# Research

> Research

- Skill: `everyinc/research` (Agent Skill, multi-file: 8 files)
- Install (CLI): `npx skillmds@latest add everyinc/research`
- Raw SKILL.md: https://api.skillmd.com/api/skills/everyinc/research/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Marketing & Growth
- Author: everyinc (https://skillmd.com/u/everyinc)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/everyinc/research

---

# Research

Conduct and synthesize the evidence required for positioning, brand character, messaging,
and art direction. The skill treats research as a decision system: every important claim
is sourced, graded, and connected to a strategic implication.

## When to invoke

- When positioning needs customer, category, or competitive evidence
- When scoping customer surveys, interviews, ethnography, or online research
- When a brand has no first-party data and needs a provisional public corpus
- When building a required competitive audit or positioning matrix
- When separating verified evidence from directional findings and hypotheses
- When deciding whether the corpus is sufficient for a first strategic recommendation

## Evidence standard

Load `references/evidence-standard.md`. Prefer recent proprietary customer and product
data. Hard evidence may also come from credible research with a sample appropriate to the
claim, usually from the last one or two years. Qualitative evidence includes ethnography,
interviews, studies, reviews, memes, and internet chatter when it is collected and labeled
honestly.

Do not flatten every source into equal “insight.” Maintain an evidence ledger with source,
date, sample, audience, method, finding, limitation, and confidence.

## Research paths

### Existing brand

Use surveys, usage and revenue data, customer and churn interviews, support logs, reviews,
sales calls, campaign results, and leadership context. Include retained, high-value,
low-usage, and churned customers where available.

### New brand

If the team cannot survey customers, build an online corpus from named sources. Use
category studies, public datasets, reviews, forums, search behavior, social posts,
competitor material, interviews, ethnography, and cultural artifacts. Mark the customer
model as provisional and add simple post-launch validation triggers.

## Required competitive work

Load `references/competitive-audits.md` and `references/positioning-matrices.md`. Audit
direct competitors, substitutes, adjacent categories, and the default behavior of doing
nothing. A positioning matrix is required for compound-brand work.

## Synthesis

Separate:

- Observed fact
- Repeated qualitative pattern
- Directional signal
- Inference
- Strategic recommendation
- Open hypothesis

For each recommendation, show the supporting evidence and the evidence that would change
the conclusion.

## Dependencies

- `foundation/marketing-os`

## Quick checklist

- [ ] Research question and decision are explicit
- [ ] First-party and public evidence are separated
- [ ] Sources include date, sample, method, and limitations
- [ ] Customer language is preserved verbatim where useful
- [ ] Direct, substitute, adjacent, and do-nothing competition are covered
- [ ] Positioning matrix axes were chosen by the human
- [ ] Findings, inference, and recommendation are visibly distinct
- [ ] New-brand hypotheses have lightweight post-launch validation triggers