High-quality data doesn't lie. It can confirm an organization's suspicions or hopes, or it can correct its assumptions so leaders can adapt their business approach.
This truth about data has shaped how Unisys approaches data work at every level, including architecting individual projects and leading full business units. Decades of hands-on experience, building data models, extracting insight before modern tools made it easy and translating raw numbers into decisions executives could act on, have produced seven critical lessons about data's role in business.
Data engineering, data architecture, and every other stage of that work points to the same conclusion: data's value only shows up when organizations know how to use it. These lessons offer practical guidance for teams working to turn data into a genuine business asset.
Lesson #1: Data answers business questions.
Data's biggest superpower is its ability to shape organizational direction and challenge assumptions. A business unit or department might earn praise for its success, but what does the data say?
Turn to data to answer some of your biggest business questions:
- Is your organization, business unit, or department meeting its business objectives?
- What KPIs are you achieving, and which are you failing to meet?
- What are the most promising growth opportunities?
- Which projects or initiatives require finetuning to yield the best results?
Answers drive action. Answering such questions can create significant value and guide future business planning, leading to promising endeavors like optimizing complex logistics.
Lesson #2: Business decision-makers must understand the data.
Consider a real-world example: managing the data behind a real estate portal used by several federal agencies to track government buildings, information critical to those agencies' decision-making. The volume of data was immense, and identifying where it lived and extracting value from it was a major undertaking.
That work took place about 25 years ago, when reporting tools weren't nearly as advanced as they are now. Reporting meant gathering data through disparate queries and exporting it into a spreadsheet. Even then, it was clear how important it was for decision-makers to make sense of the data and gain insight from it. So the team built reports that made the data easy to understand and surfaced the top insights for agency executives through reporting dashboards. Today, AI can both collect and report on data.
Lesson #3: A data modernization strategy is a must-have.
Data architecture is about more than interfacing with an application. From the earliest data architecture work, the role has involved making sure architecture can withstand the workload when applications require persistence or when someone starts querying the data. Many design principles outside data modeling have centered on data optimization, still a sound principle today.
That's where a comprehensive data modernization strategy becomes vital. Without one, organizations may end up with duplicate data or other issues that stall progress. This strategy should answer questions like which data is ripe to modernize and migrate, and what steps to take to secure it. Given the significant shift in how organizations align data foundations to structured, semi-structured, and unstructured data, the strategy should also include training and learning for machine learning models, steps for creating the right data pipeline and data ops.
Generative AI creates significant new opportunities, and even more incentive than ever to prioritize data. But it's also making clear that while data still plays a major role, the infrastructure that supports that data matters just as much. AI-ready enterprises need the right infrastructure to generate insight from data. With it, speed and agility become allies.
Lesson #4: An early warning system can decrease risk.
Machine learning ops platforms and the challenges of rebuilding data architectures and multi-cloud infrastructure to accommodate generative AI mean the underlying technology keeps evolving.
Organizations that don't pay attention to their data from the beginning risk paying for it later, literally. The 2007-2008 financial crisis demonstrated this clearly. Unisys data leaders overseeing data projects for financial services firms at the time saw consultants working long hours trying to identify why these companies were experiencing substantial financial losses.
The crisis is an example of what can happen when organizations aren't paying close enough attention to their data. There's no guarantee that closer attention would have prevented every problem that emerged. However, mechanisms that detect data anomalies early can help organizations catch trouble sooner and limit their losses.
Lesson #5: A data lakehouse is a superior storage option.
Many organizations store their data in file-based or image-based relational databases. However, cheap compute created data sprawl and led to challenges, including data governance. This is where a data lakehouse, cloud-based storage for structured, semi-structured, and unstructured data, offers a better option.
Moving data to a lakehouse enables a smooth flow of information that transforms every aspect of the organization, including personalized customer experiences and predictive maintenance. Data lakehouses are the answer to the data sprawl common with traditional storage options like data warehouses.
Lesson #6: Don't reinvent the wheel.
Before bringing in dozens of engineers and asking them to develop data solutions, perform a cost-benefit analysis of building versus buying. Today, in most cases, buying is the more advantageous route. This approach reduces the uncertainty associated with the probable success of a newly built solution. Building also often requires bringing in numerous data scientists, a challenge given the talent shortage.
With a purchased solution, the cost will likely be lower if an organization has sufficient data storage and a few experienced data engineers available, whether in-house or outsourced. For the best results, consider developing solid partnerships and integrating data with existing tools that can satisfy your requirements.
Lesson #7: Data is a business asset rather than a cost center.
The business and technology sides of an organization can have conflicting objectives. The business side is often eager for the most information at the lowest cost, while the technology side wants the best tools available.
Data leaders across industries have learned how important it is to respect data's value to the organization. The business side must recognize this and invest in technology solutions, because data can deliver a substantial payoff when used well. That requires looking beyond structured data to semi-structured and unstructured data. The right technology can help organizations derive value from all types of data.
Advance your data approach with Unisys
One constant about data: the learning never stops. Unisys has gathered these lessons over more than 25 years of data work across industries, and new insights keep emerging. AI is already reshaping what's possible, and its growing use is likely to deepen how organizations understand, and value, their data.
To optimize data, maximize its value, and prepare it for AI, explore more lessons in the “Data Visionaries: Five pivotal strategies and eight transformative stories” eBook.
See also “Data readiness for AI: A practical guide for preparing your data, regardless of your starting point”, and explore how Data and Analytics solutions from Unisys can help.
