80% Less Data, 90x Faster Operations: Jimdo's Data Cleanup Story

This customer story comes from Kai Richardson, Data Platform Manager at Jimdo, an all-in-one solution designed to help micro-businesses build their online presence. Jimdo enables entrepreneurs to create websites, online stores, manage bookings, design logos, and optimize for SEO, analytics, domains, and hosting. Established over a decade ago, Jimdo has supported hundreds of thousands of customers across various countries. Headquartered in Hamburg, Germany, Jimdo provides a comprehensive platform tailored for small business needs.

Introduction

"My role is to ensure that our data is organized, reliable, and easily accessible for everyone." Kai Richardson - Data Platform Manager, Jimdo

At Jimdo, I lead a tight-knit data platform team, working hand-in-hand with our data ops and company reporting teams. Our mission is straightforward but crucial: make sure everyone – from product teams to marketing – has the data they need to make informed decisions. Being digital-first means data is at the core of everything we do, whether it’s creating personalized experiences for our customers or driving key business decisions.

Our data team consists of a platform team, and multiple analytical teams each focusing on a specific area : Product, Marketing and Finance. Together, we manage a data infrastructure built on:

• AWS for data ingestion via DMS and APIs
• Snowflake as our primary data warehouse
• Mixpanel and Tableau for reporting
• CastorDoc as our data cataloging tool

The Challenge: Our Complex Data Landscape

“When I joined Jimdo 18 months ago, I found myself navigating through a vast data landscape. Our data warehouse contained an overwhelming number of tables.” Kai Richardson - Data Platform Manager, Jimdo

When I joined, it became clear that the sheer volume of data was overwhelming for a company of Jimdo's size. However, the issue was more than the amount of data; we also needed discoverability and ownership to truly make the most of our data landscape.

With blurred lines of ownership for tables and data pipelines; our platform infrastructure had grown without consistency, creating confusion between teams. I realized that no one had a clear picture as to what data was being persisted, or where they could find the data. It wasn’t unusual for employees to rely on a central data analytics team to find data for them, which could be a slow process over burdening analysts with ad hoc requests.

Over time, our systems were starting to weigh down data operations. ETL jobs ran slowly, analysts were becoming the sole source of truth and our every daily focus became maintaining our current platform. Different teams worked with different versions of the same data, meaning that data discovery was difficult, and reports were being made based on varying data sources.

First Steps: The Migration to Snowflake

"The migration itself went smoothly, but it revealed a bigger challenge: our data infrastructure was cluttered and needed a cleanup. It became clear that our next major task was to reorganize and simplify the entire system." Kai Richardson - Data Platform Manager, Jimdo

The first major project I joined was migrating our data platform to Snowflake. The move to Snowflake was essential to modernize our infrastructure. With data more accessible, and more people data curious, we were able to identify the next problem to solve: data discovery. With many unused or legacy tables that clouded discovery, a spring clean was overdue.

The Cleanup Project: Simplifying Our Data Infrastructure

"It wasn’t just about data confusion anymore – it was about our ability to function smoothly as a data platform. The complexity of our data infrastructure had become unmaintainable, and it was clear we needed a solution to make a change." Kai Richardson - Data Platform Manager, Jimdo

With analysts spending lots of their time answering ad hoc data discovery requests and engineers spending their time maintaining the data platform and pipelines, a change needed to be made. With support from the executive board, our team initiated a major cleanup project. Over the next three months, we focused on reducing the complexity of our data system. The goal was clear: get rid of unnecessary data, simplify our models, and make sure everything left was easy to manage and essential.

The Solution: Leveraging CastorDoc

"As we realized how much cleanup was needed, we knew we needed a systematic approach to figure out which data was still useful and what could be eliminated." Kai Richardson - Data Platform Manager, Jimdo

We had built up an unmaintainable amount of data assets. The team was eager to start from scratch, but practically, that wasn’t feasible. Instead, a systematic approach was required to identify what was still necessary and what could be streamlined or removed. The goal was to make the data warehouse more manageable, improving both the platform’s performance and the overall accessibility of data for employees.

This is where CastorDoc became essential. CastorDoc had already helped us during our Snowflake migration, and it quickly proved to be a critical tool in our cleanup process. We used CastorDoc’s unused tables report to easily identify which tables were active and which ones hadn’t been touched in many months.

We set a clear rule: every table needed a clear owner within the next two weeks. This created a sense of accountability across the company, pushing teams to work together and decide what was important and what could go.

Each week, we exported reports from CastorDoc to track our progress. Its lineage feature was a game changer, helping us understand how data assets were connected before we made any changes. The lineage feature became our safety net, it helped us avoid breaking anything downstream as we cleaned up.

The Impact: 80% Reduction in Data Assets and Major Efficiency Gains

"In the last few months, we reduced our data estate by 80%, and we couldn’t have done it without CastorDoc. The platform’s ability to help us analyze dependencies and challenge teams to clean up unused assets was critical to the success of this initiative."  Kai Richardson - Data Platform Manager, Jimdo

In just a few months, we saw dramatic improvements across the board. Our table count dropped by 80%. But the benefits went far beyond the numbers:

Our daily ETL jobs, which previously took hours to run, now operated 2.5 times faster, saving us 25 minutes every single day. Tasks that once took analysts 10 minutes, like creating a pull request, are now completed in just half the time.

Beyond time savings, this optimization brought significant financial benefits. We reduced data ingestion costs by 95% each year and cut ETL run costs by 20%.

This wasn’t just about cleaning up for the sake of efficiency. The streamlined system meant that our teams were no longer bogged down by unnecessary data or complicated processes. We could focus on what truly mattered – making informed, timely decisions based on accurate, accessible data. We’ve gone from drowning in data to surfing it and that shift is making all the difference in how we support our teams and serve our customers.

Looking Ahead: Building on Success

Our success with this cleanup project is only the beginning. As we move forward, we’re focused on increasing ownership and accountability for data models and dashboards. CastorDoc will remain a central tool in this process, helping us integrate more deeply into our workflows and ensuring that our data remains well-governed and accessible.

Our future plans include tighter integration of CastorDoc into our internal documentation systems and even exploring potential Slack integration for real-time updates.

Ultimately, this transformation has been about much more than technical fixes – it’s changed how we think about and work with data at Jimdo. We’ve learned the importance of collaboration, governance, and the right tools in building an efficient data ecosystem. Looking ahead, we’re excited to continue optimizing our infrastructure and maintaining a strong data culture across the company.