Your in-house team isn't allowed to tell you the truth

Bob the Pilot

Nassim Taleb tells a story in Skin in the Game about Bob, a freelance pilot.

You own a small airline. You've contracted Bob for a special Oktoberfest flight to Munich. The plane is full of excited passengers (lawyers) who've been waiting all year for beer, pretzels, and sausages in laughter-filled hangars.

The day before the flight, Bob calls. A Saudi Sheikh offered him more money to fly a party to Vegas. The offer is so generous it covers whatever penalty Bob owes you for breaking the contract.

Bob cancels. You're ruined.

Taleb's point isn't that Bob is a bad person. His point is that this is why companies prefer employees over contractors.

Not because employees are better. Because employees are controllable.

Employees have mortgages. Families. Reputations to protect. They can't just walk away when a better offer comes. They show up Monday regardless of what someone offered them Friday.

Bob is a wolf. He has options. He has other clients. He doesn't need you to survive.

An employee is a dog. Domesticated. Dependent. Collared.

I ran one of the top YouTube media buying agencies in the world for 8 years. I was Bob.

And I watched CMO after CMO try to turn me into a dog — or replace me with one.

What Employment Actually Buys

Taleb makes an uncomfortable note about what employment actually signals:

Not skill. Not talent. Not capability.

Submission.

Think about that next time you hear a CMO talk about "building internal capabilities" or "bringing it in-house."

They're not building capabilities.

They're building a court. A court of domesticated dogs who will never bite the hand that feeds them.

Why Dogs Can't Be Trusted

Here's where Taleb's insight gets dangerous for companies:

Read that again.

Cannot be trusted for critical decisions.

Why? Because their incentive isn't to be right. Their incentive is to not get fired. And in most organizations, you don't get fired for being wrong. You get fired for pissing off your superior.

Taleb again:

So the in-house media buyer's real optimization function isn't ROAS or CPA or any metric on any dashboard.

It's:

• Don't challenge the CMO's strategy

• Build reports that make the CMO look good

• Never say "the funnel is the problem" when the CMO owns the funnel

• Find external factors to blame when performance drops

The dog doesn't optimize for the hunt.

The dog optimizes for keeping the master happy.

The Wolf's Dilemma

I was the wolf (at least thats what i though).

When a CMO handed me a target — "Get CPA down 30%" — I had two options.

Option 1: Tell the truth. CPA isn't a media buying problem. Your funnel converts at 1.2%. Your offer doesn't justify the price point, your landers is amess. I can optimize bids all day, but I'm moving a 10% lever on a 100% problem. The issue lives in your domain, not mine.

This gets you fired for being "difficult."

Option 2: Accept the task, try your best, fail, take the blame. Exhaust every lever you control. Get 8% improvement. Not 30%.

This gets you fired for "underperforming."

The wolf tells truth because the wolf has other clients. The wolf can walk away.

But both options get you killed.

Because the CMO doesn't want a wolf. The CMO wants a dog that accepts impossible tasks and takes the blame when they fail.

I watched this playbook unfold over and over: CMO assigns a target that can't be hit with the levers you control. You try everything within your 10% of influence. You fail to hit the 30% target. CMO documents the failure. "If you can't make the change, we're bringing this in-house."

The wolf gets fired.

CMO hires in-house. They report to the CMO. Their reviews are written by the CMO. Their careers depend entirely on the CMO's perception.

They are dogs.

And now the same performance that got the agency fired becomes acceptable — even praiseworthy — when it's internal.

Because there's no wolf left in the room. No one who might tell the truth. No one whose survival doesn't depend on the CMO's approval.

"But Media Buying Doesn't Matter Anymore"

Here's the thing — the CMO isn't entirely wrong about media buying.

Algorithms did commoditize it. Google's Smart Bidding, Meta's Advantage+.... they've eaten 90% of what media buyers used to do.

The difference between the best media buyer and an average one is maybe 10-15% in outcomes. The platforms took our edge.

So the CMO has half a point: why pay an agency percentage of ad spend when an in-house person can push the same buttons?

But here's what actually happened.

Agencies aren't stupid. We saw this coming years ago. We adapted.

The value migrated upstream to creative — and we followed it. The agencies that survived shifted from "we'll optimize your bids" to "we'll produce the creative that feeds the algorithm."

So now agencies produce 100 ads. Test constantly. Find the 1 that works. Scale it until it dies. Repeat.

This was and is reality of performance marketing: 99% of things fail. That's not incompetence — that's the nature of the game.

You're buying lottery tickets. You're placing small bets hoping one hits big enough to pay for all the misses.

This is what Taleb would call Extremistan. One creative returns 100x. Ninety-nine return nothing. The hit pays for the misses. That's the only math that works.

And here's where the scam continues.

When the agency produces 100 ads and 1 works, the CMO calls it "failure rate."

When the in-house team produces 100 ads and 1 works, the CMO calls it "learning."

Same results. Different framing.

The agency's 99 failures are "waste" and "inefficiency."

The in-house team's 99 failures are "insights" and "iteration."

The CMO is applying Mediocristan logic — where 99% failure is genuinely unacceptable — to an Extremistan domain where 99% failure is the only path to the 1% that matters.

They know this. They exploit it anyway.

Because the agency's "failure" justifies firing them.

And the in-house team's "learning" justifies the CMO's empire.

The Accountability Asymmetry

The CMO holds external vendors to outcome standards they would never apply to their own team.

Agency standard:

"Decrease CPA by 30% or you're fired."

"Only 1 out of 100 ads worked — that's a 99% failure rate. Unacceptable."

In-house standard:

"Do your best. Report weekly. Don't cause problems."

"We tested 100 concepts and found a winner. Great learning. Great process."

Same work. Same results. Different standards.

Why? Because the agency's failure serves the CMO's narrative. The in-house team's "success" also serves the CMO's narrative.

It's a Bob Rubin trade — named after the Treasury Secretary who made hundreds of millions from decisions that looked good in the short term, then left others holding the bag when they blew up.

The structure is simple: heads I win, tails you lose.

If the agency succeeds → "My strategic direction enabled their performance."

If the agency fails → "I identified the problem and fixed it by bringing capabilities in-house."

If the in-house team succeeds → "Building internal capabilities was the right call."

If the in-house team fails → "The market is challenging. The algorithm changed. iOS updates. Seasonality."

The CMO takes credit for success regardless of who produces it.

The CMO assigns blame for failure — but only to people outside their court.

There is no outcome that makes the CMO look bad.

That's not strategy. That's portfolio construction for career risk.

The Court Expands

Taleb describes how organizations historically controlled people:

The CMO builds exactly this.

Once the agency is gone, look who's left in the room:

The in-house media team. Reports to CMO. Won't challenge CMO's strategy. Career depends on CMO's approval.

The analytics team. Builds dashboards CMO approves. Won't surface inconvenient data. Knows which metrics make the CMO look good.

The creative team. Executes CMO's vision. Won't say "this creative direction is the problem." Knows who signs their performance review.

Everyone in the room shares one incentive: don't piss off the CMO.

Nobody's job is to tell the truth.

Nobody's career survives saying "the strategy is wrong" or "this metric is meaningless" or "we're hiding the real problem."

The court isn't built for performance. The court is built for protection.

Why the Wolf Dies First

The agency is dangerous because:

We don't report to the CMO. Our incentive is results. If we don't deliver, we lose the account. That's real accountability — not "my boss decides if I did a good job."

We have other clients. We're not dependent on any single relationship. We can walk away. We can tell uncomfortable truths because our survival doesn't depend on your approval.

We see patterns across the market. We're running campaigns for 10, 20, 50 brands simultaneously. We know what's actually working right now — not just what's working in your echo chamber.

We might tell the board the truth. If the CMO's strategy is failing, the agency has no incentive to cover for them. We might say something in a meeting. We might put something in a report. We're a loose end.

The agency is the one relationship the CMO doesn't fully control.

So it has to die.

Not because of cost. Not because in-house is more efficient. Not because agencies can't deliver.

Because the wolf might tell the emperor he has no clothes.

And the court is designed to make sure that never happens.

The Best Slave

Taleb's darkest observation about employment:

Think about that $80k in-house media buyer.

They know they're overpaid relative to their market options. They know the CMO gave them this opportunity. They know they're one "not a culture fit" conversation away from job hunting with a mortgage payment due.

They will never challenge the CMO.

Not because they're stupid. Not because they're cowardly. Because the incentive structure makes challenge irrational.

Taleb point this:

The in-house team has mortgages. Kids in daycare. "Careers" to protect. 401k vesting schedules. Health insurance for the family.

The agency had other clients.

Which one is more likely to tell you the funnel is broken?

Which one is more likely to say the CMO's strategy isn't working?

Which one is more likely to surface data that contradicts the narrative?

The CMO knows the answer. That's why they build courts of employees instead of working with wolves.

The Dog and the Wolf

Taleb uses an ancient fable that appears in different forms across cultures — from Ahiqar to Aesop to La Fontaine:

The dog has comfort. Regular meals. A warm place to sleep. Security.

The dog also has a collar.

Taleb extends the fable:

Think about what happens when the CMO leaves.

And they always leave. Average CMO tenure is what — 18 months? 24 months? They're already interviewing for the next job while building their "transformation" narrative.

When they leave, what happens to the court?

The in-house team has been optimized for one thing: pleasing that specific CMO. They've learned to build dashboards that CMO liked. Present data that CMO wanted to see. Confirm strategies that CMO believed in.

They're domesticated to a master who's gone.

New CMO arrives. New preferences. New strategies. New metrics that matter. New court needed.

The old dogs get put down.

Not literally fired, usually. Restructured. Managed out. Given impossible tasks until they quit. Laid off in a "strategic reorganization."

The wolf survives because the wolf never needed an owner. The wolf has skills that transfer. The wolf has other clients. The wolf knows how to hunt.

The dog survives only as long as the master wants to keep feeding it.

The Metrics Corruption

There's a tradeoff in cancer treatment between laser surgery and radiation therapy. Laser surgery might have worse 5-year survival rates but better 20-year outcomes. Radiation looks good on the 5-year metric but creates secondary tumors later.

If doctors are measured on 5-year survival rates, they'll recommend radiation — even though surgery is better for the patient long-term.

The metric corrupts the decision. Skin in the wrong game.

The same thing happens with CMO dashboards.

The CMO needs to show "progress" to the board. So they pick metrics that show progress:

• Change the attribution window — suddenly ROAS improves 40%

• Redefine "customer acquisition cost" — suddenly CAC drops

• Switch from performance metrics to "brand awareness" — suddenly nothing is measurable and everything is narrative

• Compare to a cherry-picked baseline — suddenly growth looks impressive

The in-house analytics team builds whatever dashboard makes the CMO's story work.

Will they challenge the methodology? Will they say "this metric is meaningless" or "we're comparing apples to oranges" or "this attribution model is fiction"?

Of course not.

Their job security depends on the CMO's happiness, not the company's clarity.

The dashboard becomes a propaganda machine for the CMO's career narrative.

And the board sees green charts and approves of the "transformation."

The Broken Feedback Loop

When an agency fails to deliver, they lose the client.

Natural selection. Agencies that consistently fail go out of business. Agencies that deliver survive and grow. The feedback loop works. Bad performance gets eliminated. Good performance gets rewarded.

When an in-house team fails to deliver, what happens?

• Blame the algorithm

• Blame the market

• Blame iOS changes

• Blame seasonality

• Blame the previous agency

• Blame the economy

• Blame the competition

The feedback loop is broken.

Bad performance doesn't get eliminated. It gets explained away. The people who are best at explaining away failure rise to the top. The system doesn't self-correct. It accumulates dysfunction until something breaks catastrophically.

But by then, the CMO has moved on. They're at a new company, telling a new board about the "transformation" they led at the last place.

The catastrophe belongs to whoever's left holding the bag.

The Hidden Fragility

From the outside, the CMO's court looks stable.

Dashboards are green. The CMO presents "transformation progress" to the board. Headlines about "building internal capabilities." LinkedIn posts about "the team" and "the journey."

Underneath, fragility is accumulating:

Single points of failure. The one media buyer who actually knows the account leaves. The one analyst who built the dashboard quits. Everything breaks. Institutional knowledge walks out the door.

No external perspective. The agency used to see patterns across 50 accounts. The in-house team sees one account. They're optimizing in a vacuum. They have no idea what's actually working in the market.

Metrics designed to hide problems. The dashboard shows what the CMO wants to see, not what's actually happening. Problems compound invisibly until they're too big to hide.

Concentrated dependency. The entire marketing organization depends on one person's narrative. One person's strategy. One person's career incentives.

The CMO built a system optimized to hide risk, not manage it.

The fragility is invisible — until it isn't.

And when it breaks — really breaks — the CMO is long gone, interviewing somewhere else with a "successful transformation" story that doesn't mention the catastrophe they left behind.

Why I Stopped Being the Wolf

Eight years of this.

Eight years of being set up with impossible tasks. Eight years of telling truths nobody wanted to hear. Eight years of watching CMOs build courts and burn companies.

Eight years of being the scapegoat for problems that lived in someone else's domain.

You tell the truth, you get fired for being "difficult."

You accept impossible tasks and fail, you get fired for "underperforming."

You somehow succeed, they take credit and move the goalposts.

And here's the thing — I adapted. We all did.

Media buying got commoditized, so we moved to creative. We learned to operate in Extremistan. We accepted that 99% failure is the game. We built systems to produce 100 ads and find the 1 that works.

It didn't matter.

The CMO just moved the goalposts.

Our "99% failure rate" became the new excuse. Same work the in-house team does — but when we do it, it's "waste." When they do it, it's "learning."

The game was never about performance. It was about control. About credit. About building a court that can't challenge the king.

And here's what broke me:

I built an agency. I hired people. I became responsible for a pack.

Suddenly I couldn't just walk away from bad clients. I had payroll to make. People with mortgages and families depending on me keeping the lights on.

The thing that made me a wolf — the ability to tell truth and walk away — disappeared the moment I had people counting on me.

They don't just domesticate employees. They domesticate agency owners too. You take on responsibility for others, and suddenly you're wearing a collar you didn't notice putting on.

I found myself taking bad deals because we needed the revenue. Biting my tongue because we couldn't afford to lose the account. Playing the game I knew was rigged because the alternative was letting down the people who trusted me.

The wolf with a pack to feed is just a dog with better marketing.

I got tired of pretending otherwise.

So I stopped.

What I Build Now

I got tired of being the wolf that became a dog. Tired of playing a game where responsibility for others forces you into the same trap you saw so clearly from outside.

So I started building something different.

Not another agency that scales by adding headcount and becomes hostage to every CMO's power games. Not another service business where growth means more people depending on you staying quiet.

I started building the thing that exposes where the problems actually live.

That's what Bratrax is.

Not another dashboard that reports to the CMO. Not another tool the court can corrupt. Not another set of metrics designed to make someone's career narrative look good.

A system designed to surface the hidden risks the court is designed to bury.

Where is the company actually fragile? Not where the dashboard says. Where reality says.

What's actually driving performance? Not what the attribution model claims. What the data shows when you stop torturing it.

Where do problems actually live? In the media buying? In the funnel? In the offer? In the product? Somewhere no one wants to look?

I still have a pack. I'm still responsible for people. But the game is different now.

We're not looking for CMOs who want dashboards that make them look good. We're looking for founders and operators who want to see where they're actually exposed. People who understand that the biggest risk isn't hearing something uncomfortable — it's not hearing it until it's too late.

We build for people who want to be exposed to optionality, not protected from reality. People capable of hearing things they don't want to hear. People who understand that the best hedge against their own blind spots is a system that doesn't care about their feelings.

The dog tells the master what the master wants to hear.

We build for people who don't want to be the master.