McKinsey Charges $2.8M for What Prediction Markets Do for $75K

The VP of Strategy just presented McKinsey's recommendation to the board. Seven months of work. $2.85 million. Unanimous board approval to proceed with the FDA application. Everyone feels good. McKinsey said 70% chance of approval.

Eighteen months later, the FDA rejects it. $47 million burned on clinical trials. Three years of development time gone. The stock drops 12% on the news.

McKinsey moves to their next engagement. Still gets hired by three other pharma companies that quarter. Still charges the same rates. No black mark on their record. No refund. No consequence.

The Moment You See It

You start asking around. How did McKinsey get to 70%?

They interviewed five ex-FDA people. Talked to three other pharma companies who'd pursued similar indications. Built a model. Presented it in a deck with 147 slides.

But here's what you learn: one of those ex-FDA people privately told his colleague the application was "borderline at best." That never made it into the deck. One of those pharma companies had their similar drug rejected eight months before McKinsey's analysis. That got a single bullet point in appendix D.

The consultants weren't idiots. They saw the same red flags. But their incentive wasn't to be right. It was to deliver a defensible recommendation that let the client proceed. To seem confident. To get invited back.

Wrong prediction? Still paid. Right prediction? Still paid. The only thing that matters is that the client felt like they did their homework.

The Economics of Bullshit

McKinsey made $16 billion in 2024. BCG made $12 billion. Bain made $5 billion.

Not because they're more accurate than alternatives. Because they've figured out how to capture value without bearing risk.

A consulting engagement is expensive insurance for executives. When the decision fails, they can say "we hired the best." The consultant's brand is worth more than their accuracy. Nobody gets fired for hiring McKinsey.

The $2.85 million fee is cheap compared to the $47 million clinical trial. But that framing is insane. The question isn't whether $2.85M is cheap compared to $47M. It's whether spending $2.85M on people with zero skin in the game produces better decisions than alternatives.

It doesn't.

The Conflicts Run Deeper

2022 US House Committee investigation revealed that McKinsey staffed 22 consultants who simultaneously worked for the FDA on opioid regulations AND for Purdue Pharma on getting OxyContin approved.

Same firm. Same time. Opposite sides of the regulatory table.

The consultants advising the FDA on how to regulate opioids worked down the hall from consultants advising Purdue on how to navigate those regulations. Same performance reviews. Same partners approving both projects.

When McKinsey got caught, they expressed "regret" and paid $650M to settle. Then went back to work. Revenue up 7% the next year.

This isn't unusual. It's the model. Consulting firms work for everyone. Pharma company and its competitor. Regulator and regulated. Plaintiff and defendant.

They maintain "Chinese walls" between teams. But consultants are human. They can't selectively erase memories.

Your consultant advised your competitor on their FDA strategy three months ago. Now they're advising you. They're bound by NDA not to share specifics. But they know what questions the competitor asked. What data they were worried about. Where they saw weaknesses. Every question they ask you is informed by that.

You're not paying for neutral expertise. You're paying for expertise contaminated by relationships you don't know about.

What Changes When Experts Lose Money for Being Wrong

Same pharma company. Same decision about whether to pursue FDA approval.

Different approach: invite 25 experts with FDA oncology experience to a private prediction market. Former FDA reviewers who worked on similar drugs. Ex-pharma regulatory people who've submitted these applications. Consultants who've advised on them.

Each expert is background-checked. Employment verified. Conflicts screened. If you're currently advising a competing pharma company, you can't participate. If you have financial interest in the outcome, you're excluded.

The market question: "Will FDA approve [Drug X] for [Indication Y] within 18 months?"

Each expert puts minimum $1,000 of their own money on the line. They trade shares. If they think approval is likely, they buy. If they think it's unlikely, they sell. The market price aggregates their beliefs.

When the FDA makes their decision, the market settles. Right prediction? You profit. Wrong? You lose your money.

The pharma company pays $75,000 for the market. They get a probability estimate backed by 25 experts who each have real money riding on being right.

Watch what happens to behavior when being wrong costs you.

Why Privacy Actually Works Here

Traditional consulting leaks information through consultant brain osmosis. Prediction markets enforce cryptographic boundaries.

Company posts anonymized question. Experts don't know which company until invited and NDA signed. Experts from competing companies automatically excluded. Platform tracks who sees what. Full audit trail. Breach penalties: $500K+ plus disgorgement of trading profits.

HIPAA compliance for healthcare data. GDPR compliance for EU clients. On-premise deployment available for sensitive questions. Zero-knowledge architecture.

McKinsey model: professional ethics and Chinese walls enforced by HR policy.

This model: cryptographic boundaries and economic incentives enforced by code.

Why It Works

Real incentives. $5K on the line means you think harder. Call your old FDA colleagues. Review similar applications from the past 3 years. Don't just say what sounds smart in meetings.

Aggregated wisdom. McKinsey sends 2-3 consultants billing $500-1,200/hour. Prediction market aggregates 25+ experts. One consultant has blind spots. Twenty-five experts with money on the line catch them.

No systemic conflicts. McKinsey works for Pfizer, Merck, J&J, and the FDA simultaneously. Markets automatically exclude experts advising competitors on related questions.

Continuous updating. Traditional consulting delivers a report at the end. Prediction markets update real-time. Competitor's drug gets rejected for safety? Market reprices from 70% to 45% within hours.

Revealed knowledge. Consulting interviews produce hedging. "Well, it depends..." "Several factors..." Prediction markets force commitment. Buy or sell. The price reveals actual belief, not what sounds defensible in depositions.

Auditable track record. McKinsey doesn't publish accuracy. Prediction markets do. "This person predicted 73 FDA decisions with 89% accuracy" becomes a credential. Poor forecasters get filtered out.

The Pattern Everywhere

This works anywhere institutions need expert judgment on uncertain outcomes.

VCs stop arguing about gut feel. Create internal markets: "Will [Startup X] hit $10M ARR by 2026?" Analysts trade with real money. Firm invests in companies with highest probability-weighted returns. No more relying on loudest voice in the room.

Sports teams stop debating. Market question: "Will trading for [Player Y] improve win rate 5%+ next season?" Scouts, coaches, analytics team trade. GM gets aggregated judgment from people who watched tape, crunched numbers, seen the player up close.

Record labels stop trusting PowerPoint. Market question: "Will this artist hit 100M streams next year?" A&R people, producers, marketing team trade. Sign artists clearing probability thresholds. Aggregate wisdom of people who've broken artists before.

Corporations stop death by committee. Market question: "Will this product hit $5M revenue Year 1?" Product managers, sales, engineers trade. C-suite allocates resources based on highest predicted ROI. Markets are harder to bullshit than slides.

Same pattern everywhere: institutions face uncertainty, need expert judgment, traditional methods (committees, consultants, gut feel) don't aggregate information and create conflicts of interest.

Prediction markets solve both.

What Actually Changes

The expert with $5,000 on the line doesn't hedge. She calls her former colleagues at the FDA. Reviews the three most similar applications from the past two years. Reads the FDA reviewer notes that are technically public but nobody bothers to find. Thinks about what she actually believes, not what sounds defensible in a meeting.

She doesn't profit from sounding smart. She profits from being right.

The expert who advised your competitor last month can't participate in your market. The conflict is automatic disqualification, not just a "Chinese wall" nobody enforces.

The expert who's consistently wrong loses money. After a few bad predictions, he stops getting invited to markets. Poor forecasters are filtered out by economic reality, not by whether they went to the right schools.

The expert who's consistently right builds a track record. "Predicted 73 FDA decisions with 89% accuracy" becomes a credential worth more than "McKinsey alum."

This isn't theoretical. The infrastructure for this exists now. Private prediction markets where institutions can get vetted, conflict-free experts to trade with real money on the answers they need.

The Threat

McKinsey made $16B in 2024. BCG made $12B.

When you can aggregate 25+ conflict-free experts with skin in the game for $75K, it's hard to justify $2.85M for a deck from consultants who just worked for your competitor.

Consulting firms will argue they provide strategic thinking and implementation, not just probability estimates. Fine. But there's a large chunk of consulting revenue that's really "we talked to some experts and here's what they think, also we have conflicted relationships with everyone involved."

That chunk is getting replaced.

Not Just Cheaper

When McKinsey says "70% confident," what does that mean?

Three consultants' judgment, filtered through pressure to satisfy clients, constrained by billable hours, potentially contaminated by work for competitors. No feedback loop. If they're systematically overconfident by 20 points, they never know. Do 100 projects, don't track accuracy, bill whether right or wrong.

Prediction market saying "70% likely" means something different:

Twenty-five people with relevant experience assessed it. None advising your competitors currently. They put real money on that belief. Platform tracks their historical accuracy. Probability updates as new information arrives. Full audit trail for FDA or board review.

It's not just cheaper. It's better information from a better process with better conflict management.

The Choice

That VP of Strategy who watched $47M burn on McKinsey's recommendation? Next time she has a major decision, she has a choice.

Spend $2.85M on consultants with zero skin in the game. Get a 147-slide deck full of hedged language and potentially contaminated by their work for competitors. No refund if they're wrong. No consequence to their business.

Or spend $75K on a prediction market with 25 vetted experts who each lose money if they're wrong. Get a probability estimate that updates in real-time as new information emerges. See each expert's historical accuracy. Know that nobody advising your competitors is in the market.

The consultants will argue they provide more than predictions. They provide "strategic thinking" and "implementation support." Fine. But a huge portion of consulting revenue is just "we talked to experts and here's what they think, also we might have conflicting relationships we can't tell you about."

That portion is getting replaced.

Not in 20 years. Now. The first major institutions are already running internal prediction markets for key decisions. VCs are using markets instead of partner votes. Pharma companies are running pilot markets alongside traditional consulting to compare accuracy.

Five years from now, "hired McKinsey for a probability estimate" will sound as outdated as "sent a fax to coordinate the project."

The moment you can get better information for 97% less cost from people who actually lose money when they're wrong, everything that depends on the old model breaks.