As Data, Analytics, and AI reshape every industry, organizations are under pressure to turn investment into impact. But with constrained resources and rising complexity, the real challenge isn’t doing more; it’s doing the right things, better.
This article introduces the FRAMEwork, a portfolio-based approach to maximizing ROI, efficiency, and value from analytics and AI. It helps leaders classify initiatives across the following four categories and balance their investments accordingly:

• Foundational
• Research
• Applied
• Mandatory & Enabling

We also highlight the importance of fit-for-purpose measurement tracking outcomes that match the role of each initiative and introduce a simple operating model built around strategy, access, literacy, and trust. These elements help create the conditions for high-performing analytics and AI at scale.

Data strategy: From access to advantage

Modern organizations generate massive volumes of data, but volume alone does not create value. A great data strategy ensures that information is timely, trustworthy, and aligned with business needs. Building this foundation requires more than infrastructure; it demands intentional design encompassing data governance, quality, access, and flexibility.

Leading models like Davenport et al’s classic data offense vs. defense framework remind us that data is both a strategic asset and a regulated resource. Effective organizations strike the right balance, ensuring data can power AI and analytics without compromising on control.

Analytics strategy: A spectrum of value

Analytics and AI span a wide range from open-ended exploration to fully productized solutions. Understanding where an initiative sits on this Analytics Spectrum helps determine how to staff, measure, and scale it.

• Exploratory work (e.g., churn clustering) seeks new insights using experimental methods
• Applied analytics (e.g., demand forecasting) solves known problems with proven techniques
• Embedded AI (e.g., in-app recommendation engines) delivers continuous value at scale

Treating analytics as both an art and a science, blending business context with rigorous methods, helps organizations move from insights to action.

The FRAMEwork©: Managing the data, analytics, and AI portfolio

Data, analytics, and AI initiatives come in many forms: some are experimental, others are operationalized and scaled. To invest wisely, organizations need a clear way to evaluate and prioritize these efforts. The FRAMEwork offers exactly that: a portfolio lens that categorizes initiatives across four distinct roles based on confidence of success and potential value.

• Foundational: These initiatives establish essential infrastructure, processes, or data capabilities. Example: Building an enterprise data catalog to support self-serve analytics.
• Research: High-upside, exploratory efforts where outcomes are uncertain. Example: Testing a GenAI assistant to support store manager queries.
• Applied: Proven, high-impact solutions embedded into business workflows. Example: A machine learning model that predicts product sell-through and informs weekly planning.
• Mandatory and enabling: Initiatives that ensure compliance, user adoption, and system reliability. Example: Implementing model audit logs and access controls to meet regulatory standards.

A strong portfolio balances across all four categories, investing in today’s impact while preparing for tomorrow’s opportunity. The FRAMEwork helps leaders make these trade-offs transparently, intentionally, and at scale.

Measurement that matters

Not every data, analytics, or AI initiative should be measured the same way. A forecasting engine in production requires different success metrics than a research pilot or a data infrastructure upgrade. High-performing teams tailor their measurement approach to the purpose and maturity of the work.

The most useful question isn’t “Did it work?” but rather, “Did it work as intended for the role this initiative plays in the portfolio?”

A few practical principles:

• Align KPIs to intent: Exploratory projects may focus on learning velocity, while applied models should track impact on revenue, cost, or engagement.
• Use fit-for-purpose methods: Consider pre/post comparisons, forecast benchmarks, or business owner feedback, not just A/B tests.
• Visualize decision-making: Simple dashboards that guide action are more valuable than complex reports.

When measurement matches intent, it builds credibility and accelerates decisions, turning analytics from reporting tools into engines for action.

Operating for impact: Four strategic pillars

A strong portfolio sets direction, but real impact comes from execution. High-performing organizations bring their data, analytics, and AI strategies to life through four core operating pillars:

1. Aligned strategy: Clear articulation of what we’re solving, why it matters, and how we’ll get there.
2. Data culture: An environment built on access, literacy, and trust where strategy is understood, data is usable, and people are empowered to act.
3. Analytics flywheel: Scaled execution powered by the right levers: portfolio discipline (FRAMEwork), skilled talent, modern tools, and automation.
4. Impact scorecards: A habit of measurement and reflection: tracking outcomes, sharing learnings, celebrating wins, and continuously improving.

These pillars don’t just support the work; they accelerate it. Together, they form the foundation for an organization that doesn’t just use analytics and AI, but runs on them.

Conclusion and next steps

Great data, analytics, and AI work starts with purpose and a business strategy. The FRAMEwork offers a way to align that purpose with execution, balancing bold investments with practical enablement, experimentation with accountability, and innovation with trust.

As you consider what’s next, ask:

• Are we investing in the right mix of initiatives across our portfolio?
• Do our teams have the clarity, tools, and skills to execute?
• Are we building a culture that can learn, scale, and sustain?

With an aligned strategy, a strong culture, and a clear operating model, the return on data becomes more than a metric; it becomes a mindset.

About the Author:

Bhagyesh Phanse is Senior Vice President and Chief Data and Analytics Officer at Starbucks, where he leads data and analytics strategies to drive customer engagement, profitability, and operational efficiency. With over 18 years of experience in analytics and AI across top organizations like Starbucks, CVS Health, and Macy’s, Phanse has played a key role in transforming digital experiences and driving growth through advanced analytics and machine learning. He is recognized as one of the top data and analytics leaders, named to prestigious lists such as CDO Magazine’s 40 Under 40.