With all of the talk of AI and machine learning taking place today, you would be forgiven for thinking that every business is using these capabilities to gain an edge behind the scenes. While BI tools and advanced analytics are becoming more advanced and accessible for businesses, they are still a long way from being instantly valuable at the flick of a switch.
To become an organisation driven by BI and advanced analytics, we need to do more than hire talented people and purchase the right systems and tools. Before we dive head first into the exciting world of machine learning, we need to develop the right use cases for simple BI and start with the basics.
Start with current reporting requirements – Talk to department heads and managers to gain a clear picture of their current reporting requirements so you can identify where BI tools can provide immediate and demonstrable value. Having a clear idea of what stakeholders expect at a minimum is essential before you get started implementing a new tool.
Separate requirements into different sprints for each business area – Having one large deliverable alone is a recipe for disaster. Start with the basic reporting functions and work towards the most aspirational advanced analytics functions. Otherwise, when you encounter inevitable roadblocks and delays, the project has still delivered some basic milestones. Getting feedback on basic functions will also help you avoid simple mistakes when you attempt to roll out more advanced capabilities.
Prioritise sprints across the entire business – This will involve an understanding of the politics of an organisation in terms of who is sponsoring the project, and who stands to gain the most from these new tools. Some questions to ask include:
- Which areas of focus will have the greatest impact overall?
- Which areas are likely to be the most challenging?
- Which business units and roles are more likely to adopt the tools sooner?
- Which business has the most clearly defined metrics to plug into tools?
Start with basic data access and visualisations – Remember that many users might have never had simple access to data before, so don’t underestimate the power of basic BI for getting stakeholder support. Simple visualisations that the entire business can understand will also build an early culture of acceptance that will drive greater adoption as more advanced functions are rolled out.
Validate, validate, validate – A phased approach to implementing BI and advanced analytics requires constant data validation. Your user experience and visualisations can be breathtaking, but if end users lose faith in the information being presented, the project has hit its greatest avoidable hurdle. Share the data with as many stakeholders as possible early on in a “beta phase” and encourage users to spot errors so they can witness the system being refined first hand through a focus on data validation.
Roll out the tools to one department at a time – This is a staple tactic of change management, as forcing an entire company to change the way they operate overnight is a fool’s errand. By focusing your efforts on one department, you can maximise your training and support functions so that every user has the backup resources required for their adoption.
Remember that BI and advanced analytics implementations aren’t a set and forget project. As your business and your marketplace changes, these tools will require further changes and fine tuning to suit your needs. As you achieve success with lower levels of advanced analytics, you can use these wins to power adoption of machine learning and AI.