01 / EXECUTIVE SUMMARY
A finished handoff, with invented data
Custom study format example
ILLUSTRATIVE ONLY. Every number, concept label and quotation in this package is invented. No human or model study was run.
This example shows how a custom design screen can move from a decision brief to a readable report and a checkable data appendix. The four neutral concepts have no associated creatives, brands or client context.
Example decision
In this constructed dataset, Concept B has the strongest task fit and exploration scores. Concept A is easiest to understand. A reasonable example next step is to carry B forward for revision while using A as a clarity reference. This is a demonstration of reasoning, not a recommendation about a real concept.
| Concept | Records | Understanding top two | Task fit top two | Explore top two |
|---|---|---|---|---|
| Concept A | 20 | 70% | 30% | 40% |
| Concept B | 20 | 50% | 70% | 70% |
| Concept C | 20 | 20% | 30% | 20% |
| Concept D | 20 | 30% | 30% | 20% |
Top two = coded responses of 4 or 5 on a 1-5 scale. Base: 20 invented records per concept.
What the package lets you inspect
A seven-section narrative, diagnostic tables, qualitative examples, subgroup reporting and explicit limitations. The companion JSON contains the records behind every numerical table; the codebook explains each field and calculation.
No sample claim should be quoted as evidence about a product, audience or model.
02 / ANALYST BRIEF
What an analyst would explain
Separate the read from the decision
1. Read the pattern across measures
Concept A leads on ease of understanding, while B leads on task fit and interest in exploring. The useful story is the trade-off across measures. A single composite score would hide it. No composite score is calculated here.
2. Make the proposed next step conditional
For this fictional decision, retain B for another design iteration and borrow the clarity of A. A real recommendation would also depend on the agreed business objective, the actual materials and any constraints documented at scoping.
3. Keep observation separate from explanation
The tables are computed from invented coded records. The text examples on page 5 illustrate how supporting explanation may be presented, but they were written independently and cannot explain a numerical difference in the data.
4. Decide what the next study must resolve
A follow-up brief could ask whether a clearer first step improves understanding without reducing fit. That brief would need an actual design, an audience definition and a measurement plan before any study is run.
Example handoff note: keep the reasoning editable and the tables reproducible. Agree the decision rule before interpreting new results.
The sample illustrates an analyst handoff. It does not promise that every project includes the same analyses or files; deliverables are agreed in scope.
03 / MATERIALS & DESIGN
Show exactly what was compared
Neutral labels in place of study materials
No creatives are included in this public example. The register below demonstrates how a report can identify its comparison cells without displaying private source materials. These labels do not refer to any real design.
| Register ID | Public label | Materials status |
|---|---|---|
| concept_a | Concept A | No stimulus; invented example |
| concept_b | Concept B | No stimulus; invented example |
| concept_c | Concept C | No stimulus; invented example |
| concept_d | Concept D | No stimulus; invented example |
Example design
The data file has four constructed groups of 20 records. Each record has one concept label and three scale values. A real monadic design would assign one concept per participant; here, records were simply authored to make the calculations inspectable.
| Measure | Illustrative question |
|---|---|
| Understanding | How easy is this concept to understand? |
| Task fit | How well does this concept fit the task described? |
| Exploration | How interested would you be in exploring this concept further? |
These questions are newly written examples. No task or stimulus was administered. The codebook defines each scale.
What belongs in a real materials register
The agreed stimulus versions, exposure rules, question wording and analysis definitions should be clear enough for the recipient to identify what was compared. Any material sharing is agreed in scope.
04 / DIAGNOSTICS
Keep the measures visible
Means, top-two shares and response distributions
| Concept | Understanding mean / top two | Task fit mean / top two | Explore mean / top two |
|---|---|---|---|
| Concept A | 3.80 / 70% | 3.00 / 30% | 3.10 / 40% |
| Concept B | 3.40 / 50% | 3.90 / 70% | 3.80 / 70% |
| Concept C | 2.80 / 20% | 3.00 / 30% | 2.80 / 20% |
| Concept D | 3.20 / 30% | 3.20 / 30% | 3.10 / 20% |
Base: 20 invented records per concept. Means use the 1-5 codes; top two counts codes 4 and 5.
Interest in exploring: top-two shares
Understanding: full distribution
| Concept | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|
| Concept A | 0 | 2 | 4 | 10 | 4 |
| Concept B | 0 | 2 | 8 | 10 | 0 |
| Concept C | 0 | 8 | 8 | 4 | 0 |
| Concept D | 0 | 2 | 12 | 6 | 0 |
Cells are counts, not percentages. Each row sums to 20. No weighting, calibration or statistical significance testing is applied.
The full distribution helps a recipient see how a summary was formed. In this constructed example, a high top-two share does not establish real demand, conversion, sales or predicted behavior.
05 / QUALITATIVE EXAMPLES
Show the explanation layer
Invented text examples, never respondent quotations
All eight excerpts below were written for this sample. They are not verbatim responses, not model outputs and not linked to the 80 coded records.
Concept A
Clear starting point: "I can see the first step without having to search for it."
Missing detail: "I would want a short explanation of what happens after that step."
Concept B
Relevant task: "The idea fits the task I had in mind."
Next step: "A small example would help me decide whether to explore further."
Concept C
Unclear hierarchy: "Several parts ask for my attention at once."
Useful detail: "The extra detail could help after I understand the basic idea."
Concept D
Familiar approach: "The layout seems familiar, but I cannot see the main difference yet."
Comparison needed: "I would like a short comparison before making a choice."
This layout illustrates how a report can pair a concise theme with readable supporting text. A real report should identify the source of excerpts and distinguish analyst synthesis from verbatim material.
No theme frequencies are shown. The sample does not support prevalence claims, causal explanations or generalization.
06 / SUBGROUPS
Make the denominator obvious
Arbitrary scenario groups for format demonstration
Scenario 1 and Scenario 2 are arbitrary labels, not demographics or audience segments. Each concept has 10 invented records in each scenario. The distributions are intentionally equal across scenarios, so this page demonstrates layout without manufacturing a subgroup story.
| Concept / group | Base | Understanding top two | Task fit top two | Explore top two |
|---|---|---|---|---|
| A / Scenario 1 | 10 | 70% | 30% | 40% |
| A / Scenario 2 | 10 | 70% | 30% | 40% |
| B / Scenario 1 | 10 | 50% | 70% | 70% |
| B / Scenario 2 | 10 | 50% | 70% | 70% |
| C / Scenario 1 | 10 | 20% | 30% | 20% |
| C / Scenario 2 | 10 | 20% | 30% | 20% |
| D / Scenario 1 | 10 | 30% | 30% | 20% |
| D / Scenario 2 | 10 | 30% | 30% | 20% |
How to read this page
Compare groups only after checking their definitions and bases. Here, every within-concept scenario difference is zero by construction. These are not estimates of audience differences.
What to establish for a real study
Agree which subgroups matter to the decision, how records enter them and which limits apply to small bases. Document weighting and any uncertainty method where they are used.
The sample contains no personal data, demographic attributes, real participant IDs or external identifiers.
07 / METHODS & DISCLOSURE
Make the handoff checkable
What this example does - and does not - establish
| Item | Sample specification |
|---|---|
| Data origin | Manually constructed example values; no study run |
| Materials | Four neutral labels; no creatives or client material |
| Records | 80 invented records; 20 per concept |
| Measures | Three 1-5 coded scale values per record |
| Summaries | Unweighted means, top-two shares and counts |
| Subgroups | Two arbitrary scenarios; 10 records per concept/group |
| Text examples | Eight separately written excerpts; no record linkage |
| Inference | No confidence intervals, tests or population claims |
Companion files
custom-study-data.json contains the coded records, calculated summaries and separately labeled text examples. custom-study-codebook.md defines the fields, scales, denominators and formulas. custom-study-sample.html provides the complete report in accessible HTML.
Reproduce a table
Filter responses to one concept. Average a metric for its mean. Count codes 4 and 5, divide by the number of included records and multiply by 100 for its top-two share. To reproduce a subgroup row, also filter by scenario.
Use and limitations
Use this package to inspect the format and handoff. It is not a real client report, a validation result, a forecast or evidence of performance. Do not infer commercial terms, a standard deliverables bundle or a study method from this fictional example. Actual scope is agreed for each engagement.
Every public file was rebuilt from newly authored content. No original creatives, client findings, identities or source links are included.