Zego’s Data Journey: Collaborative data culture

Welcome to the second post of our series on Zego’s data culture and how we’re working to build one of the most data driven insurance businesses in Europe. This post is all about our team and how a collaborative culture is critical to a performing data function. If you want to know more about the Tech we are using, have a look at our previous post in our Data Journey series.

A common question when chatting with data professionals is, “which department are you part of?” The most common answer is tech, but operations, marketing or actuarial science (in the case of insurance) are also common responses. The reason is simply that data plays an integral part in so many functions and depending on internal organisation and individual circumstances, it can actually fit in almost any department.

In any case (at Zego we are happily a part of the Tech team) we believe that an inclusive data culture can only be achieved with a data function that interfaces with tech and most other functions in the business. The most common interface is “team A consumes data” but that’s just part of the picture.

From building the infrastructure to agreeing workflows and overseeing data usage, a data team can only succeed with a truly collaborative approach. That approach also shows that data is a company resource, avoiding the data silos that are so common in organisations of every size.

Backend engineering

Our backend engineers develop the system that supports our apps, web site and admin web interface. This is our core database where every shift, policy or transaction gets recorded. Zego launched more than 7 products in the last 12 months and often insurance products go from specs to launch in less than a week, meaning weekly or daily data structure changes.

Data engineering has to keep our core ETLs working, despite having a moving target as data source and ideally causing no overhead to the backend development process. A flexible tech solution and ability to detect changes quickly is critical, as are good communication channels and a collaborative approach. We also ensure that we are receptive to the recommendations for good engineering practises and code reviews we receive from our backend colleagues.

Tech Ops

A common problem in data teams is the lack of experience in managing infrastructure. The consequences can range from bloated servers to questionable security practises and a fair amount of grief between teams. We thought that Data should not be managing its own infrastructure but instead work together with TechOps when setting up, scaling and upgrading. By following best practises and standards, as defined by our TechOps team, Data Engineering is able to have fully managed and supported ETL servers. Backups, disaster recovery and credential security are there to make our work easier and safer.

Pricing

Pricing is one of the main stakeholders for data in any insurance business and Zego is no exception. Zego’s Pricing team needs to model data sets for innovative products where industry benchmark data is not available. We work with our actuarial and data science colleagues to make sure that they have access to the data they need, whether it’s internal data, third-party datasets like traffic and weather, anonymised log data or open source data that can help better understand risk.

Growth

Growth managers need to understand how our partnerships are performing on a daily (or hourly) basis. Their demands include plenty of ad-hoc analytics and polished visualisations that can be showed externally. Our Growth team are always pushing for a higher frequency of updates, telling us that overnight batch processing is no longer enough.

Finance

Our Finance team requires data for different types of financial reporting, reconciliations, feeds into the accountancy system and ad-hoc analytics. By collecting relevant data and automating data feeds, Data Engineering is able to save time and help reducing the risk of human error on a critical corporate function.

Customer Service

Our colleagues in Customer Service are providing the best possible support to our end customers. With new products and markets being launched regularly, it is vital that we can introduce as much automation as possible. Among other projects, Data Engineering created a data bot that sends an automated message to each advisor individually letting them know their relevant KPIs for the day before based on admin, ticketing and call centre systems. Data from ticketing and call centre systems is also available to management to access peaks in demand and other performance metrics.

Compliance

Compliance is often portrayed as a barrier to organisations using their data in the best way. At Zego we understand the importance of protecting our customers’ data while still supporting business growth. From the beginning, Data Engineering engaged with Compliance to identify potential issues in projects at an early stage. Having a tech-minded Head of Compliance is definitely a big plus!

We have just been through our first data usage survey with responses from every team in the company. Among the common themes are ease of use and update frequency. We will be reviewing our BI capability, building new dimensional data models and moving key datasets from batch to streaming. Watch out for new posts about how we are getting data to every Zegon!