5 Best Practices for a Successful BI Project
Completing a Business Intelligence and analytics project successfully could seem simple, but it is not, even now many solutions are efficient and easy-to-use. The success of this type of project relies not only on its technical aspect, but also on its human aspect. A recent study by Dresner Advisory Services called “2015 Advanced and Predictive Analytics Market Study” shows that only 35% of respondents consider their analytical project to be a real success.
1. Do not focus solely on key performance indicators.
These indicators (or KPIs) have become a tool for monitoring the performance of a company and its teams. It is essential to have reliable and indisputable figures to evaluate the activity of an organization and each of its departments and employees. But be careful, an indicator is not enough. Beyond this, it is necessary for all stakeholders to know from which data and tools this indicator has been calculated in order to understand its evolution and to make the right decisions to act.
An application of BI must therefore provide an accurate data analysis used for each indicator, which can be used in the context of business performance management, essential to monitor the operational implementation (planning, execution) of the organizational strategy at all levels.
2. Beware the myth of the self-service BI for everyone
In recent years, we have seen the growth of self-service Business Intelligence tools. Information is shown in real-time for users who want to make an ad hoc based data analysis. But these solutions might not meet the different expectations of all employees in a company. For example, less technical users prefer high performance BI applications, created specifically to meet their needs while management usually adopt the use of dashboards where data is synthesized with the ability to receive alerts when certain thresholds relating to key indicators are exceeded.
3. Prepare your data carefully
Key performance indicators distorted and misinterpretation, or a low level of data quality, or even the tools rejected by users, is often the cause of the failure of many decision-making projects. Analysts and Data Scientists may spend couple hours a day correcting errors they identify. But those they do not see are potentially a source of error for decision makers. The data quality problem can be resolved in the long term only by implementing a genuine data quality management strategy upstream of the Business Intelligence platform.
4. Think BI platform rather than tactical decision solutions
For a Business Intelligence solution, the use of data discovery tools, independently of other platforms, and allowing users to generate their own analysis or create their own graphical representations, will be inadequate because it will be too complex for a business user, and only analysts or advanced users will have the knowledge and skills to take advantage of these tools, leaving a large part of the rest of the population – directors, managers, operational staff, customers and partners – unresolved. The return on investment of this type of solution and its adoption rate will be low.
Developing a comprehensive BI strategy across the enterprise is essential. An ideal solution should set up an analytical platform offering a range of tools that can meet different needs of business users and ensure consistency in together.
5. Select the appropriate data sources
When conducting a Business Intelligence project, paying attention on information stored in the ERP (Enterprise Resource Planning) and CRM (Customer relationship management) systems or in multiple and varied databases is necessary but far from sufficient. A lot of enterprise data nowadays is unstructured data. For example, comments from social media networking, data generated by machines or by products from Internet of Things, all these data sources must be supplied into the analytical platform of the company. Only analytics based on both structured and unstructured information will help management make the right decisions.
