Hastings Direct goes fast-and-furious to the cloud with Snowflake

Insurance firm Hasting Direct has made an aggressive shift from legacy data platforms to Snowflake technology that has allowed it to deliver new experiences to employees and customers.

The Snowflake deployment began when Sasha Jory joined Hastings as CIO in September 2019. The company was introducing a new telematics-based product called YouDrive, which required detailed data from customers on their driving capabilities. As Jory’s team worked with another technology partner to develop YouDrive, they saw the cloud platform in action and liked what they saw:

We implemented Snowflake along with the YouDrive product, and we could instantly see the benefits of being on the platform. Since then, we've been on an aggressive journey to get off our legacy platforms and move everything into Snowflake.

The major element of the transition from Hadoop to the Snowflake Data Cloud took place in 2021. Hastings recently completed the final stage of the shift to Snowflake from Netezza. Today, the company doesn’t store any data on its legacy platforms and has moved 100% - raw data, cleansed data and curated data -to Snowflake. 

Transforming how the company works

Hastings deals with a huge amount of data. Being able to use Snowflake to access data, understand it, query it and find new opportunities has transformed business processes. Data analysts and scientists at the firm can now look at the firm’s data in any format, says Jory:

They can go right back to the raw data or they can use curated tables if they want to. It means we are able to enhance our quoting and our underwriting. We can do live pricing at the point of quote, and we've also been able to improve all of our pricing applications, so that we have some straight-through pricing that we can deliver. 

Having a tight grip on data means Hastings has also developed a much deeper understanding of its customers, including the products they use and the potential challenges that might come from moving into new service areas. Jory explains:

We understand our risk factors. We can understand fraud much better and hopefully give our good customers a great price and a great offering.

Snowflake has had a big impact on work processes, she adds. Rather than data scientists being frustrated with legacy ways of working, the company has a modern data platform that boosts productivity, which makes it easier to attract and retain talent:

We can now say to anybody in the marketplace who’s thinking of coming to work at Hastings that we have a very modern platform and one of the best opportunities for you to live and learn data.

Jory also points to the importance of being “consuming healthy”, which is about ensuring costs don’t balloon when you move to the cloud:

We never want to constrain our data scientists or analytical engineers from doing brilliant things. We want them to be creative. But we want them to do that in a way that they understand the cost associated to their work. We want them to consume healthy and have a sensible view on costs.

Looking for new opportunities 

When it comes to longer-term developments, Jory is interested in Snowflake’s Artificial Intelligence (AI) capabilities. At its annual user conference earlier this year, Snowflake announced new features for its unified data platform, including a Large Language Model. Hastings is also working closely with Microsoft on AI. The insurer is looking at Azure Open AI Service and Copilot and is part of Microsoft’s early adopters’ programme. Jory is keen to do more, too:

We think there's plenty of opportunity to bring Microsoft's and Snowflake’s capabilities together and we are doing some pilots together to try to bring that to life. Hastings looks at generative AI as something that will make a difference. But we're not sold that it is the be-all and end-all of everything.

As well as Snowflake, Hastings uses Databricks for some analytics processes. The firm uses Qlik Replicate to bring data from its core systems into Snowflake. Hastings also uses dbt and Azure Data Factory to assist with data management. The company is currently running a proof-of-concept study with the open-source framework Streamlit and is looking to make the most of Snowflake Marketplace, which helps bring third-party data directly into the platform. 

We are using Snowflake and all of its capabilities. We keep driving ourselves to be more innovative. And we challenge Snowflake to challenge us and say, ‘What more can we do? How can we do it better?’

That said, Jory cautions: 

There's a lot of challenges to overcome. There's a lot of stress that people have around the migration of legacy platforms and going into Snowflake. It costs money. I don't think anybody would say that it's simple. But the benefits when you get there are enormous. It feels like the shackles that you had previously around you have been thrown off. There's a lightness and a freedom you give to your organization.

Jory describes this freedom as “data alacrity”, which is about having the opportunity to create a more agile approach to data. Hastings took a rapid approach to transformation because it was keen to deliver benefits quickly. She says other CIOs should consider a similar approach, but they must be careful to ensure they keep costs in check during the journey:

We like to do things fast and furious. And so the idea of doing a lift and long-term transformation and then migration didn't really appeal to us. So, we just moved it lock, stock and barrel, knowing that – when the data gets to Snowflake – we're in a much simpler and hospitable environment where we can actually make changes. So, move fast and transform when you get there. But make sure that you don't spend all your money on the move, so you can't drive change when you arrive.