What we did, what we found, and why it matters for your hiring
We built a behavioral hiring engine. You paste a job description, answer 3 questions, and it picks the 5 behavioral dimensions (out of 15) that role demands most. Demand levels from 1 to 5. Evidence traced back to actual JD phrases. Interview questions generated automatically.
Then we got curious.
What would happen if we ran Netflix's actual JDs through the same engine? And then did the same for Stripe, Datadog, and Uber?
Here's the short version: Netflix's behavioral fingerprint looks nothing like what the culture memo would predict. The company famous for "radical candor" doesn't write job descriptions that select for conflict. The company that worships autonomy barely signals it. And the one dimension Netflix demands at maximum intensity? Not a single competitor asks for it.
We ran 9 real roles. 5 Netflix. 4 competitors. Same engine. No opinions. Here's what we found.
But first: An honest caveat
A job description is not a hiring decision. It's one signal. One layer.
The actual hiring decision involves the interview panel, who's available, who says yes, who gets a counteroffer, what the recruiter had for lunch that day. (I'm only half joking about the lunch part.)
74% of employers admit they've made a wrong hire. Not because they're bad at reading JDs. Because hiring is hard, multi-layered, and human.
So why look at JDs at all?
Because a JD is what a company signals to the market about what it thinks it wants. It shapes who applies. It shapes what the interview panel asks. It shapes who gets screened out before they even talk to a human.
If that signal is misaligned with what the role actually demands, everything downstream gets noisy.
Think of it this way. A JD is like a dating profile. It doesn't tell you how the relationship will go. But it tells you who shows up to the first date.
Netflix by the numbers
Some context so we're all working with the same picture.
Netflix had 16,000 employees at the end of 2025. Up from 14,000 the year before. That's 14% growth in a single year.
Their engineering team alone is about 3,480 people. A quarter of the entire company. They spent $3.39 billion on R&D in 2025. That's not a line item. That's a mid-sized country's education budget.
They pulled in $45 billion in streaming revenue. $11 billion in net income. Revenue per employee is somewhere around $2.8 million. For comparison, the average across the S&P 500 is about $600K.
So when Netflix writes a job description, it's not an afterthought. These are high-stakes signals for high-cost roles in a company that generates nearly $3M per person.
The culture memo (originally published in 2009, updated many times since) is probably the most famous hiring document in tech. 125+ million views on the original SlideShare deck. Key themes: freedom and responsibility, radical candor, the keeper test, context not control.
Every founder who's read it has thought: "We should hire like Netflix."
But what does Netflix actually ask for in its JDs? Let's look.
What the engine actually does (this matters)
This isn't keyword matching.
When you paste a JD into hire.korture.com, the engine reads three things at once:
Layer 1: Your JD text
Not just the requirements section. The "about us" paragraph. The team description. The way the company describes itself.
When Netflix's Compute Platform JD says "Creating clarity within ambiguity to produce and execute designs and plans," the engine maps that to a specific behavioral dimension (navigating without a clear plan) at a specific demand level.
Layer 2: Company culture signals
Every JD contains clues about the company's culture. Sometimes it's explicit ("Netflix is renowned for its culture of Freedom and Responsibility"). Sometimes it's embedded in how responsibilities are described.
The engine extracts these and uses them as evidence. For Netflix's Compute Platform role, the company culture signal boosted "Setting your own direction" because the engine recognized Netflix's known approach to individual ownership.
Layer 3: Sharpening questions
Three quick choices:
- How does this person spend their day (deep work, collaboration, or mix)?
- What's the pace (ship fast or get it right)?
- Who do they work with most (customers, internal teams, or direct reports)?
These shift which dimensions surface. A "ship fast" role surfaces "Deciding before all the data is in." A "get it right" role surfaces "Following the system, every time."
Behind all three layers, the engine cross-references against validated occupational research: 1,016 occupations, 73,308 work activity scores, and 18,797 tasks from the US Department of Labor's database.
That's what makes "strong communicator" mean something different for a PM at a 45-person startup vs. a PM at 16,000-person Netflix. Same words. Different behavioral reality.
Every dimension the engine picks comes with an evidence chain. You can trace any demand level back to the exact JD phrase, sharpening answer, or company signal that produced it. Nothing is a black box.
The experiment
We ran 9 real roles through this engine.
Netflix (5 roles):
- Senior Software Engineer (L5), Experimentation/Analysis Platform
- Software Engineer (L5), Compute Platform
- Product Manager, Ad Serving
- Product Manager, Ads Scoring and Ranking
- Creative Content Manager, AV Studio
Competitors (4 roles):
- Datadog: Senior Data Engineer (Observability), Senior SWE (Distributed Systems)
- Stripe: Backend Engineer, Core Technology
- Uber: Senior SWE, Backend Platform Engineering
Each brief uses real JD text from actual career pages. Each role gets 3 sharpening questions answered based on the role context. The engine picks 5 dimensions from a pool of 15, with demand levels and evidence chains.
Here's the pool of 15 behavioral dimensions. Plain English names, not jargon:
- Long stretches of solo focus
- Staying hands-on while doing everything else
- Navigating without a clear plan
- Rapid focus shifts
- Convincing people who don't report to you
- Navigating competing demands from different people
- Finding new ways when the obvious path is blocked
- Deciding before all the data is in
- Following the system, every time
- Working within rules that slow things down
- Setting your own direction
- Carrying other people's growth
- Being "on" for people outside the company
- Thinking long-term while putting out fires
- Walking into disagreements instead of around them
Each role gets exactly 5. The other 10 are excluded. What gets excluded is often more interesting than what's included.
The Netflix behavioral fingerprint
Across all 5 Netflix roles, here's what kept showing up:
Thinking long-term while putting out fires
Appeared in 4 out of 5 roles. This is Netflix's most common behavioral demand.
Hold a quarterly vision while the building is on fire. Level 4 every time. Consistent across engineering, product, and creative.
Navigating without a clear plan
Appeared in 3 out of 5 roles. And every time it showed up, it scored a 5 out of 5.
Maximum intensity. Netflix doesn't just ask people to tolerate ambiguity. They need people who get energy from it. Who light up when there's no playbook.
Convincing people who don't report to you
Appeared in 3 out of 5 roles.
Influence without authority. Makes sense for a flat company. But notice what this implies: if you need to convince people, there's friction. The "freedom" in freedom and responsibility has a cost.
Staying hands-on while doing everything else
Appeared in 3 out of 5 roles.
Even their PMs need deep technical chops. The Ads Scoring PM role scored a 5 on this dimension: ML model knowledge, auction mechanics, scoring algorithms. That's not a typical PM profile.
Deciding before all the data is in
Appeared in 3 out of 5 roles.
Move fast. Ship. Prototype. But ground it in data. Netflix wants people who can look at 60% of the picture and say "I've seen enough, let's go."
What Netflix doesn't ask for (this is the real story)
Four dimensions appeared in zero Netflix roles. Not one. Across engineering, product, and creative.
Walking into disagreements instead of around them: 0 out of 5
Netflix. The company that popularized "radical candor." The keeper test. "Say what you think, even if it's uncomfortable."
Not a single JD generates walking into disagreements as a top-5 behavioral demand.
Does this mean Netflix doesn't value candor? No. It means their JDs don't select for it.
They select for "Navigating without a clear plan" and "Convincing people who don't report to you." Those are related. But they're not the same thing.
A person who thrives in ambiguity might avoid conflict by reframing the problem entirely. Someone who's good at convincing people without authority might build consensus before disagreement surfaces. Neither approach requires walking into disagreements directly.
The signal and the narrative don't match. That's not a criticism. It's a finding. And it's interesting.
Carrying other people's growth: 0 out of 5
Another surprise.
Netflix talks about "talent density" all the time. But not a single role we tested demands the energy it takes to coach, develop, and give hard feedback to direct reports.
When you hire for "Setting your own direction" and "Navigating without a clear plan" (which Netflix does), you're selecting for people who figure things out themselves. The trade-off? Less natural investment in growing others.
This might not be a gap. It might be by design. Netflix famously pays top of market and expects people to arrive fully formed. But it's worth noticing.
Being "on" for people outside the company: 0 out of 5
Working within rules that slow things down: 0 out of 5
No compliance. No being "on" for people outside the company. Netflix hires people who operate in internal chaos. Not people who absorb external pressure or work within rules that slow things down.
Netflix vs. competitors: Where the real difference is
We compared Netflix's engineering roles against Stripe, Datadog, and Uber. Same level. Similar technical domains. Same engine.
What every company demands (table stakes)
Three dimensions appeared in almost every engineering role, regardless of company:
- Staying hands-on while doing everything else: 6 out of 6 roles. Always level 5. Table stakes for senior distributed systems work.
- Long stretches of solo focus: 5 out of 6 roles. The one exception was Netflix's Compute Platform, where "Navigating without a clear plan" at level 5 pushed solo focus to a 4.
- Thinking long-term while putting out fires: 5 out of 6 roles. Running tier-one systems means this is just the job.
If you're hiring a senior engineer and your behavioral brief doesn't include these three, check your JD.
Where Netflix diverges
Netflix demands navigating without a clear plan. Competitors don't. At all.
Netflix's Compute Platform SWE scored "Navigating without a clear plan" at level 5.
The evidence traces back to JD phrases like "Creating clarity within ambiguity" and "Comfort with ambiguity and ability to create structure where none exists."
Stripe? Zero. Datadog? Zero. Uber? Zero. None of the competitor engineering roles surfaced navigating without a clear plan in their top 5.
This is the actual Netflix hiring edge.
Other companies hire engineers who are good at their craft. Netflix hires engineers who are good at their craft and get energy from not knowing the plan.
Competitors demand finding new ways when the obvious path is blocked. Netflix (mostly) doesn't.
3 out of 4 competitor engineering roles included "Finding new ways when the obvious path is blocked." Datadog (both roles) and Uber all scored it at level 4.
Netflix engineering? Neither role surfaced it.
Different philosophy.
Competitors hire for "figure out a creative solution." Netflix hires for "figure out the situation first."
Subtle. But if you put a creative-problem-solving engineer in a navigate-ambiguity role, they'll get frustrated. They want a problem to solve. Netflix hands them a fog to walk through.
Same title, different demands
Two platform engineering roles. Same seniority. Same technical domain. Different companies.
Netflix SWE L5, Compute Platform
- Staying hands-on while doing everything else (5)
- Navigating without a clear plan (5)
- Setting your own direction (5)
- Thinking long-term while putting out fires (4)
- Long stretches of solo focus (4)
Uber SWE, Backend Platform Engineering
- Staying hands-on while doing everything else (5)
- Long stretches of solo focus (4)
- Finding new ways when the obvious path is blocked (4)
- Convincing people who don't report to you (4)
- Deciding before all the data is in (4)
They share one dimension at the top: "Staying hands-on while doing everything else."
After that? Completely different profiles.
Netflix wants someone who thrives in fog. Three of their five dimensions are about operating independently without clear direction. This is the "leave me alone and I'll figure it out" profile.
Uber wants someone who collaborates and innovates. "Convincing people who don't report to you." "Finding new ways when the obvious path is blocked." Fast decisions. This is the "let's solve this together" profile.
A great engineer at Netflix might struggle at Uber. Not because of technical skill. Because the daily energy demands are different.
One role drains you if you need structure. The other drains you if you need solitude.
And this is exactly what happens when a founder copy-pastes a Netflix job description for their 80-person company.
You attract the "leave me alone" person for a role that actually needs the "let's solve this together" person.
Six months later, they quit. You call it a bad hire. It was a bad signal.
The rarity problem (this is the part nobody talks about)
After the engine generates a brief, it runs a population check.
It takes the role's behavioral profile and compares it against 445 real people who have completed assessments on our platform. Each person gets a fit score: strong, good, partial, or low.
About 1 in 4 people (25%) land in the "strong" tier for any given role.
That sounds fine. Except "strong" doesn't mean the same thing for every role.
For PM and creative roles, the match quality is real
When the engine says someone is a "strong" fit for the Netflix Creative Content Manager role, that person's behavioral profile genuinely lines up with what the role demands. The match is tight. The energy patterns align.
For engineering roles, "strong" means "best we could find"
The engine still picks the top 25%. But those top people don't match as well. Their profiles are further from what the role demands. They're the best in the pool, but the pool itself doesn't naturally produce many people with this behavioral pattern.
Why? Because our current database of 445 people is mostly PM, sales, and operations profiles. People who applied for our own roles, plus friends and early testers. We have very few engineers in the mix.
And that's actually the point.
If your company's hiring pipeline also skews toward a certain kind of person (which it does, because your employer brand attracts a certain behavioral type), then the same thing is happening to you.
You're fishing for engineering energy in a pond full of PM energy. Or you're looking for "Navigating without a clear plan" in a pipeline full of people who get energy from "Following the system, every time."
The population check shows you this mismatch before you start interviewing. Not 6 weeks and $15,000 later.
Now think about what Netflix's Compute Platform role demands.
Three dimensions at level 5: "Staying hands-on while doing everything else," "Navigating without a clear plan," and "Setting your own direction."
That's someone who gets maximum energy from deep technical work and from ambiguity and from working alone. All three at the same time, every day.
Across the roles we've tested, 67% of job descriptions ask for behavioral combinations that pull in different directions.
A person who thrives on solo focus often doesn't thrive on rapid context switching. Someone who loves setting their own direction may struggle when flooded with competing demands from others.
This gives you two honest choices:
Choice 1: Keep searching
Accept that the person you need is rare. Budget for more candidate volume. Screen more people. And when you find someone whose behavioral profile actually matches, pay them what they're worth.
Choice 2: Modify the JD
If you can live with "Navigating without a clear plan" at a 4 instead of a 5, or if "Setting your own direction" is a nice-to-have rather than a must-have, the pool opens up.
The engine shows you exactly which dimension is narrowing your search, so you can make that trade-off with your eyes open.
Most companies do neither.
They write a JD that demands a behavioral unicorn, interview 5 people, hire the best available, and wonder why the person seems "off" after 3 months.
The person wasn't off. They were a partial behavioral match in a role that needed a strong one.
What this means if you're actually hiring
Three things this data makes clear:
1. Your JD already encodes behavioral demands. You're just not reading them.
"Work in a fast-paced environment" is code for "Navigating without a clear plan."
"Collaborate with cross-functional teams" is code for "Convincing people who don't report to you."
"Bias to action" is code for "Deciding before all the data is in."
These signals are already in your JD.
The question is whether you're using them to shape your interview process, or ignoring them and asking generic behavioral interview questions that don't match the role.
2. What you exclude matters more than what you include.
Netflix's most interesting signal isn't that they demand navigating without a clear plan. It's that they don't demand walking into disagreements.
The absence tells you more about the actual work environment than any culture deck.
Look at your own exclusions.
If your senior eng role doesn't demand "Carrying other people's growth" but you expect them to mentor juniors, you've got a mismatch.
That mismatch shows up as a "bad hire" six months later. But it wasn't a bad hire. It was a bad signal.
3. "Culture fit" is a feeling. Behavioral demand match is measurable.
95% of IT leaders say they've made a bad hire. The top reasons? Interpersonal issues (29%) and poor culture fit (28%). Together, that's 57% of bad hires caused by something nobody measured.
Meanwhile, the US Department of Labor says a bad hire costs at least 30% of first-year salary. For a senior engineer at $200K, that's $60K. For a Netflix L5 at $350K+, you can do the math.
And 23% of new hires leave before completing year one.
A JD-to-behavioral-brief process doesn't replace interviews. It doesn't replace judgment.
But it gives you a layer of signal that most companies don't have. And it gives your interview panel specific questions that match what the role actually demands, not what someone remembered from a generic question bank.
Try this on your own roles
Everything in this analysis came from a tool you can use right now.
Paste a JD. Answer 3 questions. In under 2 minutes, you'll see:
- The 5 behavioral dimensions this role demands, with evidence traced to specific JD phrases
- What got excluded (and why that matters more than what's included)
- A population check showing how rare your ideal candidate actually is
- Interview questions matched to each dimension, not pulled from a generic bank
Then do what we did.
Run the same role title at your company and a competitor. See where the behavioral demands overlap (table stakes) and where they diverge (your hiring edge).
Not "do we have good culture?" That's unanswerable.
But "does this person get energy from what this specific job demands, every single day?" That's a question you can answer before the first interview.
Your turn
Paste a JD. Find out if you're hiring a unicorn or a real person.