| |April 20189CIOReviewA BI makeover will be effective if the BI strategists keep the approach simple to begin with and ensure maximum buy-in from the stakeholders· Source Data & information dependencies (internal as well as external)· Appetite for Technology disruptors in the BI area (supply side from IT)· Maturity of the organization in using agile methodologies· Chalking out a budget for the programInternally, IT needs to look at below areas to infuse further inputs into the BI strategy:· Alternatives to end-to-end architecture supporting the BI strategy i.e. data warehouses / data marts, big data ecosystems, in-memory, self-serve, mobility, portal enablement· Cloud considerations i.e. analytics & reporting engine could be on cloud vs on premise · Metadata to drive BI i.e. security, user access, data taxonomies including metrics definitions, report inventory· Customization options using additional programming such as Python or SDK· BI tools capabilities required for the organization in the medium to longer termBelow, I have identified certain show stoppers typical in such journeys & tried highlighting here as lessons learnt. There could be others based on your respective organizational situations.Based on the size, business corporate structure, and complexity of data & reporting in organizations one or more BI tools can be suggested. Organizations are moving away from a lock in with single BI vendors especially given the flexibility available with licensing, scalability with cloud, reusable security models & metadata. However, a word of caution, if organizations do choose to go down multiple BI tools: Ensure these tools go through a single security/user access management module, reuse common data repositories and have tight governance on report & data change management processes (guard rails). It is never a terrible strategy to own two BI tools as long as the guard rails mentioned above are adhered to.Pull up market research from trusted sources to validate the options on the table e.g. Forrester waves, Gartner magic quadrants et al. Always map the requirements of the organization (current & future) to the capabilities you will look for in a BI tool. Never over engineer the solution as the cost of carrying redundant components could be high in longer term. Focus must be on empowering the end user in a federated model wherein self-serve capabilities are provided out of the box. Of course, this will be effective only when the users are given adequate training on the tools. Ensure training sessions are in house, where possible, with a boot strap project where instant gratification will be realized by the users.When Analytics is a requirement in the BI specifications, ensure there is a Data Science capability readily available with IT. The data scientists need to be subject matter experts with a clear understanding of the data environment in the organization. First few use cases for each business unit could be a joint development with the data scientists & the business unit where ample opportunities for data preparation & modeling discoveries are possible. Any advanced Analytics needs will need a separate program within the organization to stay successful.Finally, define metrics that will help the organization track success. A key metric I like to track is the percentage of data / information used for decisions delivered completely out of the BI eco-system. In conclusion, a BI makeover will be effective if the BI strategists keep the approach simple to begin with and ensure maximum buy-in from the stakeholders. Of course, speed to value will be another key metric to declare the program a success. Good luck with your BI makeover!
< Page 8 | Page 10 >