresh rate is optimal. Both Business Intelligence and learning processes work most efficiently when the time span between the action itself and measuring its performance as well as adjusting the action is as short as possible (Den Hertog and Huizinga, 2000). This principle is more effective if feedback about their performance is directly addressed to the employees – and Business Intelligence facilitates this – instead of via executives.
To avoid unpleasant surprises, the refresh rate and timing of reports must match the frequency and duration of activities within the business processes. An organization, which handles hundreds of orders per hour, does well to refresh sales reports on a daily basis (or even more frequent) and to display these daily or even hourly. However, there is no need for an account manager who visits four or five customers per week on average, to monitor the number of visits daily.
4. Fast response times of Business Intelligence tools
The Business Intelligence tools must ensure that information is quickly and timely available to processors within the organization. The system’s response time is of particular importance for managers since they often have little time and like action and liveliness. Three particular actions are associated with the requirement of rapidly retrievable information: finding the report, finding information, and interpreting information.
We leave out the phenomenon of interactive analysis (OLAP) whereby users can browse through data and casually surf the data ocean even though for these activities a rapid response time is often a necessity. A report that contains the required information should be able to show that information within 2 to 3 seconds after its been clicked on, but that specific BI reports first needs to be located. Once the report has been read, it has to be interpreted.
5. Storing information
In some cases, information is not immediately ‘actionable’ because an organization perhaps needs some additional information to be able to make a proper decision. If this is the case, the organization should be equipped with a memory in which they can temporarily store the information so that later, when the additional information is received, the decision can be taken rapidly.
In many organizations, the explicit storage and sharing of (newly formed) knowledge for clear decision-making is not common practice yet. What we do see is (too) many static dashboards and reports that offer hardly any (if any) options to attach a story to the information. This is a missed opportunity because the story behind the figures is often much more interesting than the figures themselves.
Business Intelligence can facilitate this aspect by temporarily storing information, which indicates a certain problem but still awaits additional information, in specifically designed place within the Business Intelligence system. For this to be effective it is wise to create an option on the dashboard which allows for placing notes and doodles with the relevant indicator and so create an action and decision log.