How CIOs can negotiate data, AI and politics while avoiding snake oil

“Data is the new oil.”

“Data is our most valuable asset.”

“AI will unlock the value of your data.”

Executives hear these refrains constantly, but the economics behind them often remain murky. A key benefit of data analytics is improved decision-making and, if we focus on the logic of improved decision-making, the essential question becomes “How can we measure the value (ROI) on investments in data infrastructure?” Knowing when your data starts delivering real value and how to avoid the vendor hype are part of the equation.

What we need is a disciplined framework for thinking about the economics of data — how to size investments, measure returns and avoid snake oil.

Infrastructure vs. asset: The two layers of data economics

One reason data investment feels fuzzy and hard to measure is that two very different layers often get blurred:

• Data infrastructure: pipelines, storage, governance tools, compute. These are the highways of your digital economy. You can’t operate without them, but by themselves, they don’t generate ROI.
• Data as an asset: the information itself. Properly managed, data can appreciate over time: cleaned, contextualized and combined with other sources, data powers more informed decision-making that can be used to increase revenues and reduce costs. Yet for data to be an asset, it must be actively managed; without governance, data quickly decays, which can impact its economic relevance.

CIOs need to make both layers work together — but not confuse infrastructure spend with asset value.

How much should companies spend?

Benchmarks vary, but most organizations overspend chasing big platforms or underspend by treating analytics as an afterthought. A grounded view is this:

• Overall IT budgets: commonly 2% to 6% of revenue, depending on industry.
• Analytics and AI: usually a small slice of IT — often 1% to 3% of the IT budget, rising in data-intensive sectors like finance or tech.

The key is not the total but the marginal dollar. If you already have reporting teams, ERP systems and cloud services, ask: What measurable improvement do we get from the next $100K invested in analytics?

Opportunity cost matters, too. If competitors deploy predictive models that improve pricing or customer retention, what value do you lose by standing still?

The hidden cost of poor data quality

Before calculating ROI, CIOs must account for the “tax” of bad data. This is not optional — it’s a recurring cost of doing business.

• Gartner estimates poor data quality costs organizations an average of $12.9 million annually in wasted effort and missed opportunities.
• A separate study reported on Anodot found employees spend up to 27% of their time wrangling data rather than using it productively.

Ignoring governance means you’ve built highways but are constantly paying fines and repair crews just to drive on them. Treating data quality as foundational unlocks every downstream use case.

The politics of ownership

Here’s the uncomfortable truth: CIOs own the budget, but they don’t own all the data.

• Data generation lives in business units: marketing generates leads, operations collects sensor data, finance owns transactional records.
• Data quality depends on frontline discipline: if sales doesn’t enter clean data, no amount of IT spend will save it.
• Data value extraction is distributed: product, HR, compliance and strategy teams all rely on the same data to make decisions.

That means every conversation about data ROI is also a conversation about politics and incentives.

• If marketing wants speed, compliance wants safety and IT wants standards, who decides?
• When the CIO has the checkbook, but other leaders control the inputs and outcomes, accountability is blurred.

The best CIOs do not just build infrastructure; they broker agreements. They create an environment of shared ownership with governance councils, shared metrics and escalation paths. They remind peers that data is a team sport — but like any sport, someone has to referee.

Measuring value: Case by case

Attempts to assign a universal “return on data assets” collapse under assumptions. A pragmatic approach measures value in context. Four categories frame it:

1. Revenue impact: Did targeting or segmentation lift sales?
2. Cost reduction: Did predictive maintenance lower OpEx?
3. Risk mitigation: Did compliance monitoring avoid fines?
4. Decision velocity: Are leaders making faster, better-informed calls?

Evidence is growing. Predictive maintenance programs cut costs by 10% to 40% and reduce downtime by up to 50%. Siemens reported an ROI of 250% from scaled predictive equipment maintenance. These are measurable, defensible returns — not abstract valuations.

Mini-formula for CIOs:
ROI = [(Δ revenue + Δ cost savings + Δ risk avoided) – (incremental spend + data quality tax)] ÷ time to impact

From ROI to return on data: A finance lens

Traditional ROI formulas are helpful, but CIOs and other executives need a structured way to explain data’s value in finance terms. One way is to adapt the DuPont model — a century-old method CFOs use to break down return on equity.

Applied to data, the model looks like this:

Return on Data (RoD) = Profitability × Efficiency × Leverage

• Profitability: How much does data move the P&L? (Revenue lift, cost savings or risk avoidance as a % of spend.)
• Efficiency: How effectively are data assets used? (% of curated datasets actually deployed in decisions, cycle time to insight.)
• Leverage: How well do governance, infrastructure and culture scale that usage across the enterprise?

This aligns directly with other methods of value mapping by looking at various dimensions:

• Revenue growth (profitability)
• Cost efficiency (profitability again)
• Asset efficiency (efficiency)
• Risk reduction (profitability + governance leverage)

In practice: A CIO can show that predictive maintenance improved margins (profitability), cut downtime by 30% (efficiency) and scaled across three business units due to governance (leverage). That’s a holistic, finance-friendly case for return on data.

Avoiding snake oil

The marketplace for data and AI products and services is crowded and many vendors lead with inflated promises. When evaluating pitches for new capabilities, four red flags are useful triggers to check and, if tripped, should set off alarms:

• Universal valuations: No one can slap a single dollar figure on your data. Value is contextual.
• Infrastructure-first pitches: Expensive lakes or fabrics with hand-wavy ROI.
• One-size-fits-all ROI models: What’s priceless to Amazon may be irrelevant to a regional distributor.
• Model hype over data hygiene: If the pitch is algorithms-first and dataset-second, stop and ask hard questions.

Snake oil stress test — Questions every CIO should ask vendors

1. What baseline are you comparing against?
2. Can you show a counterfactual (what happens if we don’t deploy)?
3. What’s the total cost of data quality, governance and retraining?
4. How do you measure model drift and maintenance costs in years 2–3?
5. Can you show published ROI beyond one pilot?
6. Who owns the risk if privacy/regulation changes?

If vendors can’t answer, walk away.

A practical playbook for CIOs

A playbook can help provide a rubric for disciplined decision-making for CIOs and other executives responsible for data economics. The playbook is intentionally streamlined with five straightforward rules:

1. Link spend to outcomes: No KPI, no project.
2. Think incrementally: Evaluate the next dollar, not the whole budget.
3. Pilot, measure, scale: Prove value small, then expand.
4. Make quality non-negotiable: Governance and cleanup come first.
5. Build culture, not just infrastructure: A literate workforce creates more value than any tool.
6. Kill failed projects fast: Don’t let sunk-cost bias drain your runway.

And always remember: the politics are real. Data ownership is distributed, but accountability sits in the CIO’s office. Success depends as much on coalition-building as it does on technology choices.

Owning the narrative vs. being taken for a ride

The economics of data are not mysterious, but they do demand rigor. CIOs don’t need vast budgets or miracle platforms. They need discipline: separate infrastructure from assets, budget for quality, measure value on a case-by-case basis and insist on accountability in ROI.

But discipline alone isn’t enough. Because data responsibility is distributed, the CIO must also navigate politics, broker agreements and rally other leaders to a shared standard.

Data may be “the new oil.” But unmanaged — technically or politically — it’s just a slick sales pitch. The CIO who can show dollar-for-dollar outcomes and align stakeholders behind them will own the narrative — and the competitive advantage.

This article is published as part of the Foundry Expert Contributor Network.
Want to join?