Legendary Consumer Brand Improves Data Discoverability, Impact Analysis, and Business Collaboration on Data
At a Glance
- Dr. Martens, an iconic global footwear brand with a six-decade heritage, evaluated the data catalog space in order to drive self-service atop their quickly modernizing data stack.
- Choosing Atlan, their data team quickly implemented a self-service catalog to provide context around their most critical data assets.
- Atlan’s implementation has accelerated time-to-insight for Dr. Martens’ internal data consumers, and is reducing time spent on impact analysis from four to six weeks, to under 30 minutes for data practitioners.
Dr. Martens is an iconic British brand founded in 1960 in Northamptonshire. Produced originally for workers looking for tough, durable boots, the brand was quickly adopted by diverse youth subcultures and associated musical movements. Dr. Martens has since transcended its working-class roots while still celebrating its proud heritage and, six decades later, “Docs” or “DM’s” are worn by people around the world who use them as a symbol of empowerment and their own individual attitude. The Company is a constituent of the FTSE 250 index.
Of late, Dr. Martens has been steadily growing and evolving its business, with 52% of their sales direct-to-consumer in FY’23. Crucial to this growth, past, present, and future, is a visionary data team that offers modern technology and insights to their business colleagues tasked with making the best decisions possible.
Among these data visionaries is Karthik Ramani, Global Head of Data Architecture for Dr. Martens.
“I started off from a user’s perspective in a Business Intelligence role, then Data Warehousing, then Data Engineering before getting into Data Architecture. I’ve had visibility into the end-to-end of data, and I’m passionate about guiding people to get the most value out of data, processes, people, and frameworks,” Karthik shared.
And responsible for ensuring Dr. Martens’ data is governed, accessible, and contextualized is Lawrence Giordano, Data Governance & Strategy.
“I found myself in Data Governance because I’m passionate about it. I’m here to prove that it’s not red tape, and it’s not about stopping people from doing stuff,” Lawrence shared. “We can offer curated data sets while also looking after our data the right way. Data Governance actually enables other functions to do their jobs better.”
Delivering Sustainable and Profitable Growth
Guiding and prioritizing Dr. Martens’ business and technology decisions is the DOCS strategy, representing four pillars of Direct-to-consumer First, Organizational and Operational Excellence, Consumer Connection, and Support Brand Expansion with B2B.
Recent examples of execution on this strategy include opening new retail stores in existing and new markets with omnichannel experiences, supported by technology modernization and supply chain improvements.
“Most initiatives at Dr. Martens will associate themselves to one of those core pillars, and we’re no different. On the data team, we can link ourselves to all four, but especially Organizational and Operational Excellence,” Lawrence explained.
Powering DOCS with the Modern Data Stack
Among the most important ways the data team supports the DOCS strategy is a new way of working, an agile, product-led delivery methodology where analysts and engineers are embedded within product teams. Interacting with their business colleagues every day, and owning the results of their work, means that Dr. Martens’ data team better understands the business problem they’re helping to solve.
Ready and able to support these business functions is a team structure composed of five core functions, Data Engineering, Data Architecture, Data Analytics, Reporting, and Data Governance, reporting into the Dr. Martens Global Data Officer, Nick Sawyer.
“It’s a matter of how we get all these functions to work smoothly with each other to solve a business problem, which might not fit neatly into each of these pillars and requires us to come together,” Karthik shared. “Our focus has always been to align to business objectives, and on how we can drive value from the data and deliver to the business.”
Continuing through its rapid growth phase, and transforming into a company that services customers across multiple channels, including digital, data plays a more important role than ever in guiding Dr. Martens’ decisions, driving their team to quickly modernize their data stack.
As part of our transformation, we recognize that data is a fundamental and a critical pillar to understanding our customers’ experiences and needs, and guides how we can improve and optimize. There’s been significant investment in modernizing our data platform to address challenges. We needed to move towards a single source of truth, and increase the reliability and scalability for delivering insights for the various departments we serve. We’re essentially removing technology as a barrier to using data and finding insights.”
Karthik Ramani, Global Head of Data Architecture
Starting with Microsoft Azure as their cloud provider of choice, Dr. Martens’ new, best-of-breed data stack consists of dbt for transformation, Snowflake as their data warehouse, and PowerBI for reporting and visualization, providing a modern foundation for further growth.
Driving Data Transparency with a Modern Data Catalog
With a new way of working that prioritized a closer relationship between the data team and their business counterparts, and with an array of new data technology, Dr. Martens’ data team needed a way to make these new capabilities and assets transparent and understandable to a spectrum of internal data consumers.
Creating a “Restaurant Menu” for a Modern Data Stack
Moving from legacy technology into a modern environment, Karthik and Lawrence sought a platform that could serve as a “data menu”, presenting crucial context about their data assets in an easy to understand manner.
“Transparency of data ownership, lineage and quality was going to be a huge driver for us if we were really going to demystify our data estate,” Lawrence explained.
In the absence of a modern data catalog, questions about data would continue to drive a costly back-and-forth, where data consumers needed to reach out to the data team each time they had simple questions about definitions, freshness, and calculations.
“There was a huge amount of time that was spent by our data team on information questions like ‘Where do I find this metric?’, ‘How is this metric calculated?’, or ‘Where does this field come from?’,” Lawrence shared.
Introducing self-service capability would mean not only significant time savings for technical teams normally tasked with answering these questions, but significantly accelerated time-to-insight for their business counterparts that were eager to make the most of Dr. Martens’ data.
Moreover, operating across dozens of markets and regions meant the data team was delicately balancing the needs of the global Dr. Martens entity with the unique, localized needs of various operating units. Metrics and KPIs in one market might be defined differently in another, making it difficult to speak a common language, and deliver common capabilities.
“You have to work to bring this together in a data layer, but there’s also the metadata layer, where you have to define knowledge and ownership for these assets,” Karthik shared. “That was another strong reason for creating not only a single data layer in Snowflake, but complementing it with a single metadata layer in Atlan.”
A Business-focused Evaluation Process
Rather than running their evaluation with data team members exclusively, Lawrence insisted on business involvement from the very beginning of their process. Dr. Martens’ data catalog would fail without robust business adoption, and the inclusion of these stakeholders in the evaluation would ensure that they understood the problem being solved, were champions for data transparency and speed of delivery, and that they provided valuable feedback on the user experience.
“How does a user touch and feel the product? How actively can they engage without a lot of direction, and how do we flatten the learning curve? How do we make sure that if we’re going to onboard 100 users when we launch the product, that it’s going to be a seamless process? Will they need hand holding across days, weeks, or months of training, or is it something they can naturally pick up?,” Lawrence shared.
Most important to Lawrence, however, was a sandbox environment of Atlan offered during the proof of concept that consumed Dr. Martens’ actual metadata, rather than well-curated samples, and ensured that when they conducted user testing with the business, that the results would closely mirror their future experience.
In a proof of concept, unless you look at it, feel it, and use it with your own organization’s data ecosystem, which can be messy and brings its own challenges, you can’t see how the tool adapts to that. You need to ultimately give your sponsors and users, who will be using this tool, the ability to get hands-on and say what they do and don’t like. It gets them more engaged in the process.”
Lawrence Giordano, Data Governance & Strategy
Finally, Lawrence and Karthik started building their evaluation criteria by considering what they did not want in a modern data catalog, rather than what they desired, ensuring they only evaluated platforms without “dealbreakers”.
Beginning by avoiding solutions that imposed costly integrations to their modern data tools, their ultimate focus was on usability, ensuring that their business colleagues could easily adopt the platform.
“We were clear that this was not a tech solution, and it wasn’t being built for technical teams. It’s for the business, and by the business,” Karthik explained.
A Collaborative Implementation of Atlan
Having chosen Atlan as their modern data catalog, Karthik and Lawrence carefully planned its implementation. To ensure Atlan was not perceived as “just another tool”, they adopted a philosophy of deep engagement with their business colleagues, opted for experiential learning where data consumers could discover capabilities of their new catalog, and carefully considered their first use cases to ensure the maximum possible early impact.
Ensuring Strong Business Engagement
Continuing the partnership they built with business colleagues during the evaluation phase, Dr. Martens’ data team began rollout with a series of workshops to better understand potential use cases, and to build champions for Atlan.
“We’re bringing in modern data tools to enhance our data journey, but Atlan could be seen as just another tool, in a kind of fatigue for end users. We wanted Atlan to be at the forefront of people’s minds so if they had a question on data, they went to Atlan,” Lawrence explained. “We wanted to bring them on board in a manner where it’s not seen as just another task they need to do, but that we engaged them in a way that they were part of the journey, and they want to get to the ‘promised land’, too.”
These workshops, supported by Dr. Martens’ senior leadership, ensured that the future users of Atlan felt empowered to contribute to, and consume the assets made available on the catalog, and understood the value of engaging further.
Finally, the early use cases built by the Dr. Martens data team were determined through value mapping sessions, identifying which business teams would yield the most benefit from the platform, which capabilities of Atlan could deliver those solutions, and that even the earliest users would receive value, then evangelize for further use.
Treasure Hunts for Context
With Atlan integrated into their critical data tooling, Lawrence began another series of workshops, energizing their business colleagues to further participate in the rollout.
Beginning with a showcase of the work they had completed on Dr. Martens’ analytics models, they conducted an Indiana Jones themed treasure hunt, where users were tasked with finding five pieces of information hidden in Atlan to retrieve a stolen gem. Offering Atlan swag like t-shirts, their business colleagues quickly got to work finding the information, meaningfully engaging with the platform and building a deeper appreciation for how they might use it in their day-to-day lives.
That was really our energizing moment. It showed how quickly you can answer questions, but the big takeaway from the workshop was that even though Atlan wasn’t in its ‘perfect state’, we were entering a phase where we were community driven. We were encouraging them to start feeding definitions into Atlan, building workflows, and approving curated data. It was brilliant to get their energy levels up and get them engaged in the process. They could see how quickly questions could be answered, and the long-term benefit of participating.”
Lawrence Giordano, Data Governance & Strategy
Early Wins through Alignment on Terms and Metrics
Informed by a trusting relationship built with their business colleagues, a value stream mapping exercise that ensured early work would be impactful, and workshops to cultivate an educated, enthusiastic user base, Karthik and Lawrence got to work building a metrics catalog, and a process for keeping it up to date.
Beginning with sourcing definitions then enriching critical metrics, the data team assigned owners to each of them, ensuring that when questions arose in the future, there was a subject matter expert that could address them.
“As our transformation project rolls on, we’re presenting our analytics models to the organization and that’s what triggers what we now call ‘The Atlan Process’, where we look at the analytics model, figure out what’s in there, define it, and establish who owns it,” Lawrence explained.
With this “phase one”, as Karthik and Lawrence describe it, underway, “phase two” will involve the drafting of more technical readmes describing transformation logic, tied to Atlan’s automated lineage, providing a rich understanding of Dr. Martens’ data pipelines.
Realizing Cross-functional Value
For Dr. Martens, self-service represents a significant shift, driving transparency not just for datasets, but the typically tribal knowledge that once existed around those datasets. While their data consumers stand to benefit the most from this work, their data team now use capabilities like automated lineage to accelerate issue resolution, and a “restaurant menu” for their modern data stack is driving greater appreciation for, and ROI from, the effort spent on the data transformation project.
“It’s about trust, confidence, value, speed to market, self-service capability, and ultimately lowering the barrier to using data,” Karthik shared. “Our business users are here to solve business problems, not to sit in front of their reports and spreadsheets spending hours sifting through data.”
Beyond the short-term wins Dr. Martens’ data team can deliver by enabling faster speed of delivery and decisions, in the years to come, Karthik and Lawrence predict that with data consumers crowdsourcing and curating metadata, a culture of self-learning and ownership will emerge.
Demystifying the Data Estate
Dr. Martens’ data stack transformation is not occurring in isolation. With a mandate to improve the way their organization operates, parallel projects to modernize anything from their ERP to their Customer Data Platform are driving constant collaboration between technical teams to ensure changes are implemented smoothly.
“Being in the Data Architecture function, I typically get bombarded by questions about the wider tech transformation that’s going on and its impact on Data & Analytics,” Karthik shared. “There’s a lot of change happening within our supply chain system, our product systems, our order management system, and our customer data platform. All these new solutions are driving change in parallel to our data transformation project.
Before the introduction of Atlan, each of these upstream changes meant a manual process of checking downstream systems for potential impacts, requiring significant human capital. But with Atlan’s automated lineage, Karthik’s team can determine these impacts in an infinitesimal percentage of the time they once needed.
“I’ve had at least two conversations where questions about downstream impact would have taken allocation of a lot of resources,” Karthik explained. “Then actually getting the work done would have taken at least four to six weeks, but I managed to sit alongside another architect and solve that within 30 minutes, saying ‘If you’re changing the column name or adding an extra column, this is what it’s going to break or impact.’”
While their focus on their business colleagues has quickly driven value from Atlan, interactions with technical counterparts that result in six-week time savings on expensive processes build more internal advocates for Karthik and Lawrence’s work, and drive even more value from Atlan.
“We did this together, and straight away the Domain Architect said ‘Can I get access to this platform, please?’ And I said ‘Yeah of course. You can get access to Atlan. Next time you don’t have to come to us.’,” Karthik shared.
Making a Technical Transformation Real for the Business
Concepts like a cloud-based data warehouse or a modern tool for data transformation may seem arcane to the data team’s business stakeholders, but their buy-in is crucial to a successful transformation. With Atlan helping to drive better access to data, and improving understanding around it, it’s far easier for stakeholders to understand the benefit of the data team’s focus on modernization.
Choosing Atlan as part of the transformation project helped us to tightly couple the delivery of a data catalog with all the new, shiny tools. But our main value driver is getting to a single source of truth, with everyone having access to the same knowledge base, which is consolidated and curated by the business. We were quite keen that the new operating model, based on a single, self-serviceable data catalog, meant changing away engineers, analysts, and end users conversing offline on chats and emails around data.”
Karthik Ramani, Global Head of Data Architecture
Through adopting Atlan, the new capabilities afforded by Dr. Martens’ transformation project are more understandable and usable to their stakeholders, providing context about data assets and their ownership for data consumers, and a fine-grained view into their data estate for data practitioners, all available via self-service.
And going forward, Atlan will be central to the delivery of new data models, with business teams required to provide definitions, descriptions, and ownership in parallel to making it available to data consumers.
“This is all knowledge that, historically, would have been sourced from conversations, or other means of a reactive nature. Now, it’s available and ready for them, and they get this as part of the transformation that they’ve been patiently waiting on,” Karthik shared. “It’s icing on the cake for them. We already see a change in behavior as Atlan almost starts to act as a gatekeeper for what’s actually going on in our production systems.”
Finer visibility into data assets, afforded by Atlan, is already driving behavioral change and more proactive fixes, most recently exemplified by Data Engineering learning that a data model hadn’t been successfully processed, resulting in metadata not yet available in Atlan. As more data consumers onboard into Atlan, Karthik and Lawrence hope to see more of this behavior, resolving issues before end users even realize they’ve occurred.
“We already see that change in culture and behavior happening, and we’re hoping to scale that up as we roll out more,” Karthik explained. “I would say it’s made a massive difference. From a data team perspective, this extended, additional layer helps us do governance proactively, and not as an after-effect of the transformation project.”
With Atlan as their “window to the data world”, the transformation project’s myriad stakeholders understand its benefits more, more assured that the data team are doing the right things, focusing on governance, security, and compliance proactively, in addition to modernizing their infrastructure and tooling.
A Foundation for AI and Data Governance
Dr. Martens’ data team are keenly focused on delivering what they’ve promised to their business partners as part of their transformation project, but have ambitious plans for Atlan, once completed. While they steadily roll out promised use cases on Atlan and monitor adoption, new technologies like Generative AI hold promise for accelerating asset enrichment, and increasing context around their data represents a strong foundation for improving governance.
“Some of the new use cases we’re seeing are around new features like Generative AI, which is really exciting for us. We’re one of the pilot customers with a hands-on trial of the feature, and we can see how it could make our curation process much slicker, then quicker. We now have a baseline that our users can start working off, then refine,” Karthik shared.
Rounding out Karthik and Lawrence’s future plans for Atlan include data profiling, classification, and implementing DataOps best practices, capabilities they have long sought, but only now can achieve with a platform that can bring them to life.
Lessons Learned
While there’s still work to be done modernizing their data technology, and democratizing access and context around their data assets and capabilities with Atlan, Lawrence and Karthik believe there are key considerations for their peer data leaders considering an investment in a modern data catalog.
Lawrence: Get Hands-on
Being hands-on is the biggest thing for me. You have to evaluate a piece of technology that’s embedded in your stack and your data if you’re actually going to know if it will work with your datasets, your culture, and your organization. This was the biggest thing when we evaluated Atlan. Then, it’s welcoming those senior stakeholders into the journey earlier, and bringing them closer to the benefits you intend to deliver.”
Lawrence Giordano, Data Governance & Strategy
Karthik: Work Agile
Atlan enables you to be agile and iterate quickly, so make use of it that way. Don’t make your implementations too tight and ‘waterfall-y’ where you’re trying to be right the first time. Then you’re not making use of the opportunity Atlan offers where you can try something out quickly. If it works, it works. If it doesn’t it doesn’t. Deliver value, and if it doesn’t work, leave it and move to the next thing and focus on that. Be agile. Test and learn. Try new things quickly.”
Karthik Ramani, Global Head of Data Architecture
Photo by Kilian Seiler on Unsplash
