River Island adopts a data mesh approach, enabling 14 departments to become self-sufficient

About

River Island is a pioneering and leading UK high street retailer, operating in over 125 markets worldwide, with over 250 stores across the UK. It has a significant online presence, while remaining a privately owned company by the Lewis family

Rivery is a no-brainer. I can just set it and forget it. We’ve probably saved around £100K on one project alone, using Rivery’s out-of-the-box Google Ad connector.

 

Ivan Isidore, Data Director, River Island

Challenge

River Island is a huge company with a small data team, and a ton of data disseminated in spreadsheets and S3 buckets company-wide.

Their data sources include apps for media ads and digital management like Google Ads, Google Analytics, Twitter Ads and Survey Monkey, data lakes like S3 and Google Cloud Storage and retail-specific sources like Rebound which manages their customer returns data. Their main storage layer is Snowflake.

River Island was struggling to ELT data from countless sources, orchestrate it, and stitch together siloed data from 14 departments and approximately 30 different teams across the organization.

Data sharing was an enormous burden and deploying new pipelines would take weeks or even months. Production was seriously affected and River Island needed a better solution to unify and streamline their data, and to accelerate PoS and data delivery.

Solution

Operating on a limited budget and with lots of non-technical users, River Island turned to Rivery to achieve seamless data integration and advance towards a data mesh implementation. This strategic investment helped them revamp their entire data architecture, empowering non-technical users to effortlessly own and manage their data and workflows from start to finish.

Custom connections to unique applications

To create a complete 360-degree customer view, River Island needs to ingest data from unique retail applications for orders and return management. To avoid coding and building their own infrastructure to access these unique data sources, River Island uses Rivery’s REST API to configure custom connections for apps like Rebound. All the configuration is done via Rivery’s Action Rivers in just a few clicks. Rivery full manages the infrastructure and monitoring of these pipelines, like other apps River Island is connected to.

There is no point in trying another solution. Rivery delivers exactly what I want it to. I don’t have to build a connector from scratch or maintain it 24/7. I have it out of the box or I can leverage Rivery’s REST API anytime to build a custom connector in a few clicks.

Implementing data mesh

With a tight budget and a lot of non-technical users who need to access siloed data, River Island turned to Rivery to reconcile the data, eliminate silos and build a complete “data platform” from the ground up. They reorganized their data architecture using a data mesh approach, whereby each department team owns its own data set and is responsible for its own “data product.”

Rivery is an enabler. It’s allowed us to adopt a data mesh approach, transforming how we share data, own it and make use of it. Today, we have 14 departments that own their own data product and are self-sufficient.

Multiple environments

River Island leveraged Rivery’s environments to roll out ownerships company-wide, so each team is now able to ingest data from various sources, transform it and orchestrate how it’s loaded to its target Snowflake database.

The “data products” are created by different department teams and are then connected by a centralized team. River Island has managed to create over 10 different environments in Rivery serving different teams and enabling a total CI/CD approach to publish tested data to Snowflake.

Rivery has been the best investment for us to onboard different data consumers across the organization. Our main users work in data analysis and science, but they are non-technical users.

They use Rivery to create data models, move data, submit tasks to reload data and modelize it so it’s accurate. Now every user is self-sufficient and can load data to their target Snowflake database

Rivery has delivered outstanding results for River Island. From data chaos to data mesh, data sharing company-wide is now easy and scalable. River Island is continuing to innovate with this data mesh approach and implementing new and complimentary BI tools to optimize data governance. Rivery has been the #1 facilitator for that, speeding up data delivery and providing the value that’s required now and in the future.