Financial services giant NatWest is using Snowflake’s technology platform to place data at the heart of the bank’s long-term business strategy.
David Charnley, Head of Strategy and Transformation, Data and Analytics at NatWest Group, explains that the firm is using information to improve customer services and experiences, to make better management decisions and to help boost sustainability. He says Snowflake plays a critical enabling role in this strategic work:
The use of Snowflake has been transformative in terms of delivering faster performance. We have 100 million-plus records that can be processed through the Snowflake engine in under four seconds. It also provides greater productivity – we have had a 60% increase in our data-modelling productivity through the use of Snowflake and other-cloud enabled tools.
Charnley gives one recent example of data-enabled work at the bank, where his team was able to push about 750 gigabytes-worth of data in three hours across the Snowflake platform instead of what would have taken three days historically. As well as providing a performance boost, he says Snowflake is easy to use:
We are deploying data-led solutions in days rather than weeks. In a world in which data collaboration is really important, we're finding it easier to collaborate across boundaries within the organisation and also bring in external data sources.
NatWest had been running another enterprise data warehouse for a number of years, which Charnley says provided an effective repository for a history of transactions, balances, holdings and other features. However, this legacy on-premises technology wasn’t scalable and struggled with higher compute, highly intensive queries. Charnley says Snowflake makes it easier for the bank’s database administrators and engineers to pick up source data, transform it and push insight out to the business:
What we found with Snowflake is that it's not only giving us faster performance, ease of use and speed to market, we also find our data engineers can come in and be productive straightaway. There isn't an extensive learning path that there might be with some legacy, on-premises technologies.
NatWest began implementing Snowflake with a “small” initiative in 2020, says Charnley. His team built some proof-of-concept projects through 2020 and into 2021. In the second year, they began building assets for some of NatWest’s key business units, including financial crime, retail and climate. From this “slow start”, the bank has expanded its ambitions:
What we're finding is that the more users get their hands on it, the more they want to do with it. One of our more recent developments is building a financial crime data hub that allows our analysts to get access to real-time, dynamic and easy-to-understand data, which influences how they monitor and manage our business to protect our customers and keep them safe.
These successes, says Charnley, means there’s going to be an exponential growth in the use of Snowflake at NatWest during the next 12 months. The bank is in the process of migrating all of its legacy data-warehousing and business-specific data into Snowflake:
We want to bring a lot more code to the data. So, instead of building long, extensive pipelines, we're relocating and moving the centre of gravity to our data lake and to Snowflake and making a much more active use of the platform’s procedures and other capabilities locally in the cloud.
Charnley believes that that shift will mean that all the bank’s enterprise data will be available to engineers and business users on the Snowflake platform. Even before that transition takes place, he says it’s possible to see some key benefits:
Our internal customers get really excited. Our commercial business, for example, is really excited by the single view of customer profitability that we created across our business and commercial books. The scale and complexity of that analytics work wouldn't have been possible before. So, delivering value at pace is probably the thing that our customers internally are most excited about.
Once NatWest holds all its data in the cloud, Charnley says he’s eager to explore some of Snowflake’s other features, especially those for commercialising and sharing data, particularly when it comes to building and co-creating data products with external organisations. He recognises the work NatWest is about to undertake will present fresh challenges because, until now, the roll out of Snowflake has been relatively organic:
It’s been about the growth of data sourced into the cloud for a particular use case. And so it's been a relatively well-constrained boundary around the work and around the complexity. The work that we're doing now, to migrate everything from the enterprise into Snowflake, is a much bigger task.
Charnley says this migration process will present some technical hurdles but he expects his team to overcome these obstacles:
Not only do we need to change the locus of the data, but we also need to rebuild our pipelines, rebuild our code, and rebuild the connections between source and Snowflake. And that's a massive undertaking. So, we expect to find challenges along the way. And we expect that to be harder; we expect to run into pressures around our pace of delivery. All of these things would be to be expected in a large transformation programme, but we're ready and up for the challenge.
In conclusion, he advises other technology and businesses leaders who are thinking about a shift to Snowflake to take a similar approach to his bank – start organically and build credibility with your business users and stakeholders:
If you can build some impactful use cases in Snowflake first, then you may get to the point where the business users are demanding more and they're asking for the investment rather than you. And that's a much easier place to be as a data or technology professional – to have your businesses making the case for you off the back of the things that you've already delivered for them. When you start to get some wins under your belt, it gives you the licence to go further and faster and to start transforming the enterprise. And then the benefits can be exponential.