Low data quality and adoption make it difficult for businesses to extract value from their data. Data experts need to step in more like business experts to make the translation from business objective to data project.
Companies receive an abundance of data. Yet they often don’t know what to do with it, or how to extract value from it. And, are the data accurate? “Data quality within organizations is a big problem,” notes Ziad A Fayad, Data Cloud Specialist Lead at Salesforce. Still sitting around the table are Christophe Robyns, Managing Partner at Agilytic, Mathias Coopmans, Cloud and Architecture Lead at SAS and Thijs Paepen, Account Manager at Ctac. Together, the experts discuss the role of data within companies, its quality and value, as well as the adoption of data and technologies within an organization.
I have data, now what?
“Data quality and data silos are still major challenges for companies, even after years of technological advances,” Fayad begins. Paepen concurs: “Many companies still struggle with their data: from a jumble of data to duplicates or incorrect data.”
That low data quality prevents companies from extracting enough value from it. “This often leaves them at a basic level when it comes to analytics,” Paepen states. “Even simple questions like: ‘What business problems can I solve by using my data?’ remain unanswered,” Robyns adds.
Data in silos
One of the common data management challenges is data silos. These are separate collections of data within the same organization that are stored in isolation, and not easily accessible by other departments.This leads to inefficiencies and barriers to using the data effectively. Coopmans sees a possible explanation for the existence of data silos. He emphasizes that in large organizations, data silos often represent power, which can lead to their deliberate perpetuation. “Not all causes of silos are logical,” he notes. “Internal politics certainly play a role as well. So the solution to this is not going to be technical in nature, but rather focuses on change management.
Beyond the hype
“Data projects have been around for at least 20 years,” Fayad states. Yet a heavyweight hangs over such projects. “When someone sells a solution that has nothing to do with data, the implementation is likely to be quick. The moment you start talking about data, most people automatically think the project will take years,” Fayad explains. Robyns agrees and adds another important first step before talking about data with customers.
“We don’t start by asking about their data, but we do want to know the company’s pain points. What are the business goals?” Robyns illustrates, “There was an organization that had developed an impressive model. Months after implementation, they found that no one in the company was using the model. The reason? It did not meet the needs of the company and thus proved useless.” This highlights the importance of clear goals and business propositions.
Adoption
Companies have a lot of data, but therefore do not know what to do with it. “A lot has to do with adoption,” Coopmans continues. “Adoption within the same company is slowed by different speeds in how teams and departments deal with data and new technologies.”
Adoption within the same company is slowed by different speeds in how teams and departments deal with data and new technologies.
Mathias Coopmans, Cloud and Architecture Lead at SAS
Paepen also notes this discrepancy. “Some teams are already working with advanced data models, while others are still struggling with basic processes such as managing Excel tables.” In other words, there is a big difference in maturity, not just between companies, but even between internal departments.
Paepen also points out the importance of change in method and mindset. “Thirty percent of the time in our projects we spend on change management so that new technologies can be successfully implemented,” Paepen states.
read also
Older workers use GenAI more often than younger colleagues
Fayad builds on this, emphasizing that organizations must focus on making data accessible and understandable to all levels of a company in order to lower the barrier to adoption.
IT as a facilitator
Robyns already stressed the importance of delineating business goals before starting the technical part. Paepen also argues that the IT department is increasingly taking on the role of facilitator and needs to make the translation between business needs and technical elaboration. “The role of the IT department has changed from a guiding force to a supporting function that helps solve business problems.”
The role of the IT department has changed from a directing force to a supporting function that helps solve business problems.
Thijs Paepen, Account Manager at Ctac
Coopmans concurs: “data scientists and engineers today not only need technical skills, but also need to be able to translate business problems into technical solutions. This requires a hybrid approach where data experts also understand the business context.” Data must ultimately contribute to business objectives.
Consider before you begin
Data quality and literacy, as well as data silos, are still common challenges. Today, media give us the idea that everything is AI and everyone is working with data, but in practice we see that many companies are not ready for this at all.
“The basic question in AI and data projects often comes back to data quality and suitability: are your data ready?” states Fayad. “It is important to define your business goals and assess the feasibility of your project. Based on a feasibility matrix, which also considers data availability and quality, we can prioritize the right projects,” Robyns concludes.
This is the first editorial in a series of three on the topic of data and analytics. Click on our theme page to see all articles from the roundtable, the video and our partners.