Main content

River Island uses Snowflake technology to fit out the business with fashionable insight

Mark Samuels Profile picture for user Mark Samuels May 16, 2023
The fashion retailer is adopting a range of Snowflake technologies to ensure its employees have integrated access to potentially game-changing information.

River Island

River Island is using the Snowflake Retail Data Cloud to bring its information together and to provide a platform for innovation on behalf of its customers. 

The Retail Data Cloud allows retail companies to tie information together from across the business and to centralize a variety of data sources into a single repository, saving time and money, and supporting enterprise-wide efforts to turn information into insight. 

According to Adam Warne, CIO at River Island: 

The Data Cloud is there as the main repository. It just makes it easy to get data in and out. And it makes it easy for us to build models on the back of the platform. It just makes what we're trying to do with data far simpler.

Warne says the platform is helping the retailer to develop a data-inquisitive culture, where people can access information and use it in a safe environment:

As a business, we've got operational reports that tell us how we’re performing. Snowflake is giving lots of people across the business the ability to go and have a look at the data, to ask some difficult questions, and to slice and dice the data. We're getting lots of insights from that inquisitiveness and we want to push harder.

He adds that one of the biggest things the platform has provided is a one-stop shop for verifiable data: 

We spent a lot of time discussing and debating the truth. Snowflake gives us a single, reliable version of the truth. That means that we now spend more time thinking about the action that we're going to drive off the back of data than anything else. The other thing it’s done for us is to increase the speed of insight.

Getting ready

River Island started working with Snowflake in 2018. When Warne joined the company in October 2021, it was already established as a foundational technology. The company continues to hone its approach by exploring other elements of the Snowflake platform.

The retailer has been trialling Snowpark, which is a developer framework that brings native SQL, Python, Java, and Scala support to Snowflake. Rather than having to take data from Snowflake and load it into a different system, Warne’s team can exploit the information where it lies:

With Snowpark, what we're doing is moving the code to the Snowflake platform. And that's actually saving time, and it means we only have the data in one place, so it saves us money as well.

River Island’s IT team is also using the Snowflake Marketplace, which is a collaboration space that allows companies to exploit a variety of third-party datasets. Warne explains: 

We’re thinking about other data, whether that's weather data or footfall information, that can we bring into our systems to inform the models. It gives us clear answers about what we should be doing and why performance might be better in certain areas than others.

River Island is ingesting these external data sources to build predictive models that are powered by machine learning. This information is being used to help manage stock in response to new customer demands. The data helps the retailer to track and trace products across its supply chain and to ensure sustainability targets, such as reducing carbon emissions, are met.

Warne has found Snowflake innovations simple to implement and use:

I think that the hardest bit was moving data from our existing platform into Snowflake. It wasn't too difficult, but we had to be careful because we have to still run the business. From start to finish, the movement of data probably took us about six months. River Island is all about constant change. And so when something new comes in, there's a process of handholding and managing that change through, but Snowflake has been very well received. It's just part of our day-to-day working life now.

Going forward

Data-led innovation continues to be at the heart of River Island’s activities. The retailer is using radio frequency identification (RFID) tagging, where each product has a unique reference number that can be recognized in smart changing rooms. Once scanned, customers gain access to the same product information they’d get online, and River Island also gleans performance insights, such as whether a product was tried on but not purchased.

Warne says he’ll be looking to Snowflake to help power innovation in other areas, too:

When I partner with a supplier, I always set myself the lofty goal of being a headache for them. I want to be the executive that's pushing them and saying, ‘Come on, I want this feature.’ What I'm seeing with Snowflake is that's actually going to be really difficult, because they're always delivering new products. They're always pushing the envelope.

Warne describes River Island, which is privately owned, as an entrepreneurial business. He says the long-term aim is to use Snowflake and the firm’s data-inquisitive culture to identify new opportunities for insight that drive operational and customer-service improvements:

I think where this ends up is we become far more predictive in our analytics as a business and we use data to drive more decisions. That’s not to say that we undervalue people's opinions, but I think there's certain things that we want the analytics to drive. And we want to create more time for people to show their creativity. So, for the more menial tasks, we want more intelligent automation. We want to be using AI more and we want to save people from doing monotonous tasks.

Warne says it’s important to make sure the platform – and the services you select – align with your overall business strategy:

For River Island, that’s very much about MACH principles – micro-services, APIs, cloud-native, and headless. And for us, Snowflake fits really well into that space. It doesn't have any problems with scale. It connects very well to lots of other systems, so you don't have lots of complexity in terms of implementation.

A grey colored placeholder image