Sales technology companies often have plenty of marketing reporting and too few decisions. A weekly dashboard may show traffic, spend, form activity, and pipeline fields, yet the meeting still ends with “keep monitoring.” The problem is usually not the absence of data. It is the absence of a review record that separates observation from interpretation and assigns a reversible next move.
This template is built for a marketing performance review with a defined audience, period, and decision owner. It does not prescribe a target conversion rate, prove attribution, or turn an analytics event into revenue. It gives a sales-technology team a shared way to answer: what changed, what may explain it, what is still unknown, and what should happen next.
Start with the decision, not the dashboard
Write one sentence before opening the report:
> Given the evidence available for [period and scope], should we continue, change, pause, or investigate [specific marketing motion]?
Name the motion precisely: a campaign family, content route, partner motion, lifecycle stage, or audience. “Review marketing performance” is a meeting title. “Decide whether the integration-partner motion receives another bounded test after its lead-quality review” is a decision.
Set the review boundary:
- Period: the dates and any delayed data expected later.
- Scope: channels, programs, pages, or lifecycle stages included.
- Decision owner: one person who can authorize the next action.
- Evidence cut: the latest reliable extraction and known gaps.
- Exclusions: records that should not be compared because definitions differ.
- Decision deadline: when the action must be chosen or explicitly deferred.
Without this boundary, a review can expand until every metric is presented and none is owned.
The decision template
Use the fields below in order. A blank is not a failure; it is an evidence gap that must be visible.
| Field | Completion guidance | Minimum evidence | |—|—|—| | Decision in scope | State the action that may change | Named motion and owner | | Intended outcome | Describe the business condition, not a vanity metric | Observable outcome and time boundary | | Audience and stage | Identify who the motion should help and where | Definition used in the plan | | Change observed | State what changed without explaining it yet | Source, extraction date, and comparison rule | | Confidence | High, medium, or low with a reason | Data completeness and definition check | | Candidate explanations | List plausible causes, including alternatives | Evidence for and against each | | Quality or fit signal | Describe whether the response is useful | CRM or review evidence, not assumed quality | | Constraint | Name capacity, permission, product, or route limits | Accountable owner and boundary | | Next reversible action | Define the smallest useful move | Owner, due point, and success evidence | | Stop or escalation rule | State when not to continue | Explicit trigger and approver | | Review state | Continue, change, pause, investigate, or close | Decision owner’s recorded choice |
The fields deliberately prevent a raw metric from becoming a recommendation. A click count can be an observation. It cannot, by itself, explain fit, urgency, or sales readiness.
Field definitions that keep the review honest
### Change observed
Write the observation in neutral language: “The share of sessions from the named account group changed in the selected period.” Do not write “the new positioning worked” until a design or evidence check supports that inference.
For website data, Google Search Console’s Performance report provides first-party query, page, click, and impression context. It should not be used to infer a visitor’s identity or private buying stage. For product or website interactions, GA4 Events document events and parameters; an event remains an observable action, not a qualified opportunity.
### Confidence
Use a short explanation:
- High: the definition is stable, the source is complete enough for this decision, and the change is reproduced by a second check.
- Medium: the signal is plausible, but one material gap remains.
- Low: definitions, timing, permission, or data quality make the interpretation unsafe.
Do not use confidence as a score for the team. It describes how far the review can responsibly go.
### Candidate explanations
Require at least one alternative explanation. A fall in qualified conversations may reflect a message problem, a routing delay, a CRM definition change, a seasonal audience change, or a reporting lag. The review should identify which checks can discriminate among them.
The NIST Information Quality Standards are a useful reminder to consider utility, integrity, and objectivity. They are a quality lens, not a certification of a sales-technology dataset.
### Intended outcome
Describe the condition the decision is meant to improve: a qualified conversation, a usable account signal, a completed handoff, or a better-supported buying decision. If the organization has not defined what “qualified” means, record that as a prerequisite rather than borrowing a number from another team.
When an event is important to the business, GA4’s conversion documentation explains how an event can be marked and used consistently across Analytics and Google Ads. That implementation fact does not prove incrementality, profit, or sales acceptance. Keep those outcomes separate in the review.
Review sequence
Run the meeting in five passes.
- Boundary check: confirm period, scope, definitions, and late data.
- Observation pass: read the change fields without recommendations.
- Explanation pass: compare candidate causes and evidence gaps.
- Decision pass: select one reversible action, an owner, and a stop rule.
- Record pass: capture the decision, dissent, corrections, and next review trigger.
This order matters. If the team starts with recommendations, the dashboard tends to become evidence for an already chosen action.
Illustrative worked example
The following is hypothetical. It is a demonstration of the template, not a benchmark or a customer result.
| Field | Example completion | |—|—| | Decision in scope | Decide whether to extend an integration-partner content motion or revise its audience boundary | | Intended outcome | Generate more partner-referred conversations that match the agreed sales-technology account definition | | Audience and stage | Operations leaders evaluating a new sales workflow; exclude general productivity research | | Change observed | Partner-referred sessions and named-page interactions were recorded; lead acceptance fields are incomplete | | Confidence | Medium: route is identifiable, but CRM acceptance timing is not reconciled with web timestamps | | Candidate explanations | Better partner context; broader curiosity; routing lag; page attracts users outside the intended account set | | Quality or fit signal | Two reviewed conversations mention the integration problem; no conclusion about repeatability yet | | Constraint | One subject-matter owner and limited partner review capacity | | Next reversible action | Run a small revised brief with an explicit account boundary and reconcile handoff records before expansion | | Stop or escalation rule | Hold if permission, account definition, or handoff timestamp remains unresolved | | Review state | Investigate |
Notice what the example does not say. It does not claim the partner route caused pipeline, assign a conversion benchmark, or hide the missing CRM field behind a confident recommendation.
Decision types and their proper use
| State | Meaning | What must be recorded | |—|—|—| | Continue | Evidence supports another bounded cycle | What remains unproven and why the cycle is reversible | | Change | The motion may work, but a specified variable needs correction | Variable, owner, and comparison rule | | Pause | Capacity, permission, or risk makes more activity unsafe | Hold owner and restart condition | | Investigate | A material explanation remains unresolved | Evidence request and decision date | | Close | The motion no longer fits the stated decision | Reason, archive state, and learnings |
Do not combine these into a green/yellow/red score. A motion can be promising and still require a pause because a permission boundary is unresolved.
Common template failures
| Failure | What it hides | Correction | |—|—|—| | Copying every dashboard tile into the review | The decision boundary | Keep only fields that can change the action | | Using “performance” as a synonym for volume | Fit and usefulness | Define the intended outcome and quality evidence | | Treating attribution as a fact | Model assumptions and missing paths | Label the model and list alternative explanations | | Reusing a stage definition without checking ownership | Conflicting operations | Name the source of truth and who can change it | | Choosing a large action before a small diagnostic | Reversibility | Make the next move the smallest action that can teach something | | Omitting a stop rule | Endless activity | State the evidence or risk that ends the cycle |
If the review contains claims about endorsements, testimonials, or product outcomes, keep them supportable. The FTC advertising and marketing guidance is a U.S. reference for truthful, non-deceptive, evidence-based advertising; it is not a substitute for jurisdiction-specific legal review.
Copy-ready decision template
“text Review date: Decision owner: Period and scope: Motion under review: Decision in scope: Intended outcome: Audience and lifecycle stage: Definitions and exclusions: Change observed (source and extraction date): Data gaps or timing caveats: Confidence and reason: Candidate explanation 1 + evidence: Candidate explanation 2 + evidence: Alternative explanation: Quality / fit signal: Constraint or permission boundary: Next reversible action: Owner and due point: Evidence that would support continuation: Stop / escalation rule: Review state: Correction log reference: Next review trigger: “
The template is complete when a person who missed the meeting can reconstruct the decision without guessing which metric mattered. That traceability is the operating value of the review.
Sources and limits
This template uses Google Search Console Performance, GA4 Events, GA4 Conversions, NIST Information Quality Standards, and FTC Advertising and Marketing as process references. It does not establish a sales benchmark, attribution truth, legal conclusion, or guaranteed commercial result.
How did this article land?
Choose one reaction. You can change it anytime.
