NTT DATA Business Solutions
NTT DATA Business Solutions | August 29, 2023 | 8 min read

7 Mistakes to Avoid When Implementing a Data Fabric

Meet Jacob Orup Lund, a seasoned expert with over two decades of SAP experience who currently holds the position of Senior Director of Data & Analytics Transformation at NTT DATA Business Solutions. With a special focus on data fabric, designing architectures and governance models, Jacob’s insights redefine how organizations unlock the true potential of their structured and unstructured data.

In the following, he shares his advice to avoid budget overruns on big projects.

The Trap

Persuading decision makers to invest in a new platform often takes longer than expected. However, at this point we are all susceptible to a common trap, we assume the hard decisions have already been made and as a result, underestimate how time consuming the implementation process will be. If you’re in the middle of a data fabric implementation, or are working towards one, we know just how long it can take to reach agreement and sign off on a project. Jacob Lund from  NTT DATA Business Solutions has guided more than 20 data fabric implementations, including many re-implementations of projects that had come off-track. While every organization is different, the mistakes made in these projects can be categorized into a handful of common themes. If you’re involved with a data fabric implementation, at any level, here are Jacob’s biggest warning signs to keep an eye out for.


The urge to ‘do it once and do it right’ can be strong, particularly for those who’ve been involved in past software projects that ran over budget. However, while he’s seen many businesses attempt to build an all-encompassing data fabric, Jacob is yet to see one successfully pull it off.

In each case, the project leaders tried to implement every capability the system might need in future. While anticipating future needs has obvious value, focusing too far into the horizon can be just as harmful as ignoring it. Money is wasted on unnecessary features and the platform takes so much time to implement, often years, that its core functionalities become superseded before launch.

To avoid the trap of overcomplexity, Jacob recommends that businesses invest a maximum of 5% of their budget on capabilities that don’t have an immediate use case.

Failure to Think Short-term

The day-to-day operations of our business can be quite boring in comparison to the exciting new project. However Jacob has seen some organizations allow their smartest employees to invest so much time in disruptive projects that regular operations have missed out on minor upgrades. As the core business becomes less efficient, less revenue is available for future projects and the disruption process becomes increasingly difficult. Ironically, preparing for the future also requires a focus on the past.

If you’re yet to dive into the story of WeWork under the leadership of Adam Neumann, this provides a classic case study in chasing a long term vision at the expense of short-term sustainability. With billions of dollars in venture capital from Softbank (who’s founder Masyoshi Son famously laid a 300-year vision plan for) Neumann chased growth relentlessly, opening dozens of new WeWork locations across the globe. In preparation for its IPO, Neumann bet that retail investors would value the business at the earnings multiple of a tech company, as opposed to a traditional real estate business. The IPO did not go as planned.

There’s another reason to avoid neglecting the short-term. Predicting the future is, quite simply, hard work. We can identify one trend and imagine the scenario when it’s taken to the Nth degree (eg. ‘What will happen when every taxi driver loses their job to Uber?’), however things often turn out a little messier than this. For example, while ride sharing crashed the value of New York’s taxi medallions several years ago, taxis are still common, and this group of workers were by no means the only people affected. In many cities, more parking spaces have become vacant and the concentration of inner-city dwellers has grown, while gig work has also impacted healthcare and taxation. These second-order effects are much more difficult to predict.

Making Data Too Public

Collaboration is a tenet of modern business, however the drive to eliminate silos and share data as widely as possible often comes with unintended consequences. While most clients Jacob has worked with have no difficulty keeping their most sensitive information private (payroll details, board meeting notes, trade secrets) it’s easy to forget that sensitive information can be inferred from metadata and other related information. For example, an executive meeting may take place behind closed doors, however if the meeting name and participants are viewable on a shared calendar, or notes are stored in a locked folder with a suggestive name, details can be assumed and acted upon as if the meeting were public. Leaks such as these don’t require concerted efforts to snoop. For example, consider the employee tasked with auditing emissions from executive travel as part of a business’ net zero obligation. If origins and destinations were visible, what could a pattern of repeated visits to the headquarters of a competitor suggest?

Making data too accessible to employees and third parties also introduces security risks. Verizon’s 2017 data breach, which exposed the personal information of 14 million customers, resulted when a contractor accidentally made customer data available on a public AWS server. With greater accessibility of data, comes a greater need to invest in data governance and security.

Building a Digital Core for the Future

Transferring the data to the customer’s SAP system in the future would enable the company to improve its planning. We’ve developed an SAP backbone for the client, the Digital Core. Now the company can use this as a framework for innovation.

Jacob Lund, Sr. Director - Data & Analytics Transformation at NTT DATA Business Solutions Nordics

Failure to Talk to Users

One of Jacob’s favorite case studies is of a global shipping client who invested 18 months of work and hundreds of thousands of euros developing a ‘control tower’ to monitor and analyze their supply chain. The company’s IT department specified the requirements to the software developers, without involving the employees who would actually use the platform in their day-to-day work. The project was delivered on-time and on-budget, but after go-live it was used by only a handful of the people it was built for, as it did not meet the actual requirements of the majority of users. While all their data was available in a dashboard that, to IT, seemed user friendly, the actual users found their original method much easier, and stuck with manually-updated Excel spreadsheets. If the business had adopted a product-thinking mindset, and communicated with the actual users, this trap could have been easily avoided.

Continuing to Invest in a Failing Project

Another cautionary tale involves a client in utilities, who needed to consolidate three data platforms into one, following a series of mergers and acquisitions. 10 million euro was invested to build a single new system that would replicate all of the functionality of the three legacy platforms. The architecture was overly complex and recreated functions that didn’t work well in the original version, essentially solving nothing. Unfortunately for this business, starting over from scratch was their most cost effective option. However, due to pressure from executives and investors to finish the project as soon as possible, the client continued to invest in fixing the platform’s shortcomings, instead of rebuilding. When trying to consolidate multiple platforms, it’s important to base the new platform on actual requirements and not the sum of all the capabilities, custom coding and bespoke solutions from the original platforms.

Another business in Jacob’s books has invested more than 50 million euro on a single SAP platform. They have a backlog of 10,000 hours to fix small problems, but understandably, nobody has the confidence to propose starting over as it requires another 40-50 million in investment.

The solution to this problem is obvious, though not so easy to implement. Employees need to feel free to admit honest mistakes and change their mind without the fear of unreasonable consequences.

Failure to Invest in Change Management

After a large investment in the technical aspects of an implementation, it’s easy to fall into the trap of not allocating enough budget to ensure the system is actually adopted. A classic example of this is the US Mint’s multiple failed attempts to introduce dollar coins, most recently in 2007. According to a Government Accountability Office report from February 2011, it was estimated that replacing the $1 bill with a $1 coin could save the U.S. government approximately $5.5 billion over a 30-year period. Despite the obvious savings, the Mint has not made the decision to force adoption of the coin, and North Americans, out of fear of loose change, vending machine incompatibility, and sheer habit, have barely taken advantage of it.

This isn’t to say that organizations should always force change on their employees, rather that it’s better to make a decision and commit to it, than endlessly test. In his environment, Jacob often hears executives list all the technical capabilities they require (data lakes, a data mesh, front-end tools, machine-learning…) without realizing these require comparable investments in changing the culture of the organization. More than a one-off series of training sessions, change management requires mechanisms for user feedback, metrics to monitor adoption rates, a culture of continuous learning, as well as the ability to iterate and improve, to name a few.

Failure to Invest in the Long-term

Perhaps the most obvious, and frequent mistake of all, is simply to never start. One of Jacob’s clients in the transportation industry has allowed their systems to run for so many years without updates, employees have stopped using company software for some tasks. Data is exported to Excel and edited manually in a much less efficient and secure manner than desired. The company operates with revenues of more than 200 million euro and attempts to carry out all financial consolidation in standard Microsoft Office products.

We Help to Overcome the Challenges

Jacob Lund’s insights highlight the multifaceted challenges in data fabric implementations. To ensure success, it’s crucial to balance long-term vision with short-term goals, foster open communication with end-users, prioritize data governance, and invest equally in technical and cultural aspects of a project.

Within the dynamic data management landscape, NTT DATA Business Solutions emerges as the ideal partner to navigate the intricacies of data fabric implementation, fortified by extensive project experience. Our commitment to supporting clients through challenges and collaboratively devising solutions solidifies our position as a trusted SAP partner.

With Jacob and his team at NTT DATA Business Solutions, you can access consulting at any stage of a data fabric implementation. Whether your organization already has a project underway and could use a second opinion, or you’re ready to begin the journey, we’re here for you.

Contact Us

You might also be interested in