Master Data Management

Lost in Translation

In December 1998, NASA launched the Mars Climate Orbiter, a robotic space probe, into the atmosphere as part of the Mars Surveyor ‘98 program. The Orbiter would serve as the communications relay for the Mars Polar Lander, to be launched a month later. Together, the two probes would enrich our scientific understanding about the alien planet for years to come.

However, in September 1999, as the Orbiter was nearing Mars, the programmed trajectory brought the spacecraft 100 kilometers closer to the planet than planned and 25 kilometers beneath the atmospheric level at which it could properly function. The navigational team had made their calculations for routine thruster firings using the metric system, while the engineering team provided acceleration data using the imperial system. The propulsion system overheated and was disabled, rendering the Orbiter unresponsive. The Orbiter would either be destroyed in the atmosphere or enter the Sun’s orbit, but regardless, because of an inconsistency in the underlying data, a $546 million investment was lost to space.

The loss of the Mars Climate Orbiter illustrates the importance of data standards and data quality within large, decentralized organizations. It’s no wonder that increasingly global corporations have teams dedicated to the governance of reference data and, by extension, master data. Master data helps businesses manage and maintain an accurate and regulated data environment throughout the enterprise. How expenses are tracked in one business unit should be consistent with how expenses are tracked in other business units, enabling organizations to leverage reliable data for key decisions.

Without robust master data management (MDM), organizations risk compromising data quality and lose opportunities to leverage data science capabilities, resulting in time wasted reconciling data and potential delays in the decision-making process. These risks increase exponentially when organizations rely on error-prone manual processes. As such, more businesses than ever recognize that fortifying MDM is key to enabling success.

Casting a Wider Net

Seeking to optimize its master data operations, a highly decentralized Fortune 50 product manufacturer asked Zenda to observe its off-shore master data team and return with recommendations to increase efficiency and accuracy in its master data processing. Zenda researchers observed a highly knowledgeable master data team skilled in handling the nuances and variability of master data, a result of the significant time spent resolving errors in master data requests. The format of the requests received from business requesters was highly variable, ranging from forms and excel files to photocopies and e-mails. 50-70% of received requests were incomplete or incorrect, necessitating consultation with requesters directly. And roughly one-third of all requests were rejected due to requesters’ misunderstanding of the master data required to support their business needs. So while the master data team was highly skilled, the team’s efficiency was correlated with transforming requestor errors into actionable requests.

Zenda’s researchers revealed that truly transformative change did not lie with making the offshore team more efficient at processing master data requests. The master data event actually extended beyond the master data team, starting with empowering requesters to send valid, complete, and accurate master data requests to the offshore team. Success would be evident if every request was appropriate and free of errors, enabling the master data team to process fewer, more accurate requests and to leverage their business acumen more strategically. The Zenda team proposed reconfiguring the end-to-end MDM process so that how data was gathered and processed reflected the differing skills, motivations, and proficiencies of each participant. Addressing the needs of both the requesters and master data team in a human-centered and systematic way would mitigate human error, normalize request volume, and build the knowledge base of requesters in a way that training alone could not guarantee.

Work Design™, Inside and Out

Zenda’s designers, in partnership with business and technology experts, were tasked to pilot an interactive

experience that would facilitate the creation, submission, and processing of master data requests. Individually, the members of the pilot team possessed the expertise that was critical to untangling the Gordian master data knot. However, within a few working sessions, it became clear that participants approached the master data experience with their own situational motivations. The data stewards were concerned with the completion and accuracy of the requests, speaking in terms of the rules defined by the data governance organization. The product owners sought to build the leanest possible request experience, speaking in terms of feasibility and minimum requirements. Zenda’s designers realized that without a common, meaningful understanding of the master data process, it would be impossible to reconcile the divergent approaches to the problem in a way that improved the current experience.

In order to advance the pilot, Zenda’s designers had to address the disconnect between the team’s diverse stakeholders. How could the Zenda team reinvent ways of working internally that were as meaningful as the solution they were seeking to materialize? The answer was to address the problem with a human-centered perspective. Echoing the approach taken in the field, Zenda designers recognized the different skills, motivations, and proficiencies of the team’s stakeholders and hosted co-design sessions to examine individual master data requirements and then express how human-centered considerations and technology would enable valid, complete, and accurate request inputs. The resultant frameworks and documentation transformed the team’s understanding of master data in a way that was no longer driven by business or technology needs independently. What was once a laundry list of requirements was made more manageable and respectful of the master data requester’s experience. The design team created the means to bring multiple perspectives together, and before long, the pilot was ready to build.

Astronomical Impacts

So what happened after the launch of the pilot? Since then, the team has extended the original interactive master data experience to include the majority of requester business needs. Within ten months of the pilot launch, request volume decreased by 89% and request rejections decreased by 86%. The off-shore master data team had expanded capacity to focus on reviewing and executing, rather than reconciling incomplete and inaccurate requests. The validations, sequencing, and guidance of the interactive experience enabled business requesters to request master data with minimal additional training, allowing the data governance organization to ensure data quality more strategically.

Beyond the tangible impacts of the program, Zenda’s scalable system of co-design enabled the broader team to transform business requirements into technical requirements quickly. Business and technology stakeholders gained an appreciation for the human concerns of the master data event, taking into account requesters’ and master data team members’ competencies and motivations. Human-centered design ultimately became a crucial component to the success of the program.

Whether the goal is to bring new products to market or launch a probe into outer space, dependable data is key to operating and scaling large enterprises. To address a situation as extensive and complex as managing master data requires a level of thinking that is as nuanced as the problem to be solved. Designers, with their ability to meaningfully bring order to seemingly disparate elements, are invaluable assets to any team. Design in this context, which is often assumed to be limited to the design of end user interfaces, actually includes design of the dynamic event of work. When adopted and embedded, design can permeate organizations, rising to challenges in unexpected ways.