Landkreditt Accelerates Process Performance, Data Loading and Development with Snowflake

Founded and still owned by a cooperative of Norwegian farmers, Landkreditt manages kr36 billion in assets for several thousand customers across Norway. Since 2002, the bank has been on a transformative journey, turning it from a modest cooperative into a large financial group — offering wide-ranging private and corporate products and services. 

Over its two-decade-long transformation journey, Landkreditt has scaled its products, services and customer base substantially. With this growth has come mountains of data, including sensitive and regulated information crucial to its operations, serving as a backbone for the organization. Adding to the stakes of having a first-class data platform are the group's environmental, social and governance (ESG) commitments, which place a significant emphasis on its regular reporting. 

“When we originally scaled our operations, we found that our existing data environment struggled to perform, so we invested in our own on-premises solution,” says Bjørn Anders Gislerud, Project Manager at Landkreditt. “After we pivoted into insurance, we started collecting and processing lots more data and had ever increasing storage and performance demands.”

This increase in data volumes meant that Landkreditt’s data infrastructure struggled to keep up, negatively impacting efficiency and time to insight. In addition, data storage and processing constraints meant that reporting had to be performed in time-consuming batches.

The group also found that its developers were spending a lot of time on optimization, maintenance and analysis, and that its outdated infrastructure risked talent loss from key roles. 

Landkreditt needed a data solution that could satisfy its demand for newer, more capable technology without ramping costs.

Drawing from experience with a comparable data environment transformation project at another group, Landkreditt's Chief Technology Officer Gudbrand Olimb proposed Landkreditt perform a business intelligence (BI) proof of concept (POC). He pitted Snowflake against the existing infrastructure, as well as other vendors’ solutions. 

“I’ve seen Snowflake prove itself before — the company’s unique, feature-rich and database-centric approach stood out to me,” says Olimb. “As a regulated group, we had to do a lot of compliance work prior to the POC, which Snowflake provided a lot of support for. Our first POC for insurance risk calculations and premium processes increased overall performance speeds by 30%, and in some calculations, 30X. So, we rolled it out to a group-wide evaluation.” 

In just over 12 months, Landkreditt migrated its entire on-premises data footprint to its new DAIRY platform — Data platform for Analysis, Insights, Reporting and Ymse (Norwegian for all that is unknown, undefined or future requirements). 

The group now utilizes Snowflake with Matillion ETL, Zeenea Data Catalog and Microsoft Power BI, supported by modern development practices using Azure DevOps and CI/CD pipelines, to deliver its group-wide data platform. This provides governance, end-to-end data lineage, data quality and discovery and common taxonomy. It also helps the group process a substantial number of changes to its data each day, with individual query time improving to one minute down from 30 minutes previously. 

Landkreditt also benefits from Snowflake’s collaboration capabilities, allowing it to transfer detailed data sets directly into its analytics platforms and share them across the entire group. “Snowflake’s data sharing is fantastic,” says Olimb. “It’s helped us eliminate manual processes and gain up-to-date data at scale. Processes that previously took months now take just minutes. We see more opportunities in this space — including using Snowflake Marketplace — to grow and evolve.”

Landkreditt uses Matillion ETL together with Snowflake to load its core enterprise data directly into its platform and analyze it on the fly. Combined with Snowpipe (Snowflake’s unique data loading tool), Tasks and Streams (used for ingesting mutual funds data seamlessly), Landkreditt has improved time-to-insight for its CRM and business intelligence platform. 

For governance and compliance, Landkreditt uses several techniques to ensure data integrity and protection. In particular, it separates role access to dictate database, schema and object permissions, and uses row-level security for comprehensive row access and Dynamic Data Masking to implement column-level security based on tags to protect its most sensitive data. 

Aside from security, having a single version of truth is very important for Landkreditt. That’s why it uses Zero-Copy Cloning to expedite and accelerate development timelines. “Snowflake Zero-Copy Cloning is a truly unique feature — we love it,” explains Gislerud. “It clones all our ETL, development and product data to ensure we always have just one version of truth. It gives us complete control. And it makes Snowflake a no-brainer because there are no duplicated storage environments or associated costs. All this modern data development has driven a huge cultural change of efficiency and speed in the organization.”

With 40 data domains and 4 billion rows of data for its insurance vertical alone, Landkreditt has scaled its data processing volumes and performance and increased efficiency in its relatively small team without incurring significant costs. 

Compared to its legacy environment, Olimb and Gislerud believe they achieved immediate ROI and can now demonstrate the “why” of data infrastructure spend to the group's Chief Financial Officer far more accurately. 

“Snowflake just works — and it’s a fair cost for a fair service,” says Olimb. “Unlike other solutions, Snowflake offers all the benefits of SaaS but without any of the management burden. What’s more, Snowflake’s documentation is great, meaning you can do most things on the platform yourself. But we also want to credit Snowflake’s Nordic and Compliance teams who have been helpful with addressing compliance, technical and functional enquiries.”

“While our old way of doing things was very time consuming, we’ve seen a threefold reduction in development times with Snowflake — the development speeds are insane,” adds Gislerud. “And while we’re not the biggest bank in Norway, we’ve achieved some fantastic things. This wouldn’t have been possible without Snowflake.” 

Looking to the future, Landkreditt plans to migrate its entire group’s data to Snowflake and continue to develop its DAIRY platform. 

The group also plans to expand new data deliveries, including reporting, ESG, machine learning and AI. In fact, shortly after migrating DAIRY, Landkreditt began experimenting with Snowpark — Snowflake’s set of libraries and runtimes for functional coding languages, such as Python. Snowpark enables Landkreditt’s machine learning model development, creating models without the need to move data out the platform, ensuring security, scalability and deployment speed. The company is also experimenting with Native Apps and Streamlit to support how it distributes insights to its agents.