---
name: tidb-messaging-review
description: >-
  Stage-gate review of TiDB's messaging platform and benchmark narrative before
  they drive external content creation. Evaluates whether all claims are
  evidence-backed, banned phrases are cleared, segment messaging aligns with ICP
  cards, benchmark methodology is disclosure-complete, and legal/compliance risks
  are addressed. Outputs Pass / Revise with Tier 1 (blocking) and Tier 2
  (advisory) issues. No content creation should begin without a messaging review
  pass.
---

# TiDB messaging review

**Role:** Messaging Gate. You are reviewing the messaging platform and benchmark narrative — the source documents that will drive everything from blog posts to sales decks to LinkedIn posts. A weak messaging platform multiplies downstream errors across every piece of content. Getting it right here saves revision cycles across the entire content program.

---

## Who runs this review

| Reviewer | What they own |
|----------|--------------|
| **PMM lead** | Overall messaging logic, ICP alignment, competitive accuracy |
| **PM** | Technical claim accuracy, capability scope, version/tier correctness |
| **Legal / Compliance** | Competitive naming, benchmark claims, customer reference permissions |
| **Brand** | Language standards, banned phrases, voice consistency |

Minimum quorum: PMM lead + PM required. Legal required if benchmarks or named competitors are in scope. Customer references require explicit customer approval documented before use.

---

## Dimension 1: Evidence quality

Every external claim must trace to a proof point. Score each claim:

| Evidence type | Score | Status |
|--------------|-------|--------|
| Named public customer case (publicly approved) | Strong | ✓ Use |
| Anonymous validated case (design partner confirmed) | Acceptable | ✓ Use with permission |
| Internal benchmark with published methodology | Acceptable | ✓ Use with methodology disclosure |
| Internal test without published methodology | Weak | ⚠ Do not use for external claims |
| "We believe" / assumed claim | None | ✗ Block |

**Review questions:**
- [ ] Every quantified claim ("10x", "5×") traces to a specific proof point
- [ ] Every customer reference has written approval for the specific use
- [ ] No claims reference roadmap features as if they are GA (check with PM)
- [ ] Version and tier are specified for all capability claims

---

## Dimension 2: Banned phrases check

Run the messaging platform through the full banned phrases list from `tidb-brand-voice` and `tidb-messaging-platform`.

**Tier 1 (block — must remove):**
- [ ] No "fully MySQL compatible" (must say "MySQL protocol compatible" with link)
- [ ] No "zero downtime" without specification
- [ ] No benchmark claims without methodology
- [ ] No "AI-ready" or "next-generation AI" without specific capability
- [ ] No "seamless", "frictionless", "effortless"
- [ ] No "industry-leading" or "best-in-class" without independent validation
- [ ] No "scales infinitely"

**Tier 2 (advisory — improve if possible):**
- [ ] "Powerful" without proof → replace with specific capability
- [ ] "High performance" without numbers → replace with measured result
- [ ] "HTAP" without explanation for non-technical audience → add one-sentence explanation
- [ ] "Distributed SQL" without explaining why it matters → add customer pain context

---

## Dimension 3: ICP alignment

| Review question | Pass criteria | Issue criteria |
|----------------|--------------|---------------|
| Does each segment's value prop match the ICP card's deciding language? | Deciding language from ICP card appears in value prop | Value props use marketing language the customer doesn't use |
| Is the primary segment prioritization consistent with the GTM brief? | Same segment priority in both documents | Messaging platform targets different segment than GTM brief |
| Are "out of ICP" users explicitly excluded somewhere in the platform? | Platform notes who this is NOT for | No exclusions → Sales will waste cycles on wrong segments |

---

## Dimension 4: Benchmark methodology disclosure

For any benchmark content in the messaging platform:

| Check | Pass | Block |
|-------|------|-------|
| Node count and instance type specified | Yes | Not specified |
| TiDB version specified | Yes | Version omitted |
| Workload generator and config specified | Yes | "Standard benchmark" without config |
| Comparison baseline named with version | Yes | "Compared to MySQL" without version |
| "What was NOT tested" section present | Yes | Scope not limited |
| TiFlash enabled/disabled stated for HTAP claims | Yes | Not specified |
| Raw data table available | Yes (link) | Numbers only, no table |

---

## Dimension 5: Legal and compliance checks

| Check | Pass criteria | Flag criteria |
|-------|--------------|--------------|
| Competitor names used fairly | Factual, not strawmen; competitor's own docs cited | Editorial characterizations of competitor weaknesses |
| Customer references approved | Written approval on file | Assumed approval; "they won't mind" |
| Benchmark claims reviewed | Legal or PM has cleared benchmark methodology for publication | Not reviewed |
| China-specific claims reviewed | Cross-regional/data sovereignty claims checked for accuracy in both jurisdictions | Untested claims about China regulations |
| Pricing claims | Accurate at publication date; include "subject to change" | Stale pricing |

---

## Review output format

```markdown
## Messaging Review — [Document name] — [Date]

### Verdict: PASS / REVISE

### Tier 1 Issues (blocking — must fix before content creation begins)
- [Dimension]: [specific issue in the messaging document] → Required fix: [what to change]

### Tier 2 Issues (advisory — fix before major campaign launch)
- [Dimension]: [specific issue] → Suggested improvement: [what to change]

### Sign-offs
- PMM lead: [name] — [PASS / NOT READY]
- PM: [name] — [PASS / NOT READY]
- Legal: [name / N/A] — [PASS / NOT REVIEWED]
- Brand: [name / N/A] — [PASS / NOT READY]

### Scope of this review
This review covers: [document names and versions reviewed]
This review does NOT cover: [e.g., "does not cover individual content pieces downstream"]
```

---

## What a PASS unlocks

A messaging review PASS means:
- Content creation (Stage 4) can begin using this messaging platform
- Sales enablement team can brief sales reps using these value props and proof points
- `tidb-channel-copy` can produce social copy drawing from approved taglines
- `tidb-tech-story` and `tidb-thought-leadership` can reference approved proof points

A PASS on messaging does NOT mean:
- Individual content pieces are auto-approved (they still go through `tidb-content-review`)
- The messaging platform is permanent (schedule review after every product release and ICP update)

---

## Pairs with

- `tidb-messaging-platform` — the document being reviewed
- `tidb-benchmark-narrative` — benchmark content in scope of this review
- `tidb-brand-voice` — banned phrases list and voice standards
- `tidb-strategy-review` — must PASS before messaging review is scheduled
- `tidb-content-review` — downstream review gate for individual content pieces
- `@tidb-gtm-brain/playbooks/messaging.md` — stores approved messaging snapshot after review