MARKET RESEARCH WORKSHEET
Companion: https://www.spaceoffice.ai/blog/market-research
Version: 2026-10-21

COPYABLE DECISION BRIEF
Business decision and human owner:
Approved offer / audience / location / relevant period:
Uncertainty that could change the decision:
Facts the business itself must verify:
Evidence that would support a test / rule it out / remain unknown:
Time, spending, access, and recruitment boundaries:
Expected packet and review condition:

METHOD PLAN
Question | existing evidence | missing evidence | suitable method | limit:
Primary research: new information collected for this question.
Secondary research: existing material, including approved historical records.
Either may contain qualitative detail or quantitative data.
Do not call an interview plan a findings report before interviews occur.

SOURCE REGISTER
Evidence ID | author/source | URL/approved file | collected/published date:
Geography | audience | definitions | narrow observation | relevant limit:
Original source behind any repeated or syndicated data:

RECRUITMENT AND QUESTION CHECK
Relevant participant criteria:
Recruitment route, invitation, and participation conditions:
Groups missed or underrepresented:
Permitted use of notes / recordings / summaries:
Pretest of question meaning and response options:
Exact question wording and reference period:
Avoid assigning population prevalence or a conventional sampling-error margin
to a small convenience interview set.

FIVE INTERVIEW PROMPTS TO ADAPT
1. Walk through the most recent time you needed a meeting or training venue.
2. What prompted the search, and what date, duration, and constraints mattered?
3. Which alternatives did you consider, why did you choose one, and who approved it?
4. What difficulties arose with timing, access, setup, or booking?
5. What happened afterward, and which details remain uncertain?
Ask without suggesting the preferred answer; keep faithful permitted quotes
separate from researcher summaries and interpretations.

SYNTHESIS LOG
Observation with source/date:
Interpretation relevant to the decision:
Other explanation or contradictory evidence:
Coverage and other limitations:
Missing fact or behavior needed next:
Proposed action, conditions, human approval, and review date:

FICTIONAL EXAMPLE — NOT RESEARCH CONDUCTED BY SPACE OFFICE
Decision: should a coworking operator test an evening meeting-room offer
for already-defined local training organizers?
20 deduplicated enquiry organizations in the preceding six months:
10 weekday-evening requests + 6 daytime + 4 unspecified = 20.
These are self-selected historical contacts, not the whole local market.
6 purposive interviews, one organizer per organization:
3 past customers + 3 noncustomers = 6.
In the invented findings, 4 report a recent weekday-evening need,
and 2 report a weekend need as the main unresolved timing issue.
3 of the 6 say they would consider the offer; this is stated interest,
not an actual booking or evidence of the accepted price.
Past-customer interviewees may overlap the historical log.
Do not add 20 and 6 as 26 independent respondents.

PROPOSED SMALL TEST
Three weeks of Tuesdays and Thursdays = 6 available evenings.
At most one booking per evening.
Operator verifies actual room conditions, access, staffing, availability,
costs, and offered price before approving a real pilot.
Fictional local decision rule: at least 3 distinct organizations complete
a paid booking during the pilot, followed by an economics/operations review.
This threshold is a teaching assumption, not a universal validation rule.
Record reservations, stated interest, completed paid bookings, returning
organizations, and cancellations separately. No profitability is established.

SPACE OFFICE REVIEW PACKET
Sodium: plan, sources, questions, synthesis, and traceable uncertainty.
Oxygen: business implications, bounded options, and decision criteria.
Hydrogen: coordination and review against the original brief.
Human: verifies decisive evidence, authorizes access and recruitment,
resolves delivery facts, and approves any test, spending, or publication.
AI cannot invent missing respondents, conversations, or observations.