Building the data organization that delivers (and creates revenue)
But most organizations overlook that a well-structured data organization isn't just a cost center that enables better decision making. For those sitting on unique, large-scale datasets, it can become a significant revenue engine.
The data monetization opportunity
Before we dive into organizational structure, let's address the elephant in the room. If you have proprietary data at scale, you're potentially sitting on a gold mine.
Consider these examples of companies that transformed their data into revenue streams:
- The Weather Company (acquired by IBM): they generate hundreds of millions in revenue by packaging their hyperlocal forecast data for industries from insurance to agriculture to logistics. Retailers use it to predict ice cream sales; airlines use it to optimize flight paths; insurance companies use it to price policies.
- Mastercard: its data and services division generates over $5 billion annually by providing anonymized, aggregated transaction insights to retailers, financial institutions and governments. They're not selling raw transaction data; they're selling intelligence about consumer spending patterns, market trends and economic indicators.
- John Deere: the company monetized the machinery and agronomic data collected from connected farming equipment, offering subscriptions for precision agriculture insights that help farmers optimize yields. What started as equipment telemetry became a data business generating hundreds of millions in recurring revenue.
Organizations can create new revenue streams by carefully packaging and delivering insights while respecting privacy and competitive boundaries. If you're sitting on unique datasets with external value, your data organization needs a dedicated capability for data monetization.
Why separating data from IT isn’t optional
According to recent research, 89% of executive teams are now directly involved in generative AI decisions, with cross-functional collaboration emerging as a key success factor. This isn't happening in IT steering committees but in strategy sessions, product roadmaps and revenue planning meetings.
The skills required for modern data work are fundamentally different from traditional IT:
- Data Engineers: build pipelines and infrastructure, but they're thinking about data quality, lineage and enabling analytics, not just uptime and security.
- Analytics Engineers: sit between data engineering and analytics, using tools like dbt to transform raw data into business-ready models. This role barely existed five years ago.
- Data Scientists and ML Engineers: they require advanced statistical knowledge and experimental design skills that have nothing to do with traditional software development.
- MLOps Engineers: combine DevOps practices with ML-specific concerns like model drift, feature stores and training pipeline orchestration.
- Data Product Owners: need to understand both technical constraints and business value in ways that traditional product managers may not.
These roles require different compensation bands, career paths and management approaches than IT roles.
Models for structuring your data organization
There are a few emerging approaches for establishing and governing data.
Centralized data organization
In this model, all data capabilities from engineering, science, analytics, operations and architecture report through a single leader (often a Chief Data Officer or Chief Analytics Officer).
This is ideal for organizations with strong data monetization potential, where data is a product and highly regulated industries requiring strict data governance.
Decentralized (federated) data organization
Here, data professionals are embedded within business units and product teams, with a small central team for infrastructure and standards.
This tends to be most successful in large, diversified companies with distinct business units and product-led organizations where data is core to product experience.
Hybrid (hub-and-spoke) data organization
This combines a central platform team with embedded analysts and data scientists who have dotted-line reporting to a central data leader. This works for organizations pursuing both internal analytics and external data monetization and balances speed with autonomy, enabling both centralized data products and business-unit-specific analytics.
Success patterns from those getting it right
After analyzing dozens of successful data organizations, several patterns emerge. Organizations that have a C-level champion (CDO, CAO or CTO) with both budget authority and a seat at the strategy table are key. If you have a data asset, leaders who monetize data successfully treat it like a product with its own P&L, roadmap and customer-facing team.
Just as companies invest in CRM systems or ERP systems, successful data organizations get platform investment that's sized appropriately (typically 2-5% of engineering budget for the core data platform).
Finally, providing career paths for data professionals requires advancement opportunities within data, not being forced into IT management or software engineering tracks.
The bottom line
The data organization that delivers isn't just structurally sound; it's strategically positioned. Whether you're building for internal analytics excellence, AI enablement or external data monetization (or all three), separating data from IT, leaning into the right structural model, and dedicating time, money and talent to data product capabilities matters.
The companies winning with data aren't just hiring data scientists and hoping for the best. They’ve constructed purpose-built organizations that treat data as the strategic asset it is. They're also discovering that their data is worth as much as their products.