Sales & Business applications: diagnosis and defense

Why B2B sales is the ecosystem most exposed to Brandolini

Three characteristics make the B2B sales cycle the prime attack zone for Brandolini's Law:

1. Long cycle with multiple touchpoints: 6 to 18 months, 7+ stakeholders involved → that many opportunities to introduce bullshit that contaminates the entire pipeline.
2. Information asymmetry: the prospect hears 100 sources (competitors, analysts, ex-employees, Reddit, community Slacks) for 1 official source.
3. Decision delegation: the final decision often depends on an executive sponsor who never spoke directly with your team — they believe what they're told about you.

A single false claim entered in W2 of the cycle can close the deal in W6, without you ever knowing where it came from.

Four families of Brandolini attacks in sales

1. Competitive FUD (Fear, Uncertainty, Doubt)

Definition: a competitor spreads a negative claim about you, partly true, partly false, not quickly verifiable by the prospect.

Diagnosis:

• Anomalous spike of similar objections in W3-W4 of the cycle
• Identical vocabulary across multiple prospects (signature of a competing battlecard)
• Repeated mention of the same article or tweet

Tactical antidote:

• Pre-positioned battlecard with 3 quantitative proofs per typical accusation
• "If you're told X, here's Y, Z, and W" — memorizable by the SDR in 30 seconds
• Proactive outreach to the critical stakeholder before they receive the attack

2. The prospect's bullshit objection

Definition: the prospect raises an objection based on a false but anchored belief. It's not an attack — it's a costly misunderstanding.

Real-world examples:

• "You're US, so our data isn't GDPR-compliant" (false: EU data residency available)
• "Your APIs are REST, we need gRPC" (false: gRPC supported since 2023)
• "You bill per seat, we have 500 users it'll explode" (false: usage pricing available)

Diagnosis:

• Objection that recurs ≥ 3 times across different prospects
• Objection that appears BEFORE any serious technical conversation
• Objection containing a verifiable factual claim (not a subjective taste)

Antidote:

• Identify the TOP 5 recurring bullshit objections via quarterly review
• Pre-bunking in the first demo ("many people think X, in fact Y, here's the proof")
• Web page "What people get wrong about us" cited in SDR emails

3. The fake review (B2B & B2C)

Definition: negative review posted by a non-customer (competitor, ex-employee, troll), often without real product experience.

Economic mechanics:

• Production cost: 5 minutes, $0 (or $5 via a review farm)
• CTR impact: -5 to -30% depending on average rating and channel
• Defense cost: flag, platform support, legal disputes (G2, Trustpilot, Capterra)

Diagnosis (weak signals):

• 1-2★ review from a recent account with no history
• Vocabulary mentioning the competitor by name
• Spike of negative reviews uncorrelated with any product event
• Author name not matching any CRM customer

Antidote:

• Mandatory workflow: immediate flag to platform + evidence preservation
• Short, factual, professional public response (never emotional)
• Programmatic mobilization of real positive reviews (NPS detractors → satisfied → review invitation)
• If coordinated attack: public LinkedIn disclosure by CEO (humanizes, contextualizes)

4. LLM hallucination in customer support

Definition: an LLM (ChatGPT, Claude, Gemini) responds to the customer with a false claim about your product that you neither wrote nor validated.

Typical cases:

• "According to ChatGPT, your Pro plan includes SSO" — no, that's the Enterprise plan
• "Claude told me you have a webhooks API" — no, it's not yet released
• "ChatGPT says your annual price is $99/mo" — outdated by 18 months

Operational cost:

• 5 to 45 min of support per incident
• Churn risk if customer feels misled
• No direct remedy: you don't control either the LLM or hallucinations

Antidote:

• Page "What AI assistants get wrong about us" — directly linkable by support
• Clear pricing section in a single place with last-update date
• llms.txt and robots.txt configured so recent LLMs read your true data
• Detection of recurring hallucinations via NPS "how did you learn X?"

The decision matrix: respond or ignore?

Not all bullshit deserves defense. Here's a decision matrix:

This matrix avoids two symmetric errors:

1. Over-defense: responding to every attack → team burnout, signal of weakness
2. Under-defense: ignoring a viral attack quickly → loss of narrative control

The hidden cost: informational debt

Every untreated bullshit accumulates. This is the informational debt, narrative equivalent of technical debt:

graph LR
    A[Bullshit published] --> B{Handled<br/>in 24h?}
    B -->|yes| C[Indexed<br/>1-2 days]
    B -->|no| D[Persists 6-18 months]
    D --> E[Re-cited<br/>in new articles]
    E --> F[Becomes 'known fact']
    F --> G[Refutation cost<br/>×100]
    style C fill:#c8e6c9
    style G fill:#ffcdd2

The 0-24h window is critical. Beyond that, cost explodes, because the bullshit incorporates into:

• Google results (and therefore new articles)
• Training data of next-generation LLMs (6-12 month cycle)
• Your competitors' presentations (memorization)

Practical rule: any signal detected > 100 views deserves a documented decision in less than 24h, even if the decision is "we ignore".

Three reusable playbooks

Playbook 1: Bullshit in demo

1. Acknowledge: "I understand why this idea circulates"
2. Replace: "the reality is X, here's the proof Y" (1 number, 1 link)
3. Pivot: "that said, what really matters for your case is Z" (return to value)

Total time: ≤ 2 minutes. NEVER exceed, otherwise the bullshit takes all the mental space.

Playbook 2: Detected fake negative review

1. Platform flag within 2h (evidence: screenshots, IP, profile)
2. Short public response: factual, professional, propose direct channel ("thanks for this feedback, we couldn't find your case in our CRM, contact us at X for investigation")
3. Mobilize 5 real positive reviews within 7 days
4. Document for pattern detection (if recurring: legal escalation)

Playbook 3: False web article

1. Don't comment publicly first (involuntary amplification)
2. Direct email to journalist with factual evidence + proposed correction
3. If not corrected in 7 days: right of reply, or in-depth article on your blog (competitive SEO)
4. Never threaten legally unless characterized defamation (massive Streisand effect)

Key takeaways

• Four attack families: competitive FUD, bullshit objections, fake reviews, LLM hallucinations
• The 0-24h window is critical to avoid informational debt
• The visibility × identity matrix determines whether to respond or ignore
• Acknowledge → Replace → Pivot is the only effective refutation protocol in sales
• NEVER respond emotionally to a public negative review