Skip to content

borghei/Claude-Skills / project-management/execution/prfaq

Write a Working Backwards press release with internal and external FAQs before anything is built. Use when a funding or go/no-go decision needs the customer outcome made concrete first; write the PRD after it passes.

references/red-flags.md

View on GitHub
# Red Flags: PR/FAQ

> Common ways this skill's output goes wrong — concrete examples, why they're bad, and how to fix them.

## How to use this document

Scan the press release, internal FAQ, and external FAQ before sending to the exec sponsor. Each red flag shows the *bad* version next to the *good* version, anchored to Amazon's Working Backwards method.

---

## Red Flag 1: PR sounds internal, not external

**Symptom.** Headline reads "ProjectCobra launches Q3 deliverable" — uses an internal codename, internal milestone language, no customer-facing benefit.

**Why it's bad.** A press release is supposed to read as if a real journalist wrote it. Internal-sounding releases reveal the team is writing for themselves; the customer perspective wasn't tested. If a journalist would not write this, customers will not care about it.

**Bad example:**
> "Headline: 'ProjectCobra: Cross-Functional Delivery of Q3 Roadmap Milestones.'
> Sub-headline: 'Engineering, Design, and PMM partner to release new capabilities to GA.'"

**Good example:**
> "Headline: 'Acme launches auto-research for sales teams — cuts pre-call prep from 90 minutes to 10.'
> Sub-headline: 'Available now to mid-market sales leaders; designed for teams running 5+ pipeline reviews per week.'"

**How to catch it.** Hand the PR to a non-team-member. Ask: "could this run in TechCrunch as written?" If no, it sounds internal.

---

## Red Flag 2: Missing customer quote (or fake-sounding quote)

**Symptom.** No customer quote. Or the quote is "Acme has been a game-changer for our entire industry."

**Why it's bad.** The customer quote is the credibility anchor of a press release. Missing it leaves the PR all-company-no-customer. Fake-sounding quotes ("game-changer", "revolutionary") signal that the team can't articulate a real outcome — readers translate this immediately.

**Bad example:**
> "(No customer quote — section skipped.)" or "'This product has revolutionized our entire approach to business and is truly the next generation of innovation.' — Made-up VP."

**Good example:**
> "'My team was spending 90 minutes per opportunity researching prospects. Last week, we ran 14 pipeline reviews and each rep showed up with a 2-page talking points doc that took me 10 minutes to generate. We have our Friday afternoons back.' — Sarah Chen, VP Sales at TechCorp (a beta participant)."

**How to catch it.** Read the quote aloud. Does it describe a specific outcome with a number? If it uses words like "revolutionary" or "game-changer", rewrite.

---

## Red Flag 3: Internal FAQ has 6 easy questions

**Symptom.** Internal FAQ skips the hard ones: "How much will this cost?", "What if it fails?", "Why now and not 18 months ago?", "Who is the team committed to it 12 months from now?".

**Why it's bad.** The internal FAQ is where weak ideas show themselves. If the PM has not stress-tested the idea against hard questions, the funding review will surface them — and the team will look unprepared. Soft-question FAQs are a tell: the author dodged the work.

**Bad example:**
> "Internal FAQ (6 questions):
> Q1: What is this product?
> Q2: Who is it for?
> Q3: When will it launch?
> Q4: How do we use it internally?
> Q5: What does it look like?
> Q6: Is everyone excited about this?"

**Good example:**
> "Internal FAQ (16 questions across 9 required categories):
> 1. Customer demand: How many customers experience the problem, evidenced how?
> 2. Business model: How do we make money? What is the 3-year P&L?
> 3. Strategic fit: Why us, why now, why does this compound?
> 4. Competition: Who else? Why will we win?
> 5. Technical feasibility: What is the riskiest technical bet, and how do we retire it?
> 6. Operational: Who supports it; what happens when it breaks?
> 7. Legal/privacy: What data; what regulations apply?
> 8. Risk/failure: What is the worst case; what is the kill criterion?
> 9. Scope: What are we explicitly NOT doing in v1?
> Each answer 1-3 paragraphs, quantified where possible, with cited sources."

**How to catch it.** Count FAQ Q&A pairs. < 10 = author skipped the hard ones. Apply the 5-readers rule.

---

## Red Flag 4: Buzzword stuffing

**Symptom.** PR uses "next-generation", "AI-powered", "revolutionary", "seamless", "transformative", "cutting-edge" in the first paragraph.

**Why it's bad.** Buzzwords substitute energy for substance. They are the prose equivalent of stock photos — placeholder content. They also age badly; "AI-powered" in 2026 reads like "cloud-based" in 2016.

**Bad example:**
> "Today, Acme launches a revolutionary, next-generation, AI-powered, seamless workflow assistant that transforms how teams work."

**Good example:**
> "Today, Acme launches auto-research — a Claude Sonnet-powered tool that generates 2-page customer talking points in under 30 seconds, cutting pre-call prep from 90 minutes to 10."

**How to catch it.** Search the PR for: revolutionary, next-generation, AI-powered (without specifics), seamless, transformative, cutting-edge, world-class, best-in-class. Strike each unless it adds information.

---

## Red Flag 5: Vague customer ("for businesses")

**Symptom.** Sub-headline says "for businesses" or "for everyone" or "for modern teams".

**Why it's bad.** "Businesses" is not a customer. Without a specific persona, the team cannot design for them; the message cannot resonate; the PR sounds like enterprise marketing template. Specificity is what makes a customer recognize themselves in the message.

**Bad example:**
> "Acme launches auto-research for businesses everywhere."

**Good example:**
> "Acme launches auto-research for B2B sales leaders at companies with 50-500 reps who run 5+ pipeline reviews per week."

**How to catch it.** Sub-headline names a persona at the role + company-stage + behavior level? If "businesses" or "users" or "teams" alone, too vague.

---

## Red Flag 6: Magnitude missing

**Symptom.** PR says "saves time" or "improves efficiency" or "boosts productivity" without numbers.

**Why it's bad.** Without a quantified outcome, the PR is generic. "Saves time" could mean 30 seconds or 30 hours. Customers and journalists scan for numbers; pages without numbers don't get clicked. Magnitude is what makes the value concrete.

**Bad example:**
> "Auto-research saves time and helps sales teams be more efficient."

**Good example:**
> "Auto-research cuts pre-call prep from 90 minutes to 10 minutes per opportunity — an 89% reduction. For a 10-rep team running 50 opportunities per week, that's 65 hours per week reallocated from research to selling."

**How to catch it.** Count numbers in the PR. < 3 quantified claims in the first half = magnitude missing.

---

## Red Flag 7: PR/FAQ written but never revised after kickoff

**Symptom.** PR/FAQ approved in March. It's now August; product is mid-build. PR/FAQ has not been touched since approval. Original answers no longer reflect reality (assumptions validated/invalidated, pricing decisions changed).

**Why it's bad.** PR/FAQ is supposed to evolve through funding review, beta, and launch. A static PR/FAQ becomes a fossil — referenced by new joiners who get a stale picture. Stakeholders read it and expect what no longer ships.

**Bad example:**
> "PR/FAQ v1, dated 2026-03-12. Approved. Last touched March 14. (Reality in August: assumption 3 was invalidated; pricing changed from per-seat to per-conversation; legal added a restriction on healthcare verticals.)"

**Good example:**
> "PR/FAQ v1.4 dated 2026-08-21. Changelog:
> • v1.4 (2026-08-21): updated pricing FAQ from per-seat to per-conversation after beta data.
> • v1.3 (2026-07-15): added healthcare-vertical exclusion per Legal.
> • v1.2 (2026-06-10): removed claim that needed assumption 3 (invalidated in concept test).
> • v1.1 (2026-04-05): expanded external FAQ to address sales objections from internal review.
> Re-read at each milestone (funding, beta launch, GA)."

**How to catch it.** Look at the last-modified date. If > 6 weeks old and the project has progressed, the doc is stale.

---

## Red Flag 8: External FAQ longer than internal FAQ

**Symptom.** External FAQ has 25 questions. Internal FAQ has 6. The PM is over-explaining to customers and under-stressing to execs.

**Why it's bad.** This typically signals that buyer questions are being mixed with reviewer questions. Internal FAQ should be 2-4 pages of hard questions; external FAQ 1-2 pages of buyer-facing answers. Reversal means the PM is preparing for customer conversations more thoroughly than for funding review.

**Bad example:**
> "Internal FAQ: 6 easy questions, 1.5 pages. External FAQ: 25 questions, 5 pages."

**Good example:**
> "Internal FAQ: 16 Q&A, 3.5 pages across the 9 required categories.
> External FAQ: 9 Q&A, 1.5 pages covering: What is it / Who is it for / How is it different / Cost / Get started / Integrations / Privacy / Without prerequisite / Cancel & refund.
> Audience-specific test: anything an exec sponsor would ask → internal; anything a customer would ask → external."

**How to catch it.** Count Q&A in each. If external > internal, the audiences are mixed.

---

## Red Flag 9: Solving for slide-deck reviewers

**Symptom.** PR/FAQ has bullets, headers, dotted lines, sub-sections, "Key takeaways" boxes. It reads like a slide deck.

**Why it's bad.** Amazon's Working Backwards discipline is *prose*, not slides. Bullets let the author skip the connective tissue (the reasoning). Forcing prose forces clarity. A PR/FAQ that looks like slides indicates the team has not done the prose work.

**Bad example:**
> "Internal FAQ Q3: What is the business model?
> Key Highlights:
> • Subscription
> • SaaS
> • Per-seat
> • Enterprise + SMB
> Important Considerations:
> • Pricing
> • Tiers"

**Good example:**
> "Internal FAQ Q3: What is the business model?
> Auto-research is a per-conversation usage-based add-on for our existing platform. Pricing: $0.30 per conversation, billed monthly. Target gross margin: 68% at scale, accounting for $0.097 of inference cost per conversation plus 4% support load.
> Why per-conversation vs per-seat: research with 22 customers showed that sales-team usage is concentrated in a few power users (top 20% of reps generate 60% of conversations). Per-seat overpriced casual users and underpriced power users; per-conversation aligns price with value.
> Cannibalization: minimal — the manual research workflow we replace is internal time, not paid software. We expect 30% of customers to expand spend; 70% net new revenue."

**How to catch it.** Skim the document. Are answers prose or bullet lists? Bullets indicate the team has not connected the reasoning.

---

## Red Flag 10: Press Release Test never run

**Symptom.** PR/FAQ goes to the exec review meeting. Three readers there see it for the first time. PM defends design choices live.

**Why it's bad.** The Press Release Test (one outside reader before exec review) catches the obvious gaps before the high-stakes audience does. Skipping it means execs find the gaps; the meeting becomes a critique session instead of a funding discussion.

**Bad example:**
> "PR/FAQ draft completed Monday. Exec review Friday. Outside readers consulted: zero."

**Good example:**
> "Press Release Test sequence (before exec review):
> Reader 1 (an engineer not on the team): 'who is this for?' — answers correctly.
> Reader 2 (a designer not on the team): 'what problem does it solve?' — answers in plain language.
> Reader 3 (someone external — friend at another company): 'would you click Learn More?' — yes, would.
> Three iterations of revisions based on their questions before exec review. By the time execs read it, the basic clarity questions are answered."

**How to catch it.** Ask: "who has read this draft outside the team?" If "nobody yet", the test hasn't run.

---

## Red Flag 11: "What we will NOT do in v1" missing

**Symptom.** Internal FAQ enumerates everything v1 will do. Does not enumerate what v1 will deliberately not do.

**Why it's bad.** Scope creep starts at funding review. Every exec reads the PR and imagines a feature that's not there. Without an explicit "NOT" list, those imagined features become commitments. The team can't ship the imagined version; trust erodes.

**Bad example:**
> "Internal FAQ:
> Q: What does v1 do?
> A: [comprehensive list of 12 features]."

**Good example:**
> "Internal FAQ:
> Q: What does v1 do?
> A: [9 features in v1 — list].
> Q: What are we explicitly NOT doing in v1?
> A: (1) No mobile interface; web-only. (2) No write-back to CRM; read-only. (3) No EU-data-residency support; US customers only. (4) No multi-language; English only. (5) No fine-tuned model option; we ship with the base. Each is a deliberate v1 cut, documented for stakeholder alignment. We may revisit in v1.1 based on demand signals."

**How to catch it.** Search the FAQ for "not", "won't", "deliberately deferred". If absent, the boundary is undefined.

---

## Red Flag 12: PR/FAQ ignored by the build team

**Symptom.** PR/FAQ exists. The PRD references it once in Section 1. Engineering builds against the PRD detail without re-reading the PR. Six weeks later, scope has drifted from the PR's promise.

**Why it's bad.** The PR/FAQ is supposed to be the design constraint, not the prologue. Once engineering stops referencing it, the PR/FAQ becomes a sunk cost. The team builds something different than the press release promised; launch communications and product diverge.

**Bad example:**
> "PR/FAQ approved March 1. PRD written March 5 referencing PR/FAQ in Section 1 ('see PRFAQ-AutoResearch-2026-03-01.md'). Build kickoff March 12. No subsequent reference. Scope drifts.

**Good example:**
> "PR/FAQ is the design constraint. Every sprint review opens with: 'does what we built this sprint advance the press release outcome?' If the team starts building a feature not implied by the PR, that feature requires justification: either re-write the PR to include it, or cut the feature. The PR is pinned in the team's planning tool. The PR is what we report against, not feature delivery."

**How to catch it.** Ask the build team: "what does the PR/FAQ promise that you are working toward?" If they hesitate, the PR/FAQ is decorative.

---

## Red Flag Quick Reference

| # | Anti-pattern | One-line check |
|---|---|---|
| 1 | PR sounds internal | Would a journalist write this headline? |
| 2 | Missing or fake customer quote | Does the quote describe a specific outcome with a number? |
| 3 | Internal FAQ has 6 easy questions | < 10 Q&A across the 9 required categories? |
| 4 | Buzzword stuffing | Search for: revolutionary, next-gen, AI-powered, seamless |
| 5 | Vague customer | "Businesses" or "everyone" or "teams"? |
| 6 | Magnitude missing | < 3 quantified claims in the first half? |
| 7 | Never revised after kickoff | Last-modified date > 6 weeks ago? |
| 8 | External longer than internal | Q&A count comparison |
| 9 | Slide-deck format | Prose or bullets? |
| 10 | Press Release Test skipped | Who has read it outside the team? |
| 11 | "Won't do in v1" missing | Search FAQ for "not", "won't", "deferred" |
| 12 | Ignored by build team | Does the team reference the PR mid-build? |

## Related Reading

- SKILL.md Troubleshooting
- references/working-backwards-guide.md
- `create-prd/` (PR/FAQ → PRD)
- `discovery/identify-assumptions/` (assumptions populate Internal FAQ "what we don't know")
- `discovery/pre-mortem/` (tiger risks populate Internal FAQ "failure modes")
- `brainstorm-okrs/` (PR's outcomes become OKR candidates)
- `outcome-roadmap/` (PR's v1 scope informs roadmap horizons)
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall