--- name: tidb-gtm-benchmark description: >- Analyzes competitor or industry-leader GTM content (blog posts, case studies, launch announcements, social copy, landing pages) to extract PMM writing patterns, messaging techniques, voice principles, and proof point strategies. Benchmarks TiDB's current GTM approach against the analyzed corpus and produces a prioritized gap analysis with skill-specific update recommendations. Use when a competitor or category-adjacent company is gaining developer mindshare, repositions, or when the GTM team wants an external quality audit. Not for feature comparison (use tidb-competitive-map) or messaging creation (use tidb-messaging-platform). --- # TiDB GTM benchmark **Role:** GTM Intelligence Analyst. You read competitor or industry-leader content the way a senior PMM would — not to copy, but to understand *why it works* and extract the underlying principles that can make TiDB's GTM content more credible, more resonant, and more effective. The output is not admiration; it is a structured improvement roadmap for the tidb-gtm-skills pack. --- ## When to run this skill **Run when:** - A direct competitor launches a new content series or repositions (e.g., PlanetScale, CockroachDB, Neon publish a wave of developer content) - A category-adjacent company is gaining developer mindshare in a space TiDB wants to own (e.g., Neon in the developer database space, Turso in edge/embedded SQL) - The GTM team wants to audit TiDB's content quality against an external standard - A new market entrant appears with unusually effective PMM execution - Quarterly PMM review: "are we keeping up with industry content standards?" **Do NOT run for:** - Feature-level competitive comparisons → use `tidb-competitive-map` - Building TiDB's messaging → use `tidb-messaging-platform` - Reviewing a specific TiDB piece → use `tidb-content-review` --- ## Reference corpus: pre-analyzed benchmarks These companies have been analyzed and their patterns are encoded in the tidb-gtm-skills pack. They serve as the quality baseline. ### Neon (Postgres serverless, developer-first) — ANALYZED 2026-04 **PMM Lead:** Carlota Soto **Corpus:** 10 blog posts (partner spotlights, customer stories, technical tutorials, thought leadership) **Benchmark standards extracted:** 1. **Problem naming precision** — coins a memorable, specific name for every problem before introducing the solution ("staging drift," "the mismatch between ephemeral infrastructure and permanent auth") 2. **Fair comparison arc** — shows alternatives honestly (legitimate strengths first, then specific limitations) before introducing Neon's approach; never strawmans 3. **Feature-to-use-case translation** — no naked feature mentions; every capability immediately connected to a named developer workflow 4. **Enabler framing** — Neon is never the hero; the partner/customer is the hero; Neon is the enabling infrastructure primitive 5. **Implicit CTA philosophy** — developer-facing content ends with proof-of-adoption metrics, not hard conversion asks; CTAs are progressive (lowest friction first) when present 6. **Trusted-source proof chain** — all proof attributed to named companies, named roles (co-founder, principal engineer), and specific metrics 7. **Multi-layer pain architecture** — problems structured as operational + technical/performance + compliance/cost (three dimensions) 8. **The reframing move** — at problem-to-solution turn: "Instead of [old mental model], treat [concept] as [new mental model]" 9. **Philosophical/structural critique** — critiques architectural assumptions, not named competitors ("inherited assumptions about permanent infrastructure") 10. **Category-language investment** — consistent use of owned terms across posts ("agent-native database primitive," "database branching," "preview environments") 11. **Ecosystem narrative** — partnership content structured as three-platform story (each solves one layer; together complete) 12. **Scale-as-proof closing** — specific, understated adoption metrics close the story (not as headlines but as matter-of-fact evidence) **Content type taxonomy used:** - Thought leadership (infrastructure philosophy) - Technical tutorial (with fair comparison arc) - Partner spotlight (ecosystem positioning; customer as hero) - Customer story / case study (architecture + developer workflow) - Developer workflow feature (workflow-as-narrative) ### Industry sources referenced in tidb-gtm-skills design - **growth-skills** — SKILL.md workflow format, stage-gate review pattern - **gstack** — brain vs. tool separation (knowledge vs. workflow) - **gbrain** — knowledge base evolution model - **pm-skills** — ICP discipline, GTM motion framing - **impeccable / agent-gtm-skills** — multi-dimension review design, field playbook format --- ## Phase 1: Corpus collection **Input:** 5–15 pieces of the target company's GTM content. Aim for content diversity across types: - [ ] At least 2 blog posts (not product announcements) - [ ] At least 1 customer or partner story - [ ] At least 1 technical tutorial or how-to - [ ] Optional: landing page, social copy, launch announcement **Output — Content inventory:** ``` ## Corpus: [Company Name] — [Date analyzed] | # | URL / Title | Content type | Approx. length | Author (if known) | |---|-------------|-------------|----------------|-------------------| | 1 | [url] | [type] | [words] | [name/role] | ... ``` **Content type classification (use these labels):** - `thought-leadership` — industry POV, trend analysis, opinion essay - `technical-tutorial` — how-to with comparison or implementation walkthrough - `customer-story` — case study, architecture story, company-as-hero - `partner-spotlight` — ecosystem narrative, integration story - `developer-workflow` — feature explained through developer workflow - `launch-announcement` — product launch, feature release - `channel-copy` — social post, newsletter, short-form --- ## Phase 2: Per-piece structured analysis For each piece in the corpus, extract: ``` ### Analysis: [Title] **Type:** [content type label] **Length:** [words / approx. reading time] **Opening move:** How does the piece open? Does it name a specific problem? What exact language is used? (Quote the first 1-3 sentences verbatim) **Problem architecture:** How many pain layers? What types? (operational / technical / compliance / cost / emotional) Quote the key problem-framing sentences. **Alternative treatment:** Does the piece acknowledge alternatives? - Yes, honestly (legitimate strengths shown) / Yes, strawmanned / No alternatives mentioned Quote the relevant passage if present. **Feature-to-use-case translation:** Are features grounded in specific workflows, or listed nakedly? Give one example of a well-translated feature (or note its absence). **Proof point strategy:** - Source type: named company + named role / named company only / anonymous / no proof - Metric type: specific number / directional ("significant improvement") / none - Scale signal: yes/no (e.g., "400k branches/day") Quote 1-2 proof points verbatim. **Reframing move:** Is there a "Instead of X, think of Y" sentence or equivalent mental model shift? Quote if present; note if absent. **CTA approach:** - Implicit (proof as close, vision statement) - Progressive (lowest friction first) - Explicit (sign up, request demo) - Hard sell (urgency language, FOMO) Note the actual closing technique. **Voice markers:** List 3-5 distinctive phrases or vocabulary choices that characterize the author's voice. Note what the piece conspicuously avoids. **Technical depth calibration:** What technical background does this assume? (none / basic / mid-level practitioner / senior engineer) Does the piece use technical vocabulary as a trust signal? Give an example. ``` --- ## Phase 3: Pattern extraction After completing all per-piece analyses, synthesize across the full corpus. ``` ## Cross-corpus patterns: [Company Name] ### Repeating structural formulas List each formula with a name and 2-3 post examples: 1. [Formula name]: [description] — seen in posts #[X], #[X], #[X] ... ### Consistent voice vocabulary **Owned terms (used across 3+ pieces):** [list] **Power words (appear consistently):** [list] **Conspicuously avoided:** [list — what they never say that most companies do] ### Proof strategy philosophy [1-2 sentences: what counts as valid evidence for this company?] [Named sources only? Does scale matter more than outcome? First-party or third-party proof?] ### CTA philosophy [Implicit by default? Progressive? Hard-sell-averse? What do they rely on instead of CTAs?] ### Category terms being invested in [List the 3-5 terms they use consistently enough to "own" in the reader's mind] ### Negative space (what they never do) [List 3-5 things this company conspicuously avoids in its content — these are often as revealing as what they do] ``` --- ## Phase 4: TiDB benchmark comparison Using the patterns from Phase 3, score TiDB's current GTM approach against each dimension. ``` ## Gap analysis: [Company Name] vs TiDB GTM | Dimension | [Company] does | TiDB currently does | Gap level | |-----------|----------------|---------------------|-----------| | Problem naming | [specific observation] | [current TiDB practice] | High / Medium / Low / None | | Fair comparison arc | | | | | Feature-use-case grounding | | | | | Proof point credibility | | | | | Reframing technique | | | | | CTA calibration | | | | | Category term investment | | | | | Enabler framing | | | | | Philosophical vs. competitive critique | | | | | Technical depth calibration | | | | **Gap level definitions:** - High: TiDB is consistently doing the opposite or ignoring this dimension - Medium: TiDB does this sometimes but inconsistently - Low: TiDB does this well in some skills but not others - None: TiDB's approach is comparable or already stronger ``` --- ## Phase 5: Skill update recommendations For each High or Medium gap, produce a recommendation: ``` ## Recommendations: [Company Name] analysis ### Tier 1 — Blocking improvements (do first) These gaps actively undermine credibility or trust with the target audience. **Recommendation 1:** - Gap: [specific gap from Phase 4] - Affected skills: [tidb-skill-name(s)] - What to add: [specific language, template, or checklist item] - Example: [show a before/after or a template sentence] ... ### Tier 2 — Quality improvements These gaps reduce effectiveness but don't actively damage credibility. ... ### Tier 3 — Excellence refinements These gaps are the difference between "good" and "industry-leading" GTM content. ... ``` --- ## Phase 6: Category intelligence brief Document signals about the market that go beyond writing quality — this feeds `tidb-icp-synthesis` and `tidb-competitive-map`. ``` ## Category intelligence: [Company Name] — [Date] ### Market positioning signals What new positioning moves is this company making? What categories are they investing in? [2-3 observations with evidence] ### ICP signals What audience segments is this company targeting that overlap with TiDB's ICP? What pain points are they naming that TiDB should also be addressing? [2-3 observations] ### Messaging pivots Have they shifted tone, vocabulary, or value proposition framing since their last content cycle? [1-2 observations, or "no significant shift detected"] ### Gaps in their coverage (TiDB opportunity) What does this company not cover or address well? Where does TiDB have standing they don't? [2-3 genuine opportunities — not wishful thinking] ### Recommended actions 1. [Specific GTM action for TiDB, e.g., "publish a Fair Comparison post for HTAP vs ETL pipelines before Neon enters that space"] 2. ... ``` --- ## Output format The full skill run produces a single artifact: ``` # GTM Benchmark: [Company Name] — [Date] ## Executive summary [3-5 bullets: top patterns extracted, top gaps identified, highest-priority recommendations] ## Part 1: Corpus inventory [Phase 1 output] ## Part 2: Per-piece analysis [Phase 2 output] ## Part 3: Cross-corpus patterns [Phase 3 output] ## Part 4: Gap analysis [Phase 4 output] ## Part 5: Skill update recommendations [Phase 5 output] ## Part 6: Category intelligence [Phase 6 output] ``` Save as: `tidb-gtm-brain/playbooks/benchmark-[company]-[YYYY-MM].md` --- ## Quality bar for the analysis The benchmark analysis is only useful if it's honest. Apply the same standards to this analysis that TiDB applies to its own content: - Do not cherry-pick the competitor's best content and ignore weak pieces - Do not rate TiDB generously — accurate gaps are more valuable than flattering scores - The "negative space" section (what the company never does) is often the most instructive — take it seriously - Recommendations must be specific enough to implement in 30 minutes; "improve proof points" is not actionable; "add named company + role + metric to every claim in tidb-tech-story's Framework A" is --- ## Pairs with - `tidb-competitive-map` — for feature-level competitive analysis; this skill covers GTM/content patterns - `tidb-icp-synthesis` — category intelligence brief feeds new ICP signals - `tidb-content-review` — gaps identified here should inform Dimension scoring calibration - `tidb-tech-story` — most structural improvements will land here first - `tidb-brand-voice` — voice and vocabulary gaps feed into brand guidance updates - `@tidb-gtm-brain/playbooks/` — output saved here for team reference