First things first. Giving a north to your data prediction project is crucial on the success of it. Addressing the next points will give you a clear vision of the project and most importantly, will help you make your predictions useful towards achieving your desire KPIs.
- Set your business objectives: What is your objective? What are your goals? You need to know what you're searching for, and what you want to get out of the data. What aspect of your product are you looking to understand or improve? Once you know what your objective is, go a step further and think about what data or events are necessary to help you achieve it.
- Use only data that matters: Only track the events that are necessary to achieve the objective that you defined, the predictions you want to make. This will make your priorities clear, and it will make it easier for the team to understand and use the data in a variety of ways within EPICA. It can be difficult for the team to understand what the data means and where data points are triggered, so starting with only the necessary tracking will make it easier to use. As goals change, you can add more events.
- Track your data properly: Before you start arranging your events, it is important that you understand how users are identified and tracked in EPICA, and that you are able to do so properly.
- Structure your prediction events: Create a document to keep your events organized. A spreadsheet is usually the best way to do this. List each event and its properties, and keep the names simple, clear and intuitive, so they are easy to remember.
- Consider adding offline data sources afterward: Integrating offline data sources can be a long process, so make sure if you necessarily need the data coming from the physical touch points in order to achieve your objective.