Many Research projects end with a Final Presentation.
There is an Executive Summary.
There are Charts.
There are Findings.
There are Recommendations.
Then the real decision happens in the next meeting.
That is often where the gap appears: everyone sees the same Evidence but applies different decision criteria.

Marketing may say: “The score is strong enough to launch.”
Finance may say: “The Margin still does not work.”
Sales may say: “Customers like it, but we do not yet know why they would buy.”

The problem is not necessarily a weak Research Report.
The problem is that the Evidence has not yet been connected to a clear Decision Logic.

ESOMAR advises users of Research to begin by asking what impact the evidence will have on the final decision, reinforcing the principle that Research should start from the Decision Need rather than collecting information first and deciding what to do with it later.

A Report explains the Evidence. A Decision Rule explains how that Evidence changes the Decision.
Research Reports remain essential because they document the Method, Evidence, Findings, Limitations and Interpretation. But when Research exists to support a Business Decision, the final output should ideally go beyond statements such as “consider proceeding” or “the result is generally positive.”
‍
A Decision Rule defines the conditions under which Evidence changes an Action. For example, a Concept Test might specify Go if Purchase Intent, Relevance and Value all clear agreed thresholds, but Revise if Appeal is strong while Value Perception remains weak.
Decision Thresholds are used in Evidence-to-Decision frameworks to make explicit the point at which judgment changes and to focus interpretation on effects that matter for decisions rather than on Statistical Significance alone.
A practical business sequence is: Business Decision → Uncertainty → Evidence → Decision Rule → Action

Reports and Decision Rules do different jobs

A Research Report answers:
How was the evidence collected?
What did we find?
What patterns appeared?
What are the limitations?
How should the findings be interpreted?

A Decision Rule answers:
If the Evidence looks like this, what will we do?
If it falls below this level, what changes?
If the result remains uncertain, do we gather more Evidence or stop?

A Decision Rule should therefore not replace the Report.
Decision-oriented Research needs both.
Report = Evidence Record
Decision Rule = Evidence-to-Action Logic

A broad Recommendation is not yet a Decision Rule

Statements such as:
“Continue developing the product.”
“Focus on Segment A.”
“Improve the communication.”
are still Recommendations.

They do not specify:
What Evidence is required to continue investing?
What level counts as sufficient?
What happens if a criterion fails?
Who owns the final Decision?

A more explicit Rule might look like:
If X ≥ Threshold A and no Critical Risk exists → Go
If X passes but Y fails → Revise
If a Critical Assumption fails → Stop
If Evidence is insufficient → Test Further

Research on Go / No-go criteria in New Product Development argues that decision criteria should be formulated before project review and matched to the information, cost and timing available at each decision stage.

Why define the Rule before seeing the result?

Suppose a Concept Test produces: Purchase Intent = 63%
Without a predefined threshold, the discussion may become:
Is 63% good?
What did competitors score?
Would it look stronger with a different Base?
Would it pass if we removed a less relevant Segment?
Interpretation can begin moving with the Outcome.

A predefined Decision Rule reduces this flexibility by making clear in advance what type of Evidence would change the Decision.
Decision Threshold literature makes the same point: explicit thresholds improve transparency and consistency compared with applying implicit criteria after seeing results.

A Threshold should not exist simply because a number looks scientific

Another mistake is creating a rule such as:
“Above 70% = Go” without knowing why 70% matters.

A useful Threshold should connect to Business Meaning.
For example:
Minimum Conversion required for viable Unit Economics
Minimum Retention required to support Customer Acquisition Cost
Minimum Demand required to justify Fixed Investment
Maximum Defect Rate Operations can tolerate
Minimum Customer Acceptance management requires before investing further
Decision Threshold frameworks emphasize effects that are meaningful for decisions rather than relying only on Statistical Significance.

Therefore: Statistically Significant ≠ Commercially Important
and: High Score ≠ Decision-ready automatically

A useful Decision Rule usually needs more than one Metric

Many Business Decisions should not be reduced to one KPI.
A New Product decision may need to consider:
Customer Acceptance
Market Opportunity
Strategic Fit
Technical Feasibility
Financial Performance
Research on Go / No-go criteria in Product Development found that firms use multiple dimensions and that their relative importance changes across development stages.

A Concept with high Purchase Intent may still be a poor Go decision if:
Production Cost is unworkable
The Target Market is too small
No viable Channel exists
The Product cannot reliably deliver its Promise
Decision Rules should reflect critical assumptions in the Business Model, not only the Metrics easiest for Research to measure.

Example: A Concept Test should not end with “Concept B wins”

Suppose Research compares Concept A with Concept B.

Concept B produces higher Overall Appeal.
A Report may conclude: “Concept B is the preferred Concept.”

The Decision needs more.
For example:
Go If Concept B clears minimum thresholds for Need, Relevance, Purchase Intent and Value with no Critical Barrier in Qualitative Feedback.
Revise If Appeal and Need are strong but Value or a fixable Barrier remains weak.
Test Further If Segment Differences are substantial and current Evidence is insufficient for the Decision.
Stop If the Core Need is not confirmed or the economics require Demand beyond what the Evidence supports.

Now the Research does more than show that B beats A.
It states under what conditions B deserves further investment.

Example: Customer Experience Research should not end with “Satisfaction declined”

Suppose Customer Satisfaction falls from 82 to 77.
The Report can show:
The decline
Affected Segments
Problematic Touchpoints
Customer Feedback themes

A Decision Rule adds:
If Satisfaction declines but Retention remains stable → investigate Drivers before making a major investment.
If Satisfaction, Complaints and Repeat Purchase deteriorate within the same Journey → prioritize investigation.
If a Process Failure affects a large number of customers → fix the Process.
If the decline exists only in a small Segment → consider a targeted response rather than a company-wide change.
This prevents every Score Movement from automatically triggering a large Action.

Go does not have to mean “full-scale Go”

Decisions are not limited to: Go orNo-go
‍
Useful intermediate actions include:
Go → Scale
Conditional Go → Pilot with a Budget Cap
Revise → Change the Proposition and retest
Test Further → Resolve a Critical Unknown
Stop → Do not invest further under the current assumptions

This is particularly useful for SMEs because Research often does not need to justify a full investment.
It may only need to justify the next bounded commitment.
Decision Rules should therefore align the size of the commitment with the strength of the Evidence.

When Evidence is insufficient, the Rule should say what needs to be learned next

Good Research does not always produce certainty.
Sometimes the correct conclusion is: UNKNOWN

For example:
Demand looks promising, but an important Target Segment remains underrepresented.
Customers like the Concept, but Willingness to Pay is unclear.
Trial is strong, but Repeat Purchase has not yet been observed.

In these cases, the Rule should clarify:
What remains Unknown?
Could that Unknown change the Decision?
What minimum Evidence would reduce the uncertainty?
How much is it worth spending to learn more?

Research on matching Study Designs to Decision-maker questions reinforces that Research Design should correspond to the specific question that needs to be answered. More data is useful only when it reduces uncertainty that matters to the Decision.

Do not turn the Decision Rule into an automatic formula

A Decision Rule should not become an Algorithm that replaces judgment.

Important factors may sit outside a single Metric:
Regulatory Risk
Brand Risk
Strategic Timing
Operational Constraints
Ethical Concerns
Data Quality
Unexpected External Events

A useful Rule makes assumptions more explicit. It does not eliminate judgment.
The final Decision may therefore consider:
Quantitative Thresholds
Qualitative Evidence
Critical Risks
Management Judgment
Limitations

Use Evidence Labels before applying the Decision Rule

Before making the Decision, separate:
FACT What the Evidence directly supports
BEE INTERPRETATION A reasonable meaning derived from the Evidence
HYPOTHESIS An explanation or expected effect that still needs testing
UNKNOWN What the current Evidence cannot answer

Example:
FACT: 68% of the Target Sample selected Concept A over the Current Product under this Test Design.
BEE INTERPRETATION: Concept A shows encouraging Acceptance within the studied population.
HYPOTHESIS: Concept A may increase Trial if launched with suitable Pricing and Channel support.
UNKNOWN: Actual Purchase Rate after launch has not been established.
The Decision Rule can then determine whether this Evidence supports a Pilot, Scale-up or Additional Test.

Write the Decision Rule into the Research Brief

Do not wait until the Final Presentation to ask: “So how do we use this result?”
A decision-oriented Research Brief should include:
Business Decision
Decision Owner
Critical Uncertainty
Research Question
Critical Metrics
Decision Threshold / Criteria
Possible Actions
Key Risks
Acceptable Unknowns

Example:
Business Decision: Should we invest in a Product Pilot?
Critical Uncertainty Are Customer Need and Willingness to Try strong enough?
Decision Rule Go to Pilot if Need and Trial Intent clear the agreed thresholds with no Critical Barrier related to Usage or Price.
Revise if Need is confirmed but the Value Proposition remains weak.
Stop if the Core Need is not supported.
When the Rule is designed before the Method, the Research team knows what Evidence is needed and does not need to build a long Questionnaire merely to “cover everything.”

Framework: Make Research end with a Decision, not a Presentation

1. Define the Business Decision
   What exactly needs to be decided?
2. Identify Decision Uncertainty
   What is still unknown that could change the Decision?
3. Write the Hypothesis
   What would need to be true for the Decision to make sense?
4. Define the Evidence Needed
   What Evidence would support or challenge that Hypothesis?
5. Set Decision Criteria
   What would trigger Go, Revise, Test Further or Stop?
6. Design the Research
   Choose the Method and Sample to answer those criteria.
7. Separate Fact from Interpretation
   Do not jump directly from Finding to Recommendation.
8. Apply the Decision Rule
   Identify which conditions the Evidence satisfies.
9. Record Exceptions and Unknowns
   If management decides differently from the Rule, document why.
10. Define the Next Action
    Scale, Pilot, Revise, Research Further or Stop.

This follows the BEE sequence: Business Decision → Decision Uncertainty → Research Question → Hypothesis → Evidence Needed → Method → Analysis → Interpretation → Action Rule → Decision

The takeaway: The Report should preserve the Evidence, but the Research Process should reach a Decision Rule

The answer is not that Research should end with either a Report or a Decision Rule.
The Report remains essential.

A business needs to know:
Where the Evidence came from
How the study was conducted
What the findings show
Where the limitations sit

But if Research exists to support a Decision, it should go one step further: Under what conditions should this Evidence change the Action?
Decision Thresholds are designed to make the point at which judgment changes more explicit and to keep evidence interpretation focused on meaningful effects rather than Statistical Significance or post-result intuition alone.

A useful distinction is:
Report → What did we learn?
Decision Rule → What will we do if this is true?

Good Research therefore should not end with the final slide saying: “Thank You.”
It should leave the decision-makers able to say: “Based on this Evidence, we will Go, Revise, Test Further or Stop—and this is why.”