Cover case study

How Cover uses BoostKPI to quickly detect anomalies and opportunities in their subscription funnel

Background

Cover is what getting insurance in the 21st century should be. Users can choose from over 30 different providers and have bespoke policies produced for them in minutes.

Use case

With over 400,000 unique combinations of events happening across the board, Cover uses BoostKPI to:

track its subscription funnel

understand a user's journey on their platform

detect performance anomalies in their ad-campaigns on different channels

Solutions

BoostKPI also alerts the data team when there is a statistically significant and meaningful change in any KPIs. All without writing a single line of code!

In one of such events, while monitoring their user conversions, the data team at Cover was alerted of an abnormally low conversion rate. With all hands on deck, it was a race against time to slice the 400,000 of unique combinations in data to identify the reason behind the drop in KPI and recover from the event. The team under Morgan Bugbee ( VP, R&D and Analytics @ Cover) was able to identify the root cause driving the drop in conversion, a task that could have otherwise taken many hours and analysts.

The team also receives timely machine learning guided alerts, so that they do not have to spend time monitoring the dashboard.

An example alert message is included below.

The first alert is on Cover's “installs” KPI. A single alert monitors for changes in daily install per device_bucket, with an added business rule of only alerting if the change exceeds 25%.

The second alert is on Cover's “quote_completed_percent” KPI, broken down by media source, depending on the media source. Only “important” media sources, where they contribute heavily to Cover's KPIs, are highlighted.

Morgan Bugbee, VP, R&D and Analytics at Cover says:

“Building customized queries to examine the data for root causes is time-consuming and can stifle the data team's ability to solve bigger and better challenges. The ML-based tool from BoostKPI is like having an insurance policy on our data: we can rest easy knowing that if a KPI fluctuates, the tool will immediately notify us with the true reasons. We are no longer continually seeking for the "what, why, and where" behind our business KPIs in the war room."