Gmkt Competitor Content Audit

---
name: gmkt-competitor-content-audit
description: Maps three to eight competitors' public content, with channels, themes, formats and posting rhythm, their claims sorted into table stakes, rare and specific, and the gaps nobody fills; a map, never a ranking. Use for "run gmkt-competitor-content-audit", "competitor content audit", "what are competitors doing on social", "competitor social media analysis", "audit competitor content", "how do we stand out from competitors", "everyone says the same thing", part of the Claude for Marketing Graduates Pack by Polar Bear.
---

# Competitor Content Audit

## When To Use
Your manager asks "what are competitors doing on social" by Friday. Use this when you need a clear picture of what the brands next to yours publish, which claims are so common they say nothing, and where the space is.

## When Not To Use
If the question is what your own brand should do about it, take the gaps into the SWOT Analysis; this audit only maps competitors. If you can name fewer than three competitors, the map will be thin: say so, or scroll their feeds for twenty minutes and note themes by hand.

## Inputs
- 3 to 8 competitors you name, and the channels that matter (for example Instagram, TikTok, LinkedIn, newsletter, website)
- One time window you set (for example the last 30 days)
- Their public posts, captions, page copy or screenshots, pasted (with Claude in Chrome on a paid plan, I can read the public pages you name; I do not go looking for more)
If you have none of this, I start from the competitor names and what you remember of their content, and mark the output as a first draft.

## Approach
A content audit (practitioner convention, no single originator) combined with a claims sort: every claim a brand makes about itself goes into one of three piles. Table stakes are claims half or more of the set make; rare are claims one or two make; specific are claims that come with something a reader could check. The judgment is that the table stakes pile is noise, not a lie. The failure this prevents: a 30-slide deck of screenshots ending "they post a lot of reels", with no sentence your team can act on.

## Workflow
1. Ask: which competitors and channels; what time window; is there a question behind the request (a launch, a new channel, a dip)?
2. Per competitor and channel, record themes, formats (carousel, short video, static image, text, link) and posting rhythm counted from what is visible in the window. Engagement goes in only if you observed it on the post; I never estimate reach or likes.
3. Extract every claim each brand makes about itself and rewrite it as a plain statement ("loved by thousands" becomes CLAIM: popular). Mark whether it came with evidence a reader could check (a named source, a number with a source, a described practice).
4. Sort the claims into piles: table stakes (half or more of the set make it), rare (one or two), specific (comes with something checkable, in any pile). You can move the half threshold.
5. Find the gaps: themes, formats or audience questions nobody covers, and the hard truths nobody admits (price, delivery times, what the product does not do). If a persona or Jobs to Be Done Map exists, note which forces nobody speaks to.
6. Write it up as a map: what is said, how often and by whom. No best competitor, no scores, no league table.

## Output Format
```markdown
# Competitor Content Audit
Set: [competitor A, B, C...] · Channels: [list] · Window: [dates] · Collected by: [name]
## Content map
| Competitor | Channel | Main themes | Formats | Posts in window |
|---|---|---|---|---|
| [competitor A] | [channel] | [themes] | [formats] | [count seen] |
## Claims
| Claim (plain statement) | Made by | Count | Evidence a reader could check |
|---|---|---|---|
| [claim] | [A, C, D] | [n of n] | [yes: what / no] |
## Piles
- Table stakes (half or more): [claims]
- Rare (one or two): [claims]
- Specific (checkable): [claims]
## Gaps
- Nobody covers: [themes, formats, questions]
- Nobody admits: [hard truths]
## Decision
[Name] decides which gap to explore in the SWOT or the next brief, by [date].
```

## Done When
- Every competitor row covers the same channels and the same window.
- Counts come only from what was visible; no estimated reach or engagement.
- Every claim has a pile and an evidence note.
- At least one gap is stated, or the map says plainly that none was found.

## Quality Bar
- Public brand content only; never collect anything about named employees, followers or reviewers.
- A thin set (under three competitors, one channel) is labelled thin at the top.
- Table stakes are described as noise, never as lies.
- Never invent a post, a claim or a count for a competitor whose content you were not given.
- Claude maps what competitors publish about themselves; it never invents their numbers or ranks them.

## Next
Run gmkt-swot-analysis (SWOT Analysis) to turn the gaps into opportunities for your brand.

## About the makers

This pack is made by Polar Bear, a consultancy built by ex-McKinsey founders with a dream to make AI work for People, not instead of them. We help our clients build people systems and AI-first ways of working, and we run our own company on Claude. If your team has outgrown the self-serve version, message Pauline (linkedin.com/in/paulinebertry).