In Brief
A chief data officer’s first 90 days should produce more than an understanding of the organization. They should produce a written mandate that the CEO or executive sponsor will sign.
For
Incoming data leaders and the executives appointing them
What the quarter must produce
01 The signed mandate.
02 The priced decision inventory.
03 The delivered proof point.
04 The operating model and governance floor.
Proof in this piece
“Customer data is inconsistent” describes a condition. “The pricing committee delays account decisions while finance and sales reconcile conflicting revenue definitions” describes a business decision, an ownership problem, and a measurable consequence.
Evidence base
Gartner (May 2026, April 2026) · Forrester (April 2026) · MIT Sloan Management Review (2022) · HHS · PCI Security Standards Council · NIST AI Risk Management Framework
A chief data officer’s first 90 days should produce more than an understanding of the organization. They should produce a written mandate that the CEO or executive sponsor will sign. The listening tour, data landscape review, and stakeholder map matter only if they help define what the data function is accountable for, what authority it has, and how its work will be measured.
Most CDO first 90 days plans reverse that logic. They turn the quarter into a curriculum: meet the stakeholders, map the stack, assess the team, and build a roadmap. The result is often a larger backlog, not a clearer mandate.
A backlog records demand. A mandate establishes accountability. Without one, the new CDO can quickly become the head of an executive-level service desk, responsible for many requests but authorized to make few consequential decisions.
This guide shows how to use the first 90 days to identify the business decisions that data should improve, negotiate a one-page mandate, deliver an early proof point, set a governance floor, evaluate the inherited data team, and establish the measures executives will use to judge progress.
The Mandate Gap Facing a New Chief Data Officer
At the Gartner Data & Analytics Summit in London in May 2026 72% of data and analytics leaders were hired to act as organizational change agents, while only 12% were fully prepared to execute that transformation mandate. Gartner also found that 61% had experienced changes to their strategic priorities by executive mandate .
Together these figures describe the conditions in which a new data leader must operate: high expectations, limited readiness, and priorities that can move after the appointment.
The EWSolutions interpretation is practical. A mandate that remains distributed across private conversations is difficult to defend when budgets, leadership, or priorities change. Writing it down forces the unresolved questions into the open: Which business decisions will the data function improve? Who owns those decisions? What authority and resources does the CDO need? What is outside the function’s scope?
The time available to settle those questions may be short. MIT Sloan Management Review reported in 2022
The mandate is also unlikely to be purely technical. Forrester wrote in April 2026
That negotiation should begin in week one.
Why Standard 90-Day Plans Fail Data Leaders
General executive transition advice often follows a simple sequence: learn, diagnose, then act.
For a CDO, that sequence creates four problems.
Use the Quarter to Build a Mandate
Use the quarter as a mandate ledger. Each 30-day block adds an artifact that narrows the function’s scope and makes the next executive conversation more concrete.
This is different from a milestone plan. A milestone can say that interviews are complete. A ledger records what those interviews established: the decisions at risk, their owners, the cost of current workarounds, and the expectations that conflict. By day 90, those entries resolve into a signed mandate, an operating model, and evidence that the function can deliver against both.
01 Days 1-30 Build the Decision Inventory
Do not begin with an exhaustive audit of every data asset and platform. It will identify what is broken without establishing what is worth fixing first.
Run a focused data landscape review instead. For each candidate business decision, identify the systems that supply its data, the people who control access, the team that produces the analysis, and the points where definitions or numbers diverge. The aim is not a complete enterprise map. It is to expose what prevents an executive or operator from making a recurring decision with confidence.
The central artifact for month one is the decision inventory, the instrument EWSolutions uses to open a data management program: a ranked list of recurring business decisions made with incomplete, contested, or late data. Each entry should name the decision, its business owner, the data problem, the current workaround, and the cost or delay created by that workaround.
The distinction is important. “Customer data is inconsistent” describes a condition. “The pricing committee delays account decisions while finance and sales reconcile conflicting revenue definitions” describes a business decision, an ownership problem, and a measurable consequence. The second statement can support a mandate and a funding request.
Use the same four questions in each stakeholder interview:
01 What recurring decision do you make that data should inform?
02 What do you do when the data disagrees with itself?
03 What would you do differently if you trusted the number?
04 What does the current workaround cost in time, delay, rework, or write-offs?
A consistent interview structure makes answers comparable across functions. It also reveals governance problems. When two leaders describe the same measure differently, the issue is no longer an abstract data-quality concern. It is a definitional dispute with two business owners.
Weight the calendar toward two groups: executives who own the decisions and budgets, and operators who handle the failure every day. Include system owners whose cooperation will be required, but do not let the stakeholder map become a reason to postpone analysis.
Two other assessments run alongside the interviews. First, evaluate whether the inherited data team has the capabilities and structure required by the emerging decision inventory. Second, map the political conditions around the role: who sponsored it, whose responsibilities now overlap with it, and where authority remains ambiguous.
By day 30, the CDO has a ranked decision inventory, a focused view of the supporting data landscape, an initial team assessment, and a list of mandate questions that require executive resolution.
02 Days 31-60 Negotiate the Mandate and Deliver a Proof Point
Month two has two parallel tracks.
Draft the mandate. Keep it to one page and write it in business language. It states:
the decisions the data function is accountable for improving;
the outcomes expected from those improvements;
the authority required to deliver them;
the resources attached to the work; and
what is explicitly outside the function’s scope.
The exclusion list is not a formality. If the mandate excludes nothing, every request can be treated as a priority and the CDO remains accountable for demand that was never funded.
Circulate the draft to the sponsor and every executive whose budget, headcount, or decision rights it affects. A marked-up mandate is evidence that the negotiation is working. It is better to expose disagreement on one page in month two than through competing expectations after the annual plan is funded.
Deliver one proof point. Select an item from the decision inventory that meets three conditions:
it can be improved within the quarter using resources already available;
it matters to an executive who owns the affected decision or budget; and
it leaves reusable capability behind, such as an approved definition, named owner, monitored pipeline, or working control.
The third condition separates a proof point from a favor. A favor resolves one request. A proof point demonstrates that the mandate can create a business result and a stronger data foundation at the same time.
03 Days 61-90 Make the Mandate Operational
The final month converts agreement into an operating system. Structure, governance, funding sequence, and review cadence are settled here.
Set the operating model. Decide how accountability will be distributed across the central data function and business units, then publish the reporting lines and decision rights. A centralized, federated, or hub-and-spoke model can work, but an accidental mixture makes ownership difficult to enforce. Gartner reported in May 2026 that 85% of data and analytics practices were not architected for enterprise scale . The choice of operating model is therefore a deliberate design decision, not an inheritance.
Lay a governance floor. EWSolutions treats this as the minimum control set required before a program scales. Establish the controls that would be costly to retrofit later. Keep the scope tied to the data used by the first proof point and the decisions already named in the mandate.
Publish a business-aligned sequence. Show which priced decisions will be addressed in which quarter and connect each one to a strategic priority the executive team has already approved. This is a funding sequence, not a list of technical projects.
Establish the review cadence. Create a monthly or quarterly forum where executives review data outcomes for the decisions they own. The discussion covers changes in cycle time, cost, risk, and adoption, not the percentage of a technical roadmap completed.
Then communicate the mandate, structure, reporting lines, and known capability gaps to the data team. State what has been decided, what remains open, and when the next decisions will be made. Ambiguity that is tolerable in an executive draft becomes damaging when the team must allocate work against it.
Choose Proof Points, Not Quick Wins
Quick-win advice often optimizes for ease. Ease determines whether something can be delivered quickly; it does not determine whether the result will change executive confidence in the data function.
Select for exposure and reuse. The right first initiative addresses a failure that a named executive already experiences, has a measurable baseline, and creates an asset the next initiative can inherit. That asset may be a governed definition, a clear owner, a monitored data flow, or a tested approval process.
Use the decision inventory to rank candidates by expected benefit and delivery effort. Name the executive owner on every line. When two initiatives appear similar, prefer the one with greater business or risk exposure, provided it can still be delivered credibly within the quarter.
Avoid chaining together unrelated quick wins. Responsiveness can earn goodwill, but it can also lock the data function into a service role. Every early delivery demonstrates part of the mandate rather than distracting from negotiating it.
Build Trust With Business Stakeholders
Trust is not a soft workstream beside the technical plan. It determines whether business leaders will accept shared definitions, fund changes to their processes, and support the authority written into the mandate.
Operational trust depends on what the data function reports, how often it reports, and what it declines:
Report in business terms. Lead with the decision improved, the delay reduced, the cost avoided, or the risk controlled. Explain the metadata repository or pipeline only when it helps the reader understand the result.Communicate on a fixed rhythm. Use a short written update between formal reviews. A predictable cadence keeps outcomes, decisions, and unresolved dependencies visible.Name what the function is not doing. Declined or deferred requests make the mandate credible when the reason is clear and tied to agreed priorities.
The CDO sequences relationships by the decision inventory, not by the organization chart. Include the finance partner who will test the business case, the risk or compliance leader who will evaluate exposure, and the operations leaders closest to the failure. Repeated working conversations reveal constraints that a maturity survey may miss and give the CDO relationships to use when priorities conflict.
The CDO also builds data literacy where the definitions will be applied. Business units need to understand what a governed definition means, why a number changed, and who can approve an exception. Gartner’s annual chief data and analytics officer agenda survey found that only 26% of organizations incorporate culture and communication into their data and analytics governance strategy
Establish a Governance Floor, Not a Bureaucracy
New CDOs face two opposing risks.
The first is deferral. Governance is politically difficult, so ownership and approval rules are postponed. If AI or analytics initiatives then depend on poorly governed data, the organization must add controls after design and delivery choices have already been made. Gartner stated in May 2026 that organizations with low data governance maturity are significantly more likely to be among those that fail to realize AI value abandon 60% of AI projects that are unsupported by AI-ready data
The second risk is overbuilding. A new CDO launches a council, publishes a policy library, and requests headcount before delivering a business result. Stakeholders experience governance as new overhead rather than as a way to improve a decision.
The first-quarter governance floor sits between those extremes. Establish:
Named accountability for the critical data elements used by the proof point. A documented rule becomes a control only when someone has authority to maintain and enforce it.A working definition of approval. State what makes a data definition official and who can approve or change it.One documented escalation path. Test it on a real definitional dispute between business units. An untested path fails at the moment it is first needed: no one can say who decides, and the dispute returns to the CDO as an escalation about the escalation.A scoped compliance perimeter. For US organizations, determine which obligations apply by entity, dataset, jurisdiction, and use case. HIPAA applies to covered entities and business associates PCI DSS applies to entities that store, process, or transmit cardholder data NIST AI Risk Management Framework
This floor is intentionally narrow. It applies governance to the decisions already prioritized, proves that ownership can work, and creates a pattern that can be extended. It does not attempt to govern the entire enterprise in one quarter.
Evaluate the Data Team Against the New Mandate
The inherited team was built for the previous roadmap and operating model. Evaluate it against the decisions in the proposed mandate, not against the volume of work it currently processes.
Assess what the team can do, how it is holding up, and how it is organized:
Capability: Does the team have the data management, governance, engineering, and analytics skills required by the prioritized decisions?Morale: Which work has repeatedly been delayed or shelved? Who is already carrying responsibility beyond their formal role? What uncertainty is affecting the team’s ability to deliver? Those answers usually describe the previous operating model more than the people in it: work stalls where no business owner was named, and individuals absorb accountability the structure never assigned.Structure: Do reporting lines allow someone to own a business outcome, or only a queue of technical tasks?
Share the direction with the team before every staffing detail is settled. Explain the mandate, the decisions the function will support, the capabilities required, and the questions that remain open. Stated plainly, that direction prevents people from filling the gaps with unsupported assumptions.
Close capability gaps in the order the mandate requires. Place ownership and expertise where the business decision is made, rather than defaulting to additions in the central team. The operating model should determine the organization chart, not the other way around.
The Day 90 Artifact Set
90
Day 90 — Artifact Set
At day 90, the CDO should be able to place four artifacts on the table:
01 The signed mandate02 The priced decision inventory03 The delivered proof point04 The operating model and governance floor
The signed mandate. One page covering accountabilities, outcomes, authority, resources, and exclusions, signed by the CEO or executive sponsor.
The priced decision inventory. A ranked set of business decisions with named owners, current constraints, and a baseline for cost, delay, rework, or risk.
The delivered proof point. One improvement measured against its baseline, with reusable governance or data-management capability left behind.
The operating model and governance floor. Published reporting lines, decision rights, named data owners, an approval definition, an escalation path, and a business review cadence.
Together, these artifacts show that the CDO has moved beyond onboarding. The function has an agreed purpose, evidence of delivery, and a way to make and review decisions after the first quarter ends.
Metrics the Board Will Read
Define the measures before delivery begins. Otherwise, the team may complete the work without being able to demonstrate what changed.
Cycle time, avoided cost, coverage, and value realization translate data work into executive terms:
Decision cycle time: the elapsed time from a business question to a trusted answer, measured before and after the intervention.Avoided cost: remediation, rework, manual reconciliation, or delay no longer incurred because the underlying data problem was addressed.Coverage: the share of critical data elements supporting funded decisions that have a named owner and current documentation.Value realization rate: the proportion of data initiatives that reach production and demonstrably improve a business outcome.
These measures carry weight only when each claim states its scope, its baseline, and the criterion that defines success. A performance number missing any of the three cannot be evaluated.
The emphasis on financial and operational results matters. Gartner reported in April 2026
Report foundational improvements in the same business context. A governed definition, retired shadow spreadsheet, or accountable owner matters because it reduces friction or exposure for a prioritized decision. Report the capacity it creates, not merely the task completed.
EWSolutions has delivered more than 155 data management programs since 1997 with a 100% client project success rate, for Global 2000 companies and U.S. federal agencies, under the leadership of David Marco, PhD. Our data governance services client case studies
Five Ways the First 90 Days Go Wrong
Most of these failures are failures of sequence. The CDO begins with tools, requests, or structure before securing agreement on the decisions and outcomes that justify them.
The Two-Week Test
The first test of this approach does not require the full quarter. Use the opening two weeks to conduct focused conversations across finance, operations, and a revenue-facing function. Ask the four decision-inventory questions and record the cost of each workaround in time, delay, rework, write-offs, or exposure.
At the end of the test, assess whether you can name a recurring decision, its executive owner, the data constraint, and a measurable consequence. If you can, you have the beginning of a fundable mandate. If you cannot, do not fill the gap with a platform review or generic maturity program. Determine whether the interviews reached the people who own material decisions and whether the organization can articulate what it expects the data function to change.
Either result is more useful than a stack of unstructured notes. Discovery earns its place in the first 90 days when it produces evidence for a decision.
From First Quarter to Written Authority
The first 90 days do not need to produce a complete data strategy. They need to establish the authority, evidence, and operating conditions from which a credible strategy can be funded.
A signed mandate gives the roadmap a boundary. The decision inventory gives it an economic basis. The proof point demonstrates that the function can improve an outcome without abandoning foundational data management. The operating model and governance floor make that progress repeatable.
The mandate will be shaped during the first quarter whether or not the CDO writes it down. Stakeholders will otherwise define it through requests, budget decisions, and inherited responsibilities. The better course is to put the terms on one page, expose the disagreements, and ask the sponsor to decide.
Incoming data leaders and the executives appointing them can request an Executive Briefing David Marco, PhD
