Detect anomalies with root causes. Across all subsets of your data.
BoostKPI does the heavy lifting, so you and your team can be rest assured that any business opportunity in your KPIs don’t go unnoticed.
Businesses today collect their KPIs across multiple dimensions. A worldwide eCommerce business, for example, may measure daily revenue (KPI) across countries, states, channels (iOS/Android/Web), device hardwares, and payment processors. To guide decisions, they want to detect KPI anomalies in any subset of data – changes that are both statistically significant and meaningful. Detecting such anomalies (and the accompanying root-causes) is currently expensive and highly manual, relying on skilled data analysts slicing & dicing through millions of possible combinations, arising from the cross product of unique values of each dimension. With a limited supply of such analysts, businesses often miss detecting anomalies or detect them too-late.
While talking to many businesses, we heard of numerous such instances where anomalies went undetected or were detected too late. The potential dollar impact of each instance ranged from 10s of thousands of dollars to hundreds of millions of dollars. In one example, a global e-commerce business did not know for 2 months that the revenue in a particular country had jumped 3-fold. The jump was almost 5% of the overall revenue but no one knew because they typically tracked just the global revenue and the revenue of the top-5 countries. It was an “unknown unknown” that had slipped by. In another example, the payment processor in a particular country was down but no one knew.
BoostKPI leverages machine learning to solve this problem. It makes the data analysts and the operations team more productive by proactively alerting them when there are any anomalies in the data. If you would love to learn more or see the tool in action, please book a demo.