Real-time and batch analytics

Data has a shelf life

Depending on the questions you have and the answers you are looking to get, you might be required to analyze data at different speeds. It is important to understand that data has a shelf-life. The older data becomes, the less useful it is in helping you with timely reactions. To derive accurate insights, consider the type of answers you want. Align your questions to the speed with which you gather and analyze data.

For example, consider if you made recent enhancements to your game like a new weapon or downloadable content (DLC), such as a new game level. You want to quickly assess if players react positively. On the other hand, metrics such as daily active users (DAU) or monthly active users (MAU) must draw from data that is compiled over a day or calendar month.

This is why there are primarily two types of analytics pipelines you want to focus on: batch and near real-time. Batch analytics is analyzing data collected over a period of time, where near real-time analytics is analyzing data as it comes in.