How do you actually measure competency?
In a rant with my cofounder, who in parallel leads important strategic initiatives at a Fortune 50 company, he said the following:
"Ten percent of the people do ninety percent of the company. Not eighty. Ninety."
The crazy part is that he was just stating it like a weather report. And the thing is, if you've ever run a large company or led a huge team, you already know he's right. You've seen it. You have people on your roster who make everything move, and you have people who… are there.
So here's the question that kept bothering me: if we all know this is true, why don't we have a way to identify the ten percent before we hire them?
The proxy problem
Right now the private sector measures competency with proxies. Resumes. Degrees. Years of experience. Interview performance. Certifications. Maybe a take-home project if someone's feeling ambitious.
None of these predict who becomes the ten percent.
A Leadership IQ study tracking over 20,000 hires found that 46% of new employees fail within 18 months. Only 11% of those failures were due to lack of technical skills. The other 89%? Coachability. Emotional intelligence. Motivation. Temperament. The exact things a resume doesn't measure and an interview barely touches.
These numbers might be tolerable when you're hiring for a B2B-SaaS company. They're not tolerable when you're trying to reshore semiconductor manufacturing, build defense systems that work, or bring critical infrastructure online on a timeline that doesn't forgive delays and lives might be lost.
We've built an entire hiring infrastructure around signals that don't actually measure the thing we actually care about: can this person perform when it matters?
Reindustrialization has a capital problem, sure. It also has a talent problem, sure. But, measuring competency of roles that have not existed in America for decades? Oof.
Here's what we found when we tried.
We studied the people who can't afford to get it wrong
When I started building the Arklight Cognitive Index (ACI) to be a part of the Trade School 2.0 that we're building, I didn't look at what other tech companies were doing with talent assessment. I looked at the military and the intelligence community.
Here's why: the military sends people into environments that are high-stakes, highly reactive, and often wildly unpredictable. The variables are irrational. Plans fall apart on contact. And yet the military has been doing this for decades, sending people into chaos and expecting them to perform. They've built a system that, in many ways, works well.
I spent a long time studying how military and intelligence training programs measure readiness. Three principles stood out that the private sector almost universally ignores.
Principle 1 — Measure performance under ambiguity, not ideal conditions
Most private sector assessments test whether you can do the job when everything goes right. The military tests whether you can function when nothing goes according to plan. That's a fundamentally different question, and it's a far more useful one.
Apply this: stop giving candidates clean case studies. Give them an incomplete brief with conflicting constraints and watch what they do. See how their mind works under ambiguity and stress.
Principle 2 — Observe behavior. Stop trusting self-reporting
The military watches what people do under pressure, over time. Not a self-assessment. Not "tell me about a time when." They put you in the scenario and watch. Remember: 89% of hiring failures are behavioral, not technical.
Apply this: build trial periods or working sessions into your process. Two weeks of real work tells you more than ten interviews.
Principle 3 — Build a body of evidence. Kill the single snapshot
One test tells you very little. A pattern of performance across escalating complexity tells you almost everything. The military doesn't certify readiness based on one exercise. They progressively increase pressure, evaluate at each stage, debrief, and feed it into the next round. They build a body of evidence.
Apply this: create a hiring progression where each round increases complexity — screen, working session, simulation, team exercise. Each one reveals more about how someone actually operates.
Why we built ACI the way we did
These three principles became the foundation of ACI — the engine that will drive our entire Trade School 2.0.
ACI will measure how a student processes information, responds to pressure, and recovers from errors — and then it never stops. Every contract they fulfill, every simulation they run, every decision they make under real constraints feeds back into their profile. By graduation, we will have months of longitudinal performance data on every student. We call it the Talent Dossier.
Then we added a layer nobody else has. Through our growing employer partnerships, we will assess top-performing engineers and operators already working in industry — building a cognitive and behavioral benchmark based on the best talent in the field. We will know what elite performance actually looks like in the data. How the best operators make decisions under ambiguity. How they recover from errors. How they calibrate confidence. Once you have that benchmark, you can train directly to it — and push beyond it.
Think about what that means for employers. Instead of a diploma that says "this person completed coursework," you get a competency profile built from real performance data, measured against the best operators in the industry. No university in the country can hand that to an employer. We will.
Here's the takeaway for anyone building a team right now: if you're not measuring behavioral and cognitive performance over time — longitudinally, under real pressure — you're making talent decisions with a fraction of the data available. Start capturing it. Even rough performance data across multiple touchpoints beats a perfect resume every time.
Based on our internal results — it works. That's how we're going to produce the most capable industrial operators in the country.
The bigger question
Go back to the ten percent. The people carrying ninety percent of the weight at every company.
What if you could find them before you hire? What if you could build entire teams of them? What if the bottleneck on your next factory, your next product launch, your next critical hire wasn't capital or strategy — but the fact that you never had a way to measure what actually matters?
P.S. — If you read to the end, reach out. You can use ACI for free. We built it for our Trade School 2.0, but we want to make it available to anyone building in critical industries.