How to Run a Cross-Functional Demand Review With Marketing, Sales, Product and Finance

A 60-minute demand-review system that reconciles marketing, sales, product and finance evidence and ends with recorded decisions, owners and learning needs.

Documented fact: Google documents that report comparisons can differ when dimensions, identities and processing choices are incompatible; reconciliation requires understanding what each dataset represents.

CDM inference: The Reconciled Demand Review Pack and 60-minute agenda are CDM operating-system proposals, not universal governance standards.

Define the forum by what it decides

A demand review is a recurring cross-functional meeting that decides how the organisation will respond to changes in customer demand, pipeline, product use, capacity and commercial outcomes. It is not a longer marketing dashboard presentation.

Publish a short charter: decisions in scope, permanent participants, evidence owner, cadence and escalation path. Typical decisions include whether to change the market thesis, reallocate discretionary investment, resolve a capacity constraint, repair a data definition or commission customer research.

Give one executive the final decision right while preserving specialist accountability. Finance should not approve product claims; marketing should not redefine recognised revenue; sales should not unilaterally change what qualifies as demand. The forum connects these systems without erasing their controls.

Build one reconciled evidence pack

Send the pack before the meeting and make every metric traceable to a definition, owner, source, time zone and refresh date. Put comparable periods together. Distinguish leading indicators from outcomes and observed values from modelled estimates.

Do not require every system to produce the same number. Require each number to explain its population and use. Google’s guidance on data compatibility in Google Analytics illustrates why metrics cannot always be combined across incompatible dimensions and scopes. The wider lesson is that reconciliation is a mapping exercise, not a search for one magical dashboard.

Use a 60-minute decision agenda

1. 0–5 minutes: control check. Confirm data refreshes, definition changes and unresolved corrections.
2. 5–15 minutes: material changes. Review only variances above agreed thresholds and the assumptions they affect.
3. 15–30 minutes: constraint diagnosis. Compare market, marketing, sales, product and finance explanations.
4. 30–48 minutes: decisions. Choose actions, stops, tests or escalations and state the evidence basis.
5. 48–55 minutes: learning plan. Assign missing evidence, customer research or instrumentation.
6. 55–60 minutes: playback. Read back decisions, owners, dates and dissent.

Do not spend the meeting discovering what a chart means. If a definition dispute cannot be resolved beforehand, show it as an explicit issue with its effect on the decision. A known uncertainty is safer than a quietly blended number.

Rotate deep dives according to the decision calendar. A pricing change needs a different evidence module from an AI-discovery review or a distribution expansion. Keep the core pack stable so changes remain visible.

Record the decision and its evidence

For every decision, capture the question, options considered, chosen action, owner, evidence used, dissent, expected sequence, guardrails and next review date. Link to the source pack version. This creates an approval trail and allows the team to evaluate judgement fairly later.

Separate three statuses: decided, where an authorised owner has chosen; assigned, where evidence work has an owner but no decision exists; and escalated, where the forum lacks authority. Calling an action item a decision creates ambiguity.

At the next meeting, review whether the predicted leading signal appeared. Do not score a decision only by the final outcome; implementation and external changes matter. The purpose is better repeatable judgement, not retrospective blame.

Avoid the common failure modes

Departmental theatre

Every function presents success in its own language and no trade-off is resolved. Replace status slides with a shared variance and decision format.

Metric coercion

One function forces all systems to adopt its definition. Preserve legitimate differences and build documented bridges between them.

Activity substitution

The meeting ends with more campaigns, calls or features but no statement of the diagnosed constraint. Require every action to name the mechanism it addresses.

Consensus without authority

Participants agree in principle, but nobody can release budget or change a dependency. Put decision rights and escalation routes in the charter.

Limits and risks

A meeting cannot compensate for absent customer evidence, unreliable systems or incentives that reward conflicting outcomes. Start with a narrow portfolio if the data landscape is immature.

The forum may expose sensitive employment, customer or financial information. Apply access controls and data minimisation. Legal, accounting and security decisions remain with qualified owners.

CDM verdict

A cross-functional demand review is valuable when it produces an auditable choice from reconciled evidence. Standardise the pack, preserve differences in meaning, focus the hour on material variances and record who decided what. Alignment is not identical dashboards; it is coordinated action despite known uncertainty.

Related guides

• When Buyers Delay Decisions: How Should Marketing Adapt to Longer Sales Cycles?
• How to Build a 13-Week Rolling Marketing Plan When Demand Is Volatile

Frequently asked questions