SOP-048 — Research plan

SOP · owner: Design Research (Design Research Director). Layer: SOP/Runbook (docs/canon.md). Method definitions and the spine’s role live in docs/methodology.md and the JD — not restated here. State note (2026-06-01): the repo holds zero research evidence and no participant seam; until a live seam exists, plans resolve to simulated studies at a stated confidence penalty. This procedure is executable today (the plan is produced; the study it plans may be simulated).

Interface Contract

Inputs: phase / gate context ← from CPTO · trigger: de-risk commissioned
Inputs: concept to de-risk ← from Innovation Director · trigger: de-risk commissioned
Inputs: framework / spec for Director review ← from <Instructional Designer / UX Designer> · trigger: review requested
Outputs: research plan + brief → to UX Research Associate · trigger: research brief issued
Outputs: research activities (status) → to all agents · trigger: briefing posted
Outputs: recruitment request (personal data) → to Personal Assistant · trigger: personal-data scoped
Outputs: research plan → to Archivist · trigger: artifact approved
Handoff trigger (in): a phase gate approaches and gate evidence is needed, or Innovation hands a concept to de-risk
Handoff trigger (out): the plan is approved and the brief is issued to the UXR (or the study is run/simulated by the Director)
Definition of Done before handoff:
- [ ] The gate question is named and traced to a docs/methodology.md exit criterion
- [ ] The leanest method that answers it is chosen (and justified — no method for its own sake)
- [ ] Instruments specified (screener 3–5 criteria; open-ended, task-framed guide; debrief template separating quote / behavior / interpretation)
- [ ] Confidence ceiling stated (simulated vs. live) before any study runs
- [ ] Recruitment touching personal data routed through the PA (never direct)
- [ ] Learnings captured to docs/knowledge/ if the planning surfaced one
- [ ] Plan handed to the Archivist (0 drift)

Registered triggers (gate evidence ready, research brief issued) — added to docs/handoff-triggers.md (Wave-2 block, Operations-approved 2026-06-01). briefing posted, personal-data scoped, and artifact approved already existed. The recruitment-request edge is absorbed by the PA’s personal-data scoped fan-in sink (SOP-040, §3.1 R1); the CPTO / Innovation / UX-Research / PMO edges reconcile as pending (those roles are not Interface-Contract identities here — §3.1.1 R3).

Purpose

Establish, at each phase boundary, what evidence the gate needs, by what method, on what timeline — so the venture advances on evidence, not assumption, and runs the leanest study that produces a real signal.

When to use

Whenever a phase gate (Ideate / Plan / Build / Grow / Operate) approaches and gate evidence is owed, or when the Innovation Director hands a concept that needs de-risking, or when in-market/telemetry signals call for a structured read.

Flow

flowchart LR
  A["Identify the<br/>gate question"] --> B["Choose the<br/>leanest method"]
  B --> C["Draft the<br/>instruments"]
  C --> D{"Live participants<br/>available?"}
  D -->|yes| E["Brief UXR<br/>(SOP-051)"]
  D -->|no| F["Plan as simulated<br/>(state confidence penalty)"]
  E --> G["Sequence with PMO;<br/>route recruitment via PA"]
  F --> G
  classDef s fill:#1d4ed8,color:#fff,stroke:#1e3a8a
  classDef d fill:#0f766e,color:#fff,stroke:#134e4a
  class A,B,C,E,F,G s
  class D d

Text fallback: name the gate question (traced to a docs/methodology.md exit criterion) → choose the leanest method (JTBD interview, concept test, usability walkthrough, telemetry synthesis, short survey) → draft the instruments → are live participants available? (yes → brief the UXR via SOP-051; no → plan as a simulated study and state the confidence penalty) → sequence the activities with the PMO and route any personal-data recruitment through the PA.

Steps

Identify the question. Work from docs/methodology.md phase deliverables and exit criteria. State the one question the evidence must answer for the gate.
Choose the method. Select the leanest method that produces a real signal. Justify it; reject method-for-its-own-sake and over-long studies (a one-week question gets a one-week study).
Draft the instruments. Screener (3–5 criteria); interview/usability guide (open-ended, task-framed, no leading questions); debrief template that captures verbatim quotes, observed behaviors, and interpretations separately.
Set the confidence ceiling. If no live participant seam exists, the study is simulated — state this and the confidence penalty up front, in the plan.
Sequence and route. Hand the research activities to the PMO for delivery sequencing; route any recruitment that needs personal contact/calendar data through the PA (personal-data scoped).
Brief execution. For evaluative/telemetry work, issue the brief to the UXR (SOP-051, research brief issued); for generative work, run or simulate it yourself.
Hand off the plan to the Archivist; capture any learning.

Roles / approvers

Run by the Design Research Director. The Director owns method selection and exit criteria; the UXR executes within the plan but never modifies it. The CPTO + Principal are the downstream consumers of the evidence the plan produces (via SOP-049).

Guardrails

Lean research — the smallest study that answers the question; no method for its own sake.
Simulation is labeled — no live seam means a simulated study with a stated confidence penalty; never present a simulated plan as a live one.
No leading instruments — open-ended, task-framed; a leading question is a research defect.
Personal data is PA-only — recruitment routes through the PA; never accessed directly, never archived.
Stay in your lane — the plan produces evidence, not concepts (Innovation), market sizing (Growth), or builds (Engineering).
Do not restate docs/methodology.md method definitions — link to it.

See also: SOP-049 (gate evidence) · SOP-051 (usability protocol) · docs/units/design-research/charter.md · docs/handoff-triggers.md.