Reflections from CDOIQ 2023

Shifting the CDO to a Profit Center - The first decade of the CDO roles has focused heavily on integrated data, providing unified views across an organization and providing leaders with trustworthy KPIs. Now that many organizations have the foundations, we are seeing a pivot to the CDO holding accountability for a P&L as any other business unit or line of business leader. CDOs must evolve their organization to innovate with new ways to leverage and monetize their data. While we have a ways to go in applying standard methods to value our data like any other asset, we have all the necessary frameworks today to build partnership around data exchanges, monetize our data through marketplaces and build new data driven consumer products that bring new revenue.

Experimenting With Impact, Not Research - This was a theme that came up several times. Shifting the culture mindset away from research to experimentation with new methods, technology, partnerships and products to measure the impact and make informed decisions about future investments. This directly influences how a CDO can become a profit center, by iterating quickly and leveraging feedback to improve both approach and products.

Data Literacy as Mainstream L&D - Data Literacy has traditionally been part of data governance programs, ensuring that teams are effectively using technology to protect data while using that data to tell compelling stories about business performance. Now that data is part of all jobs, we see corporate Learning & Development taking on this challenge to educate and mature the organization's skills. Career rotations are common as a tool to provide visibility for data teams across a variety of teams and business functions to inform their own thought process and spur creativity.

Data Products as Glue Between Data Domains - Many organizations have gone on a journey to define their data domains. These often serve as the anchor for data governance responsibilities, while extending into technical implementations across complex organizational structures where a single data platform is not feasible. But Data Domains do not provide value by themself. Data must be consumable for action. The Data Product has become that consumable component and measure of value, where value is maximized when spanning multiple data domains for visibility deep into a business's performance.

Move from Data Quality to "fit for use" - Many data quality programs have gotten stuck as they work to define endless metrics that measure data completeness, accuracy and other elements at multiple stages through the data transformation lifecycle. While these checks are valuable and necessary to ensure failed processes do not affect downstream data consumers, they are missing the larger objective. Every data product has a defined set of consumers and users for this data. That use should drive data quality standards and inform levels of investment for improving data quality. For use cases in the medical field, the bar is much higher for "fit for use". Patients' lives could depend on it. But for targeted marketing use cases, some incomplete data may very well be acceptable without impact to outcomes. Always work to ensure you define "fit for us" as your priority over endless data quality measures.

Growing Data & Compliance Obligations - For years now, data teams have focused efforts around CCPA, CPRA and GDPR as the bar for how consumer and employee data is managed, tracked and purged. These collectively created a set of obligations for companies to acknowledge the true owners of data being collected and analyzed. While many regulations today fail to account for the capabilities and risk inherent in the application of AI, that will change quickly as the courts weigh in and governments implement new laws. One particular area of risk is the "ghost" of an individual; this is the impact that an individual's data has when being used to train a specific analytical model. With the growth of LLMs the potential for these "ghosts" to expose data or derived data only grows. Organizations should begin thinking now about how they protect consumer privacy, not just with data collection and storage, but also with analytical models.