Cover Uses Complex Data Analysis to Empower Customers

San-Francisco based Cover is a straightforward, app-based way to get insurance on anything from cars to personal property. Customers use Cover’s mobile application to simply take pictures of an item they want to insure, answer a few short questions and the Cover team then finds them the best insurance policy from 30+ carriers to match their situation, completing the entire process via text message. 

When Chris Stanley arrived to lead business operations in 2018, Cover was collecting an incredible amount of raw data on its interactions with customers, but that data wasn’t being piped into any analytics solution. Instead, engineers were regularly writing SQL queries against the raw database to get a sense of how the company was operating.

“The first thing we needed to do was effectively get all these different data sources and pipe them into one place where it could easily be analyzed by writing SQL queries,” said Stanley. “Sisense for Cloud Data Teams provided us the ability to link our data sources to an analytics tool in a very easy way.”

“We can create dashboards, create email reports, everything we need, so Sisense for Cloud Data Teams really has the full scope of what we needed to effectively drive business insights from our raw data tables. Sisense for Cloud Data Teams has also scaled a lot with our needs as we’ve done some more complex analysis. It’s easy to get started, great for scaling, and gives you all the functionality you need right at your fingertips.”

Chris Stanley
Chris Stanley, Business Operations at Cover

Stanley has worked with several different tools in his analytics workflow throughout his career. That included one workflow with manual, custom-made ETL tools supported by a competing BI solution platform for visualization, and another workflow using another competitor. At Cover, he realized that recreating those complicated workflows wouldn’t fit in his new position.

“At a startup, we don’t have the luxury of going to our devops team with specific requirements for analytics and have them go build it,” said Stanley. “All our engineers are focused on trying to build and enhance the actual consumer-facing product, so our priority was to find a solution that would help us get up and running without requiring any bandwidth from our engineering team.”

As he began evaluating tools, he determined that building a manual ETL tool would have been too big of an effort, and one competing solution would have been too expensive to provide value. He also determined that the other competitor’s implementation would have meant spending his time in the initial weeks and months of his new job trying to teach his team a new language, with added layers of complexity involved to maintain that over time.  

With Sisense for Cloud Data Teams, Cover has been able to get up and running quickly. Stanley also says the SQL Views capability was a big win for Cover, because it meant they didn’t have to use their own ETL tool. Instead, they could use SQL Views as a way to write all the glue between their tables. The simplicity of the “write SQL, get charts” mantra played perfectly into Cover’s needs, because it didn’t require multiple steps to get SQL queries ready to be visualized. Finally, the flexible reporting functionality in Sisense for Cloud Data Teams allowed Cover to maximize the value of their analysis by empowering the rest of the company with data insights.  

“We can create dashboards, create email reports, everything we need, so Sisense for Cloud Data Teams really has the full scope of what we needed to effectively drive business insights from our raw data tables,” said Stanley. “Sisense for Cloud Data Teams has also scaled a lot with our needs as we’ve done some more complex analysis. It’s easy to get started, great for scaling, and gives you all the functionality you need right at your fingertips.”

Stanley helps tackle analytics questions around every step of the customer flow for Cover. The company tracks how it acquires customers through paid marketing, organic traffic, and social media, then collects data related to how well customers are navigating within its app. And with text messaging as its main communication medium, Cover then analyzes text from those conversations to make further predictions on conversion of customers.

On the product side, Cover does all its AB testing analytics within Sisense for Cloud Data Teams. They have event-based data piped in and run tests in the platform to inform the business on whether they should fully roll out a new product feature or not. 

“We’re always keeping a pulse on how engaged our customers are with our insurance agents. At the end of the day, everything is a conversion funnel, so we’re always just trying to optimize the conversion points through the Cover experience,” said Stanley. “We’re constantly pushing new features and always using Sisense for Cloud Data Teams to measure the impact that our changes are having on the conversion funnel.” 

From there, Stanley and the Cover growth team use Sisense for Cloud Data Teams to email reports out across the company and externally on a daily basis, highlighting metrics on new customers and the amount of policies sold. Cover’s executive team is among those who often create some of their own dashboards in Sisense for Cloud Data Teams, and in the future Stanley hopes to create workspaces for those non-technical users to dig deeper and analyze data on their own.   

Stanley also owns the financial and operating models that are crucially important for the company to run the business every day and communicate to investors. By piping financial data into Sisense for Cloud Data Teams, he can instantly see how many customers came in, how many people were on the app, track conversion rates, analyze premium bookings, monitor ad spending, and determine gross margin figures. 

“Working with Sisense for Cloud Data Teams on our financial models is a really big time saver,” said Stanley. “I don’t have to go collect data from everywhere, I just have literally one canned query that I run on the first of the month that basically takes care of all our financial reporting right out of the box.”

To make all this data aggregation and analysis possible, Cover’s production databases are copied into an analytics database that Sisense for Cloud Data Teams points directly to – all Cover’s raw data tables are cached and glued together through SQL Views.

“Sisense for Cloud Data Teams effectively plugs into all these data sources, and we’ve built ETLs and SQL views and everything to see all this data together and provide analytics for the entire life cycle,” said Stanley. “And because we have data sources all over, we sometimes have to upload CSVs directly and the CSV upload feature is great too.”

Empowering complex data analysis

As Sisense for Cloud Data Teams opens up more basic analysis tasks to be done by others, Stanley has advanced to work on more complex projects. He’s currently working on scaling a machine learning model to predict the likelihood of customer conversion. 

“We have a lot of people on our app every day, we know there’s a lift on our sales conversion if we’re able to prioritize the people who are most likely to become customers,” said Stanley. “When someone comes on the app, we run them through our machine learning model via Python in Sisense for Cloud Data Teams and put a prediction score on them, and they go to the top of the queue for our salespeople if they are likely to convert.

With the integration of Python and R capabilities directly, Stanley can easily insert Scikit-learn to run a Random Forest model on Cover’s customers, create predictions, and then use that information downstream to optimize sales efforts. 

“The integration of Python and R in Sisense for Cloud Data Teams is a huge win for us,” said Stanley. “We don’t need another script running outside and talking to multiple different databases in order to do our machine learning work, we can just do it right in the system.” 

Stanley says he’s also working on using Sisense for Cloud Data Teams to handle more advanced dashboarding and complex data transformations with SQL Views, and is excited about new features that will allow him to better manage data recency and control when views run.  

“With regards to business reporting and analytics, I don’t really see another tool that’s going to be better for us than Sisense for Cloud Data Teams,” said Stanley. “It’ll scale very well with the organization, we will just have to add a few extra seats over time. The fact that everything is through Sisense for Cloud Data Teams’s Amazon Redshift database means that we don’t actually have to manage database scaling, so that’ll be pretty great for us moving forward.”