AskCraft / AUDIENCE RESEARCH REPORT
Saved recipes. Different dinner choices.
How to organise recipes around a reader’s circumstances rather than dish categories alone.
Creator blog
AC-REPORT-EN · 15 September 2026 · Version 2.0
Teaching example. All accounts and findings are fictional.
Contents
Help readers choose a feasible dinner first. There is no basis for shortening every piece.
03 Decision and research design
08 From quotation to interpretation
11 Questions for further research
12 Limits and using the material
AC-REPORT-EN · 15 September 2026 · Version 2.0
Decision and research design
The creator must choose between changing categories, recipe cards and content length. Research question: what happens between saving an idea and deciding to cook?
JTBD examines the progress a person seeks in a specific situation. Questions about a past choice reconstruct circumstances, alternatives and decision criteria. A job statement is a researcher’s interpretation to be checked against the material [1].
Who to recruit in a real project
• A recent, specific recipe decision, including abandoning a recipe.
• Contrast weekday meals, entertaining and learning a technique.
• Include people who chose a familiar dish, delivery or a friend’s advice.
• Recruit beyond the creator’s readers; apply the same eligibility criteria to followers.
Collection and analysis
Agree the decision and eligibility criteria first. Use a short screener to check specific experience. Reconstruct the sequence of choice in interviews. Record observations separately, compare patterns and examine exceptions. Keep links from recommendations to the source accounts [2, 3].
Example size is not a sample recommendation |
Six accounts were written to illustrate differences. They cannot establish prevalence, saturation or expected impact. Real scope depends on the diversity of situations and the decision risk. |
Jobs and their basis
A saved recipe can be mistaken for an intention to cook. These accounts separate two moments: an attractive idea enters a collection, while the meal decision happens later under different constraints.
J01 · R01, R02
Get weekday dinner done with less effort
With low energy and limited time, choose feasible food from available ingredients to complete a household task.
J02 · R03, R04
Prepare a meal for others with confidence
When cooking for guests, understand sequence and risks beforehand to complete the meal successfully.
J03 · R05, R06
Learn a principle and apply it independently
With time to learn, understand a technique well enough to adapt rather than follow steps rigidly.
A difference that must remain visible |
R01 wants less decision-making while R05 wants time to understand. We therefore do not recommend making everything shorter. The distinction concerns situations, not permanent categories of people. |
J identifies a job hypothesis; R/E/C identifies a fictional account. A reference makes the reasoning traceable, but does not turn fictional material into empirical evidence.
Evidence: accounts 01–02
R01 / J01
Late weekday evening; ingredients are already at home.
“I save what looks good. In the evening I need something I can make with what I have.”
Opened saved recipes, could not find substitutions and made a familiar omelette. Saving did not mean intending to cook that recipe.
Interpretation | Situational navigation may reduce search effort, but a familiar meal remains an alternative. |
Alternative explanation | Fatigue may have made any unfamiliar dish unsuitable. |
Next question | What would a visible substitution have changed? What happened on a comparable evening? |
R02 / J01
The family needs dinner before evening commitments.
“Twenty minutes turned out to mean time at the stove. I noticed the extra hour later.”
Chose a video by its stated time, discovered marinating and ordered delivery. Time until eating mattered more than active cooking time.
Interpretation | The card should support the decision before cooking, not just describe active work. |
Alternative explanation | Delivery may have been preferable regardless of timing accuracy. |
Next question | When did you realise dinner would be late, and which option was still feasible? |
Evidence: accounts 03–04
R03 / J02
Cooking for friends for the first time; wants to avoid letting them down.
“I asked my friend which step usually goes wrong.”
After browsing, chose a recipe a friend had tested. Needed confidence in execution rather than another idea.
Interpretation | Guest-oriented content needs risk guidance, but a label cannot replace personal trust. |
Alternative explanation | The friend’s recipe may have been more familiar or better suited to the guests. |
Next question | What exactly did your friend say? Which answer ended the search? |
R04 / J02
Several dishes need to be ready at the same time.
“Every recipe has its own steps. I need one plan for the whole evening.”
Saved a collection, then asked a friend to sequence the work. The barrier sits between recipes rather than inside one.
Interpretation | The useful unit of support is the whole evening, not one recipe card. |
Alternative explanation | Equipment, skill or serving count may have been the binding constraint. |
Next question | Which two tasks competed for time or equipment? |
Evidence: accounts 05–06
R05 / J03
A free Saturday; learning to handle dough.
“I want to understand why it behaves that way. A short clip does not explain it.”
Moved from short clips to a detailed explanation. Length was acceptable because learning was the goal.
Interpretation | Length is evaluated against the job: explaining causes may have value in its own right. |
Alternative explanation | This account does not establish successful learning. |
Next question | What did you do differently afterwards, and how did you judge the result? |
R06 / J03
Wants to apply a technique to different ingredients.
“If I understand the principle, I can substitute. If I only know the steps, I am afraid to change them.”
Chose a long explanatory article and returned during another attempt. The material supported transfer to a new situation.
Interpretation | Reuse may reflect transfer of a principle rather than entertainment value. |
Alternative explanation | The return may have been caused by an unclear initial explanation. |
Next question | Which part did you revisit, and what could you now do without assistance? |
From quotation to interpretation
One fragment and the limits of inference
Interviewer
What happened before you ordered delivery?
R02
I had taken the ingredients out. Then I saw: marinate for an hour.
Interviewer
What did you need to have done by then?
R02
Eat before my child’s evening activity.
Interviewer
What options were still available?
R02
I could have made something familiar quickly, but I was frustrated and ordered.
What the follow-up changes |
The fragment establishes a deadline and an alternative. It does not prove that correcting the time label alone would solve the problem: frustration had already changed the subsequent choice. |
Reading the analysis chain
Source | R02: Time information did not match the deadline |
Hypothesis | The card should support the decision before cooking, not just describe active work. |
Competing explanation | Delivery may have been preferable regardless of timing accuracy. |
Missing evidence | When did you realise dinner would be late, and which option was still feasible? |
In a real study, the excerpt would reference a recording or transcript location. No such recordings exist here; timestamps have not been invented.
What to change
These are specifications for small prototypes, proposed for testing rather than changes already implemented.
D01 / First · J01
Before-you-cook card
At the top: ready-to-eat time, active work, waiting, equipment and substitutions. Show waiting before the steps.
Usefulness criterion: Before acting, the reader can judge feasibility and reject an unsuitable recipe for a stated reason.
Suggested owner: Recipe editor
D02 / Next · J02
Guest-menu schedule
Join dishes into one sequence: advance preparation, overlapping tasks, likely mistakes and recovery options.
Usefulness criterion: The reader can plan preparation for their own kitchen; unknown conditions remain explicit.
Suggested owner: Collections editor
D03 / Separately · J03
Technique explainer
Retain the explanation of causes. Add a task that changes an ingredient or condition and asks the reader to explain how the method changes.
Usefulness criterion: The reader transfers the principle to a new case rather than repeating the author’s words.
Suggested owner: Creator / learning designer
What to defer |
Do not scale before checking the mechanism. Start with one piece and one scenario, comparing with the current approach. The order reflects dependencies, not a calculated return on investment. |
How to test the changes
Test behaviour, not approval of the idea
Use comparable tasks for the current version and prototype, varying presentation order. Do not lead the answer. Record actions, errors and explanations separately. Recruit for the situation, not simply follower status.
T01 / D01
Give a deadline and available ingredients. Ask the person to choose a feasible option and explain it.
Record: Feasible recipe selection; understanding of waiting time; reason for rejection.
Reason to reconsider: If the card is understood but fatigue prevents cooking, presentation is not the whole issue.
T02 / D02
Ask someone to sequence preparation and resolve a specific mistake.
Record: Preparation sequence; dependencies identified; requests for help.
Reason to reconsider: If personal advice remains necessary, written instructions may not provide enough confidence.
T03 / D03
Introduce another recipe and check whether the person can explain how to adapt.
Record: Application in a new situation; transfer errors; return to the explanation.
Reason to reconsider: Repeating the explanation without applying it means the learning job is not yet supported.
Decision rule before testing |
Continue when the person independently completes the intended action and correctly explains its limits. Revise when prompting is required or a critical error occurs. If the task is irrelevant or resources are the constraint, revisit recruitment and the offer. Quantitative thresholds and test scope require a separate agreement; no results exist yet. |
Questions for further research
Screening before invitation
Ask for a specific recent episode, the person’s role and the outcome. Check eligibility against page 2. Record recruitment source and missing perspectives. Do not require a positive opinion of the brand [2].
Core interview questions
01. Recall the last time you selected a recipe but made something else. What happened?
02. What ingredients were available, and when did you need to eat?
03. When did you save the recipe, and when did you return to it?
04. What alternatives did you consider after leaving the recipe?
05. What helped you judge whether the meal would work?
06. When was a detailed explanation useful, and when was it a burden? Describe two episodes.
Follow-ups for any episode
What happened immediately before this? What did you do next? Which options did you actually consider? Can you show material you used? How did the situation end?
Avoid substituting speculation for evidence |
Replace “Would this feature help?” with a reconstruction of past choice and observation of prototype use. Future intention, praise and willingness to leave contact details are not usage or purchase. |
Limits and using the material
No prototype use, job prevalence or audience growth has been measured. Saving, cooking and returning are different behaviours and should not be collapsed into one usefulness metric.
Teaching example. All accounts, quotations and interpretations are fictional. No field study or prototype tests were conducted.
What accompanies this report
The Excel workbook contains six sheets: orientation, source material, interpretations, decisions, tests and questions. Account, job, decision and test identifiers match this report. CSV remains available as a simple exchange format for the source material.
How to use it in your work
First define your own decision and recruitment criteria. Copy the structure and replace all fictional material with real data under agreed access rules. Do not transfer these hypotheses to another audience. Preserve the distinction between quotation, interpretation and proposed action.
Method references
[1] Christensen Institute — Jobs to Be Done
https://www.christenseninstitute.org/theory/jobs-to-be-done/
[2] GOV.UK — Finding participants
https://www.gov.uk/service-manual/user-research/find-user-research-participants
[3] GOV.UK — Analyse a research session
https://www.gov.uk/service-manual/user-research/analyse-a-research-session
These sources explain the approach; they do not validate the fictional accounts or findings in this example.