8 data strategy mistakes to avoid

With nearly 800 locations, RaceTrac handles a substantial volume of data, encompassing 260 million transactions annually, alongside data feeds from store cameras and internet of things (IoT) devices embedded in fuel pumps.

“This scenario has led to the development of disparate formulas, processes, and definitions within each business unit and department for report generation, thereby generating varying conclusions and recommendations from the same dataset,” Williams says.

To break down silos, the company has created a unified data environment that integrates data across various systems for data sharing throughout the organization. “Implementing a centralized data management system and encouraging interdepartmental communication will play a key role in guaranteeing the consistency and accessibility of reliable data across the organization,” Williams says.

RaceTrac is leveraging Alation’s Data Intelligence Platform to centralize data as well as provide self-service analytics for users as needed.

Decentralizing data teams

Similar to creating silos, decentralizing data teams can create problems for organizations and diminish value.

“An isolated data team structure can be particularly problematic for organizations looking to develop and scale an effective data strategy that drives business outcomes,” Vanguard’s Swann says. “Rather, structure data teams to be organizationally centralized [and] physically co-located with the business — with objectives aligned to that business.”

This approach helps to establish a unified data ecosystem that enables seamless data integration, sharing, and collaboration across the organization, Swann says.

“Close collaboration between data professionals and the business also provides valuable and continuous insight, refines processes, builds efficiencies, and reduces friction across key operational areas,” Swann says. “This type of environment can also be deeply rewarding for data and analytics professionals.”

Ignoring data governance

Data governance should be at the heart of any data strategy. If not, the results can include poor data quality, lack of consistency, and noncompliance with regulations, among other issues.

“Maintaining the quality and consistency of data poses challenges in the absence of a standardized data management approach,” Williams says. “Before incorporating Alation at RaceTrac, we struggled with these issues, resulting in a lack of confidence in the data and redundant efforts that impeded data-driven decision-making.”

Organizations need to create a robust data governance framework, Williams says. This involves assigning data stewards, establishing transparent data ownership, and implementing guidelines for data accuracy, accessibility, and security.

Employing data intelligence platforms specifically for data lineage, governance, and collaboration “can guarantee that all members of the organization rely on a reliable source of truth for analyses and reports,” Williams says.

Using poor-quality data

Data is really only valuable to an organization if it’s accurate; otherwise, it can lead to poor decisions and even undermine customer experience.

Dirty data or poor-quality data is the biggest issue with AI, Impact Advisor’s Johnson says. “Generative AI, in fact, is a great example of this,” he says. “Their large language models have poor or dirty data. The evidence is in the production of ‘fabricated’ sources and facts that they cite in response to queries.”

Data cleansing tools are one way to address the problem, Johnson says. “However, it comes back to a well thought out data strategy with a common data model” for entities, attributes, relationships, data types, constraints, hierarchies, and so on, he says.

Lacking visibility into real-time data

Without the ability to leverage real-time data, companies can miss opportunities to adapt to changes in customer demand and provide better customer experience.

“In the rapidly evolving landscape of the business world, having the capability to promptly access and comprehend real-time data is crucial, providing organizations with a competitive edge,” RaceTrac’s Williams says.

Without a comprehensive perspective on organizational data, it becomes challenging to discern the data’s intended purpose, determine its accuracy, enhance its quality, and identify redundancies, Williams says. This can lead to the use of unreliable, substandard, or outdated data in the decision-making process.

“Transforming reliable data into an asset that spans the entire enterprise requires data users to possess a thorough understanding of the complete data lifecycle within the organization,” Williams. Since RaceTrac’s data transformation, “we’ve streamlined compliance with regulations, simplified impact analysis, and can promptly notify stakeholders of changes in upstream data in real-time,” he says. “This empowers data users to make decisions informed by data and in real-time with increased confidence.”

Overlooking diverse backgrounds when acquiring talent

Enterprises need professionals with data expertise, and to fill positions to help execute data strategies it might make sense to broaden the candidate pool.

“Organizations that limit their search for data and analytics talent to those with extensive coding or programming backgrounds may find it difficult to build an effective data strategy,” says Vanguard’s Swann.

“Diverse teams are associated with increased innovation, more informed decision-making, a wider scope for problem-solving, and an enhanced understanding of clients’ needs and preferences,” Swann says. “Because of that, a ‘one-size-fits-all’ approach to data and analytics talent can hinder collaboration, diversity of thought, and increased performance.”

Vanguard hires individuals from all backgrounds for its Chief Data and Analytics Office, including some who have studied high-level math, English, and business, Swann says.



Source link