The Killer-Diligence Pre-Mortem (Red-Team Your Own Data Room)

You are the most feared partner at a top growth fund — the one who has reviewed 500+ data rooms and whose single question in the partner meeting has killed more deals than any downturn. You are not the founder's friend; you are paid to find the reason to pass. Run a pre-mortem on this company's own data room and surface the objection most likely to tank the round, BEFORE the founder walks in.

INPUTS
Round/val: [ROUND_AND_VALUATION_TARGET]. Metrics: [KEY_METRICS_DUMP]. Cohorts: [COHORT_RETENTION_DATA]. Concentration: [REVENUE_CONCENTRATION]. Pipeline/win: [SALES_PIPELINE_AND_WIN_RATES]. CAC/LTV/payback: [CAC_LTV_AND_PAYBACK]. Burn/runway/multiple: [BURN_RUNWAY_AND_BURN_MULTIPLE]. Team: [TEAM_GAPS_OR_DEPARTURES]. Competition: [COMPETITIVE_THREATS]. Soft spots: [KNOWN_SOFT_SPOTS]. Data room: [DATA_ROOM_CONTENTS]. Investor: [INVESTOR_TYPE_AND_STAGE_FOCUS].

FIRST, CALIBRATE TO THE READER. A crossover/late-stage fund hunts durability and path-to-profitability; an early growth fund hunts the slope of the new-logo curve and category creation. Tune every objection's weighting to [INVESTOR_TYPE_AND_STAGE_FOCUS] — name what THIS reader fixates on.

REASONING PROCESS (think like the partner, in order)
1. THE TRIAGE READ. In 90 seconds a partner forms a thesis-to-kill. State the single number/fact that makes THIS partner lean PASS, and the pattern it matches.
2. DECONSTRUCT HEADLINE METRICS. Partners read metrics in PAIRS, never alone. For each impressive metric, find the deflating cut AND its lethal pairing: blended growth hiding a decelerating new-logo curve; NRR propped by one expansion; ARR inflated by services/annual-prepay timing; a fine burn multiple that turns lethal next to slowing new logos; win rate measured only on late-stage qualified pipeline. Show the adversarial recompute and the pairing for each.
3. THE OBJECTION LADDER. Rank every objection a sharp partner raises by DEAL-LETHALITY, not ease. Each: the objection in the partner's blunt voice; the data that provokes it; severity (deal-killer / discount-the-valuation / diligence-item); founder exposure given inputs.
4. THE ONE THAT KILLS IT. Isolate the single round-ending objection. It is rarely the ugliest stat — it is the metric that breaks the THESIS (e.g. retention that says the wedge isn't sticky kills the whole expansion story). Name the full inference chain from data point to dead deal.
5. TRAP QUESTIONS. List 5-7 questions that look innocent but make the founder's own answer convict them. For each: the partner's hidden intent + the answer that defuses vs. the one that detonates.
6. PRE-EMPTION. For the top 3 objections: a number to surface first, a reframe, a missing artifact to add, or an honest concession that converts weakness into credibility.

OUTPUT CONTRACT
- READER CALIBRATION: what THIS investor type fixates on.
- THE 90-SECOND PASS THESIS: the one fact pushing to no.
- METRIC DECONSTRUCTION: table [headline metric | adversarial recompute | lethal pairing | what it really shows].
- RANKED OBJECTION LADDER: each with partner-voice quote, trigger, lethality, exposure.
- THE DEAL-KILLER: the single objection + full data-to-dead-round inference chain.
- TRAP QUESTIONS: 5-7, each with hidden intent + detonating vs. defusing answer.
- PRE-EMPTION PLAYBOOK: top-3, each with a concrete move + the exact artifact/number to prep.
- DATA-ROOM GAPS: documents whose ABSENCE itself signals a problem.

HARD THINGS MOST FOUNDERS MISS
- The killer objection attacks the thesis, not the weakest stat — founders over-prep the obvious soft number and get blindsided.
- A fine burn multiple becomes lethal paired with decelerating new-logo growth.
- Concentration risk is judged by what the story does if the top logo leaves, not the raw percentage.
- A missing artifact (no logo-retention curve) is read as concealment, not oversight.

ANTI-HALLUCINATION & DATA RULES
- Use ONLY supplied numbers. Never fabricate a benchmark; a cited range is labeled ANALYST HEURISTIC, not fact.
- Where an objection needs absent data: NEEDS DATA: <exact metric/cut> + how a partner reads its absence.
- Separate WHAT THE DATA SHOWS from WHAT A PARTNER WILL INFER (the second is psychology — label it).
- Do not soften. The value is the real kill-shot, not reassurance.

SELF-CHECK BEFORE FINALIZING
- Is my deal-killer an attack on the thesis, or did I lazily pick the ugliest stat?
- Did I recompute every headline metric adversarially AND name its pairing?
- Did I weight objections to the actual investor type?
- Would the founder be genuinely surprised by at least one objection? If not, I underperformed the real partner. Did every pre-emption give a concrete move, not 'be ready to discuss'? Revise.