Positioning Audit

The standalone positioning diagnostic EntireCommerce AI runs on DTC brands with traffic and revenue but unclear or stale messaging. Pulls apart founder vision, current-state messaging, ICP language, and competitor positioning. Finds the overlap, whitespace, and misalignment. Outputs three positioning options with full messaging systems. Runs end-to-end from a single Claude Code prompt in 20 to 30 minutes.

Purpose

Surface the positioning opportunities where ICP language, founder conviction, and competitor whitespace overlap. Produce three positioning options (Safe, Bold, Bet-the-brand) with a complete messaging system for each. The audit is the artefact a founder uses to decide which positioning move the brand commits to next and what messaging to roll out where.

Positioning is the highest-leverage decision a founder makes. Pricing changes, creative refreshes, and channel shifts all compound on top of the positioning frame. A sharp frame doubles the return on every subsequent marketing dollar. A blunt frame halves it. This playbook exists so the positioning decision stops being a 90-day rebrand project and becomes a one-session diagnostic with three concrete paths forward.

Distinct from the GTM Audit's Path B intersection pass (which is a single sub-section inside a broader audit). This is a standalone, 20 to 30 minute run targeted at founders who already have traffic and revenue but are unclear on messaging.

When to run it

• Pre-rebrand. Founder has been told the brand needs a messaging refresh. Wants a data-grounded starting point before briefing a designer or agency.
• Post-launch, positioning drift. The brand has 3+ months of trading and the founder cannot explain in one sentence why the buyer picks them over the competition.
• Pre-raise or pre-exit. Investors or acquirers ask for a one-page positioning doc. The audit produces the raw material.
• Stalled growth. Traffic is flat, conversion is flat, ads are not scaling. Often a positioning problem dressed up as a performance problem.
• Pre-launch clarity check (Path B). The founder has a product and an audience hypothesis but has never put positioning on paper. Run before the site build.

Who this is for (ICP)

Premium DTC brands, AOV $500+ or 2x category median, founder-led, US / UK / western-European markets. Traffic 5,000 sessions per month or more, or pre-launch with a clear ICP and 3-5 named competitors. The playbook works outside that ICP but is tuned for it.

Two paths

Path A. Existing brand with live site, ads, content

The brand has a live site, running ads, a social presence, and an email list. The current-state messaging audit (Section 2) catalogues every explicit positioning claim across those surfaces. Sections 4 and 5 compare that current state to ICP language and competitor claims.

Path B. Pre-launch or new brand, founder vision only

No live site to audit, no running ads, no email corpus. Section 2 (current-state messaging audit) compresses to a founder-vision capture only. Sections 4 (competitor positioning scan) and 5 (overlap, whitespace, misalignment) still run in full. The output is three positioning options the founder commits to before building the site.

Path detection is automatic: Claude checks whether a primary URL is provided and renders any public copy. If the site has fewer than three page surfaces (homepage, one product page, and an About page), Path B applies.

Tool-availability gate

The audit runs entirely on the External layer. No client logins required.

Gate behaviour

For every check the audit plans to run:

1. If the tool is listed above, run the check. All tools are External and always available.
2. If WebFetch fails on a specific URL (JS-rendered page returns empty, common for single-page apps), fall back to Serper for the meta/title pair and note the limitation in Section 2.
3. If Meta Ad Library WebFetch fails (JS-rendered), use the fallback chain: Playwright script if available, third-party ad-library indexes (adscan.ai), manual screenshots. File any unobtainable scan as an appendix gap.

Never fabricate a positioning claim. A missing data point is always better than an invented one.

Inputs the playbook accepts

Prompt the user for these at the start of the run. Accept defaults from CLAUDE.md where the fields already exist.

If any required input is missing, ask. Never proceed with invented inputs.

Output: one consolidated report, two file formats

The audit produces exactly one report in two formats side by side:

{client_folder}/docs/positioning-audit/{YYYY-MM-DD}/positioning-audit-report.md
{client_folder}/docs/positioning-audit/{YYYY-MM-DD}/positioning-audit-report.docx

The .md is the source of truth. The .docx is the polished Word deliverable, produced via pandoc after the voice gate passes. Length target: 3,500 to 6,000 words.

Supporting evidence lives alongside:

{client_folder}/data/positioning-audit/{YYYY-MM-DD}/
  audience-voc-mining.md   (15-quote verbatim appendix)
  competitor-positioning-scan.md   (per-competitor H1/subhead/claims capture)
  current-state-capture.md   (screenshots and text of the brand's current claims)

Report structure (fixed order)

# Positioning Audit: {Brand Name} ({YYYY-MM-DD})

Path: {A / B}

## Executive summary
3-4 sentences leading with the narrative the intersection reveals. Names the
recommended positioning option and why.

## The intersection map (front-of-report)
One visual block: Founder Vision | ICP Language | Competitor Whitespace, with
the overlap, whitespace, and misalignment regions called out.

## Key findings
5-7 one-sentence bullets. Each surfaces the single most important finding from
one of the eight sections. Skim-readable in 90 seconds.

## The three positioning options (skim layer)
Three blocks: Safe, Bold, Bet-the-brand. One sentence on the category frame for
each, one on the core claim, one on the expected reach-and-risk profile.

## 1. Founder vision capture
## 2. Current-state messaging audit (Path A only; compressed to interview summary on Path B)
## 3. Audience insights
## 4. Competitor positioning scan
## 5. Overlap, whitespace, misalignment synthesis
## 6. Three positioning options (detailed, with messaging systems)
## 7. Recommendation with reasoning
## 8. Messaging rollout plan

## Cross-cutting tensions
3-5 bullets. Each names a tension that surfaces across two or more sections.
Produced by the synthesis pass.

## ICE action list
Every candidate positioning move (including sub-moves like a homepage rewrite,
an ad angle pivot, an email welcome refresh) scored on Impact x Confidence x
Effort. Ranked descending. Grouped into: This week, Next 30 days, Backlog.

## Conclusion
2-3 short paragraphs. What the founder does this week. Where to learn more
about ongoing support if they want the rollout managed. Book-a-call CTA.

Section 1. Founder vision capture

Interview-driven. Captures the things audience data and competitor scans cannot reveal. Same structure as the GTM Audit's Section 2 Vision block, expanded for positioning-specific fields.

Fields to fill

How to execute

• Conduct the interview live, or ask the founder to fill a structured form. A live interview produces sharper material because the founder's first phrasing is the raw positioning material.
• Transcribe verbatim where the founder's phrasing is distinctive. Exact words become positioning language in Section 6.
• Capture the emotional register, the metaphors, the phrases the founder reaches for when explaining the brand to a sceptic. Those phrases are load-bearing.

Minimum output

One page. Seven fields completed. At least three verbatim founder phrases captured in quote marks and stored in the data folder at founder-vision.md.

Section 2. Current-state messaging audit (Path A)

Catalogues every explicit positioning claim the brand currently makes across its public surfaces. On Path B, this section compresses to a single paragraph noting "Pre-launch, no current-state surfaces to audit" and proceeds to Section 3.

Surfaces to audit

Method

1. WebFetch every listed URL. Extract clean text with metadata (H1, H2, meta description, title tag).
2. For each surface, produce a 2-column table: "Surface" and "Claim made". One row per claim. Verbatim.
3. Categorise each claim: Feature (what the product is or has), Benefit (what the product does for the buyer), Identity (who the buyer becomes by using the product), Category (what kind of thing this is), Social proof (numbers, reviews, press).
4. Count the frequency of each category. A brand over-indexed on Feature claims and under-indexed on Identity claims has a positioning gap. Surface the imbalance.
5. Pull the 10 to 15 most-used phrases across all surfaces. Those are the brand's current lexicon.

Minimum output

One table of 20 to 40 rows (surface, claim, category). One frequency chart of categories. A list of the brand's 10 most-repeated phrases.

Section 3. Audience insights

Reference the canonical audience prompt at EntireCommerce/playbooks/audience-insights-master-prompt.md. Run it in full. Do not rewrite the prompt inline.

Feed it the three required inputs:

1. Target audience (from Section 1, the founder's own-words description)
2. Their #1 struggle (from the CLAUDE.md brief or asked at kick-off)
3. What the brand sells and how it helps (from Section 1)

Deliverable per the master prompt: a one-page brief with nine sub-headings (Avatar, Core tension, What they tried and why it broke, Market gap, Recurring language, Three headlines, Top three objections and responses, Positioning angle, Belief shift). Plus a 15-quote verbatim appendix at data/positioning-audit/{YYYY-MM-DD}/audience-voc-mining.md.

Why this section carries extra weight on the positioning audit

On a GTM Audit, audience insights feed multiple downstream sections (SEO, ads, email). On the Positioning Audit, Section 3 is the direct input to Section 5 (overlap, whitespace, misalignment). The exact phrases in the Recurring Language block are compared line-by-line against the brand's current lexicon from Section 2 and against competitor claims from Section 4. Drift between the three reveals the misalignment. Overlap between ICP language and competitor claims with no brand representation reveals the whitespace.

Run the master prompt with the same discipline used on any engagement. 15 verbatim quotes. No speculation. Every phrase in the Recurring Language block traces to at least two quotes in the appendix.

Section 4. Competitor positioning scan

Step 0 — competitor-set + benchmarking opt-in. Run the competitor-set decision tree from PLAYBOOK-TEMPLATE.md. If a competitor set exists, render BT-VOC-trustpilot-google.md (review counts and ratings per competitor — quickest, cheapest signal of customer-felt positioning) alongside the per-competitor narrative below. If a paid-ads BT was already produced by a parallel competitor-creative-audit run, reference it; otherwise leave that pillar to the dedicated playbook. Skip cleanly if the user opts out.

For each of the 3 to 5 competitor URLs, capture the positioning elements a founder would see in 60 seconds on their homepage plus a 30-day ad snapshot.

Per-competitor sub-section

4.1 Homepage hero. H1, subhead, hero subcopy. Verbatim.

4.2 Elevator pitch. The one-sentence description the brand gives of itself on its homepage or About page. Sometimes this is the H1, sometimes it is a block further down.

4.3 Primary promise. The single core benefit the brand leads with. Usually one of: faster, cheaper, easier, more premium, more effective, more ethical, more specialised.

4.4 Secondary claims. The three to five supporting claims that flesh out the primary promise. Captured verbatim from the homepage.

4.5 Tone of voice. Three adjectives. Sources: the homepage copy, the About page copy, any founder-voice content. Example registers: clinical, warm, technical, playful, luxurious, combative.

4.6 Visual identity flavour. One sentence on the visual register. Minimal, dense, photography-heavy, illustration-forward, retro, high-contrast. Feeds Section 8 rollout (visual direction influences messaging system).

4.7 Ad-library angles. 30-day Meta Ad Library snapshot. Top 3 angles in rotation. One sentence each. Use the Section 3.4 fallback chain from the GTM Audit if WebFetch on the ad library fails.

Cross-competitor positioning map (2x2 grid)

After the per-competitor teardowns, produce a 2x2 positioning grid on the two dominant axes the category plays on. Examples:

• Jewellery: Craft-heritage (y-axis) vs Modern-everyday (y-axis low) crossed with Price-accessible (x-axis) vs Luxury-exclusive (x-axis high).
• Short-game training: Data-driven (y-axis) vs Feel-based (y-axis low) crossed with Solo-practice (x-axis) vs Coach-led (x-axis high).
• Luxury resale: Transparency-first (y-axis) vs Opacity-traditional (y-axis low) crossed with Breadth-of-stock (x-axis) vs Curated-rarity (x-axis high).

Pick axes that force the competitors to sort into distinct quadrants. If three competitors land in the same quadrant, the quadrant is commodity territory and the client should avoid it. If one or two quadrants sit empty, that is whitespace and Section 5 investigates whether the ICP actually asks for something in that territory.

Minimum output

One sub-section per competitor. One 2x2 grid. One paragraph of "what every competitor claims" (the commodity territory). One paragraph of "what no competitor claims" (the potential whitespace).

Section 5. Overlap, whitespace, misalignment synthesis

The intersection section. Load-bearing. This is where the three-way data from Sections 1, 3, and 4 becomes a strategic call.

Four-part output

5.1 Overlap: founder vision matches ICP language. The phrases the founder uses (Section 1) that also appear in the Recurring Language block from Section 3. This is the strongest ground. When founder conviction and audience vocabulary already agree, the brand has a positioning asset it may be underusing. Write: "Founder says X. ICP says X. Current site says [Y / not at all]." If the current site says neither, that is the fastest positioning win available.

5.2 Commodity territory: current claims overlap with all competitors. The claims in Section 2 (current-state audit) that also appear in Section 4 (competitor claims). If every brand in the category says "premium craftsmanship" or "expert guidance" or "easy to use", the claim carries zero positioning weight. The brand is burning homepage real estate on a tie. Flag every claim that appears in the brand's top-10 phrases AND in at least 2 competitors' top claims.

5.3 Whitespace: ICP asks for X that no competitor delivers. The Recurring Language phrases from Section 3 that do not appear in any competitor's homepage, ad copy, or positioning claims from Section 4. This is the positioning territory the brand could claim. Rank whitespace opportunities by: how many ICP quotes support the phrase (frequency in the appendix), how far the founder's vision already leans toward it (alignment with Section 1), and how commercially viable the territory looks for the 12-month revenue goal.

5.4 Misalignment: founder vision drifts from ICP language. The phrases the founder uses in Section 1 that the ICP in Section 3 does not use. Drift here is dangerous. The founder is writing copy for an audience in their head that differs from the audience actually buying. Name each drift explicitly. For each one, offer the founder two options: swap the founder's phrase for the ICP phrase, or build a campaign that teaches the ICP the founder's phrase (harder, slower, occasionally right when the founder is creating a new category).

Minimum output

One labelled block per sub-section. A single consolidated intersection map at the end: Overlap in the centre, Commodity territory in the bottom-left quadrant, Whitespace in the top-right, Misalignment as a red-flag list. This map is what appears at the top of the executive summary.

Section 6. Three positioning options (detailed)

Produce three positioning options. Each option has a complete messaging system. Each option represents a different risk-reward profile. The founder picks one.

Option structure (same shape for each)

1. Category frame. One sentence. What category does the brand say it belongs to? Category claims drive which competitors the ICP mentally benchmarks the brand against. Safe options keep the current category frame. Bold options shift to an adjacent category. Bet-the-brand options create or claim a new category.

2. One-sentence positioning statement. Template: "For [ICP] who [struggle], [brand] is the [category] that [primary promise], because [reason to believe]." The positioning statement is internal-facing, the scaffolding every piece of downstream copy hangs from.

3. Headline + subhead. Public-facing. ≤12 words for the headline, ≤20 words for the subhead. Uses phrases from the Recurring Language block of Section 3. The headline is the one-line version a visitor sees above the fold.

4. Five primary promises. The five core benefit claims the brand will lead with. Each a single sentence. Each traceable to either an ICP quote or a founder-vision phrase. These become the H2s on the homepage, the subject lines on lifecycle email, the ad primary text variants.

5. Three objection-handling lines. The three objections the brand hears most (captured at kick-off, expanded in Section 3) each paired with a one-sentence response the brand deploys on the site, in ads, and in email.

6. Tone-of-voice brief. Three adjectives plus three words the brand uses and three words the brand avoids. Feeds every downstream copy decision.

7. Three example ad headlines. Three Meta or Google ad headlines written in the option's voice. Concrete, deployable.

8. Example social bio. A 150-character Instagram / X / LinkedIn bio written in the option's voice.

The three options

Option A. Safe. Iterates on the current positioning. Sharpens the headline, cleans up commodity claims, doubles down on the overlap region from Section 5.1. Low risk. Low ceiling. Good fit for brands doing $1M-$5M who want to compound current performance without a rebrand.

Option B. Bold. Leans into the strongest whitespace opportunity from Section 5.3. Category frame shifts. Primary promise shifts. Carries 6 to 12 months of messaging work to land. Medium risk. Medium-high ceiling. Best fit for brands at $2M-$10M with a 12-month revenue goal that requires meaningful re-acceleration.

Option C. Bet-the-brand. Reshapes the brand around a category-creating claim. The category frame is new (or new to the ICP). The primary promise is something no competitor currently makes. The founder becomes the chief evangelist for the new category. High risk. High ceiling. Best fit for brands with founder appetite for a year of category education, or brands facing an existential positioning problem (current frame is commoditising fast).

Minimum output

Three full messaging systems. Eight elements each. All traceable back to Section 3 ICP language, Section 1 founder vision, or Section 5 whitespace analysis. No positioning option produced from thin air.

Section 7. Recommendation with reasoning

The operator picks one of the three options and defends the pick. Not optional. Recommendations without conviction are diagnostics the founder has to turn into decisions on their own, which is the slow path.

Structure

• The recommendation. Option A, B, or C. One sentence.
• Why this one. Three bullets. Each bullet links the recommendation back to specific evidence from Sections 3, 4, and 5. Example: "Option B wins because the Section 5.3 whitespace (ICP asks for trains the shot I hit on the course) is supported by 8 of the 15 audience quotes, no competitor claims that language, and the founder's vision in Section 1 already leans toward it."
• Risks of each option. One paragraph per option. Safe: ceiling too low to hit the 12-month goal. Bold: 3-6 months of traffic turbulence while the brand teaches the new frame. Bet-the-brand: 12 months of founder evangelism before paid acquisition becomes reliable.
• Proof the founder needs before committing. Concrete tests the founder runs before the full rollout. Example: run a one-week Meta Ads test with the recommended option's headline against the current-state headline. If the recommended option's CTR is within 70% of the current-state headline or better, proceed with the rollout. If it is below 70%, return to Section 6 and tune.

Minimum output

One page. Clear call. Evidence cited. Risks named. Test designed.

Section 8. Messaging rollout plan

Sequence of touchpoints to update. The rollout plan is the difference between a report the founder files and a report the founder uses.

Rollout sequence (default order)

1. Homepage hero first. The H1 and subhead are the single highest-leverage surface. Update before anything else. One-week build, 48 hours to measure the conversion delta.
2. Meta and Google ads second. New primary text and headline in rotation alongside the current-state control. Allocate 30% of the budget to the new variants for two weeks. Compare CTR, CPA, and ROAS against the control.
3. Social bios and primary pinned posts. Cheap, immediate, compounding. The visitor landing on Instagram from an ad sees the new positioning in the bio.
4. Email welcome series. Subject lines, H1s in the first three emails of the welcome flow, and the primary promise in email #1. Email is the first high-trust touchpoint after capture. It has to match.
5. Product pages last. The most expensive surface to update, and the one where conversion sensitivity to copy is highest. Do this after the homepage test has passed.

What to test before full rollout

• Two-week ad test between the new headline and the current-state headline. Pass criterion: CTR within 70% of control or better.
• Two-week homepage A/B test if the brand has a CRO tool (VWO, Optimizely, Shopify's native experiments). Pass criterion: PDP add-to-cart rate holds or improves.
• Two-day founder sanity check with five customers or five prospects from the ICP. Show them the new headline, ask what it means, ask who they think it is for. If three of five describe a target different from the intended ICP, the headline is failing.

KPIs to watch

• Leading indicators (first 14 days): homepage bounce rate, ad CTR, email open rate on welcome flow, time-on-homepage.
• Mid indicators (30-60 days): add-to-cart rate, welcome-flow revenue per send, branded-search volume.
• Lagging indicators (90 days): revenue growth rate, customer acquisition cost, organic traffic on branded queries.

The positioning change is not proven until the lagging indicators move. Keep measuring for a full quarter.

Minimum output

One sequenced rollout plan. Three tests with pass criteria. One KPI table.

Synthesis pass (mandatory before writing to disk)

Same pattern as the GTM Audit. After drafting all eight sections, do not write to disk yet. Read the full draft end-to-end in one pass. Ask:

• What cross-cutting tension surfaces in three or more sections that no section captures on its own? (Example: founder vision in Section 1 leans toward craft-heritage, ICP language in Section 3 uses modern-everyday, competitor whitespace in Section 4 shows craft-heritage is commodity territory. The three combined reveal a misalignment the three-option frame has to resolve.)
• What is the single positioning move that would collapse the most ambiguity in the brand? That move becomes the headline of the Executive Summary.
• Is any ICE-ranked action in the action list solving only one surface? Consider demoting in favour of moves that touch multiple surfaces.

Produce 3 to 5 cross-cutting tension bullets. Revise:

• Executive Summary to lead with the positioning call the intersection reveals.
• The three positioning options to cross-reference the synthesis tensions where relevant.
• ICE action list to promote multi-leverage moves (anything that updates multiple surfaces from one source of truth).
• Write the synthesis bullets into the Cross-cutting tensions block immediately after Section 8.

Voice enforcement gate (mandatory before writing to disk)

Positioning copy invites every banned pattern. The gate matters double on this playbook.

Before writing the file, run a mechanical scan on the full draft. Grep for each banned pattern and rewrite every match:

Positioning writing is full of "X not Y" temptation. The Safe / Bold / Bet-the-brand option frame invites contrast copy. Resist every "this option is X, not Y" instinct. State what each option IS and let the comparison surface from the specifics.

The voice gate must hit zero matches before the report is written to disk.

ICE action list

Every candidate positioning move becomes a candidate initiative. Examples:

• Rewrite homepage H1 and subhead to Option B (headline + subhead)
• Run two-week ad test: Option B primary text vs current-state
• Update Instagram bio and pinned posts to Option B social bio
• Rebuild email welcome flow with Option B primary promise
• Commission founder-voice About page rewrite with three verbatim quotes from founder-vision capture
• Test Option B against five ICP prospects (qualitative comprehension check)
• Update product-page H1s for top-3 revenue SKUs to Option B primary promises
• Brief the designer on Option B tone-of-voice and three adjectives

Score each on Impact x Confidence x Effort (inverse). Rank descending. Group into three buckets:

• This week: top 5 items. Homepage test usually leads.
• Next 30 days: next 10 items. Ads, email, product pages.
• Backlog: long-tail. New content, PR rollout, investor deck refresh.

No P0/P1/P2 labels. Use bucket names only.

Modes

The audit supports four modes. Default is Full on Path A.

State the chosen mode at kick-off. Record in the report header.

Handoff to execution

The positioning audit is the decision artefact. Execution happens in three places:

1. actions.md in the client folder. The "This week" items merge in as P0/P1 per the existing cap logic. Homepage test usually enters as P0.
2. Downstream playbooks. The homepage-copy rewrite uses the copywriting skill seeded with Option B's primary promises. The ad variations use the ad-variations playbook. The email welcome series uses the email-sequence skill.
3. Weekly review cadence. The weekly marketing review checks leading indicators (ad CTR delta, homepage bounce rate, welcome-flow open rate) in week 1, mid-indicators in week 4, lagging indicators in week 12. If lagging indicators fail to move by week 12, return to Section 6 and tune.

The audit is one artefact. The positioning change is a quarter of disciplined rollout.

Voice conventions (strict, applied end-to-end)

• No em-dashes. Period, comma, or colon.
• No negative parallelisms. State the positive claim alone.
• Four-word sentence floor.
• Spelling: "ecommerce" one word.
• No AI clichés.
• ICP language discipline: use the buyer's exact phrases in Sections 5, 6, 7, 8. Founder phrases appear only where traceable to the Section 1 capture.
• No unearned certainty. The three options are hypotheses, the audit ranks them but the founder runs the test.

Output formatting: bullets over paragraphs (mandatory for every report and summary)

When this playbook generates a report, executive summary, brief, or any deliverable with multiple findings or actions, the default output structure is bullet lists, not prose paragraphs.

• Any time the text enumerates three or more items (findings, fixes, gaps, recommendations, criteria, things-they-do-better) those items become separate bullets. Do not bury enumerated points inside a paragraph.
• Reserve paragraphs for genuinely connected reasoning where prose flow carries the argument. Cap those paragraphs at four sentences.
• Multi-clause sentences with semi-colons or commas separating discrete points are a signal to break into bullets.
• The canonical pattern is: heading + one lead-in sentence + bullet list. The lead-in frames the list; the bullets carry the load.
• Reports ship scannable. The reader skims on a phone between meetings.

Run a paragraph-density check before commit: any paragraph longer than four sentences, or containing three or more enumerated items inside prose, is a candidate for bullet conversion. Convert it.

Applies to: full reports, executive summaries, email overviews, weekly reviews, dashboards, action-item lists, and any narrative deliverable. Does not apply to: opening framing paragraphs (one paragraph allowed), conclusion paragraphs (one or two allowed), or quoted founder/buyer voice.

Reference material

• audience-insights-master-prompt.md: canonical audience research prompt, runs as Section 3.
• gtm-audit.md: parent audit. Section 2 of that spec is the predecessor of this playbook's Section 1.

Acceptance criteria

• Runs end-to-end from a single Claude Code prompt. No manual copy-paste between steps.
• Detects the correct path (A or B) and adapts sections accordingly.
• Runs the canonical audience prompt as Section 3 by reference to file path. Does not rewrite the prompt inline.
• Produces three positioning options with full messaging systems. Each option traceable to Section 1, 3, or 5 evidence.
• Names one recommended option with reasoning in Section 7.
• Ships a single consolidated markdown file plus the pandoc Word doc.
• Writes the 15-quote audience appendix and the per-competitor capture files to the supporting data folder.
• Voice enforcement gate run before write. Zero banned-pattern matches in final output.
• Merges the top 5 "This week" items into actions.md with the existing priority caps honoured.
• Output folder pattern: {client_folder}/docs/positioning-audit/{YYYY-MM-DD}/.