“How would you weigh a 747?”
This is one of the cleaner brainteasers in the Microsoft canon — cleaner because, unlike “shrunk to a pencil in a blender”, the question has multiple structured approaches that test real estimation skill. The 747 weighs something, the candidate has to figure out a method to determine what, and the methods that work draw on physics, engineering decomposition, or proxy reasoning. The pure question is dead in tech, but the genre survives in PM Fermi estimation, senior+ system design rounds, and engineering capacity planning.
The actual answer
A Boeing 747-400 weighs about 178,000 kg (392,000 lb) empty, and up to 412,000 kg (910,000 lb) at maximum takeoff weight. The interviewer is not testing whether the candidate knows these numbers; they are testing whether the candidate can reach a defensible estimate through structured reasoning.
Approach 1: parts decomposition
Decompose the plane into major components and estimate each:
- Fuselage and wings (structural metal). Aluminum airframe; rough estimate based on dimensions and density.
- Engines (4 of them). Each large engine weighs 4,000–5,000 kg.
- Fuel. A 747 holds about 200,000 liters of jet fuel; jet fuel density about 0.8 kg/L; so fuel ≈ 160,000 kg at maximum.
- Passengers and cargo. About 400 passengers × 100 kg (with luggage) = 40,000 kg.
- Interior fittings. Seats, galleys, lavatories — maybe 10,000 kg.
Total at maximum takeoff: 60,000 + 20,000 + 160,000 + 40,000 + 10,000 = ~290,000 kg. Off by about 30% from the actual 412,000 kg, but in the right order of magnitude. Decent answer.
Approach 2: dimensional / volumetric
The 747 is a thin-walled aluminum tube with wings. Approximating it as a cylinder of metal:
- Length: 70 meters; diameter: 6 meters; surface area: ~1,300 m²
- Aluminum skin thickness: maybe 5 mm = 0.005 m
- Aluminum density: 2,700 kg/m³
- Skin mass alone: 1,300 × 0.005 × 2,700 = 17,500 kg
- Add structural reinforcement (frames, spars, longerons) — maybe 3x: ~50,000 kg structural
This gets the structural part to ~50,000 kg, then add engines, fuel, payload as before. Less accurate than approach 1, but tests whether the candidate can reason about volume and density.
Approach 3: proxy reasoning
“A 747 carries about 400 passengers plus cargo plus fuel. The published maximum takeoff weight should equal the empty weight plus payload plus fuel. If the empty weight is the same order of magnitude as the payload-plus-fuel, then maximum takeoff is roughly twice the empty weight. Empty 747 should be in the 150,000–200,000 kg range.”
This is the laziest approach but often the most accurate. It uses general knowledge about aircraft economics — that a plane with 400 passengers and trans-Pacific range needs to balance lift and weight in known proportions — to back into the answer. Real engineers reason this way often.
Approach 4: physical measurement methods
Some interviewers want the candidate to also describe how to actually measure the weight, not just estimate it. Methods:
- Drive it onto an industrial scale. The most obvious answer; airports have scales for this.
- Put it on a barge and measure displacement. Archimedes’ principle — measure how much water the barge displaces with and without the plane, multiply by water density.
- Cut it into pieces and weigh each. Less practical; mentioned for completeness.
- Calculate from lift-to-weight equilibrium during cruise. If you know the aircraft’s wing area, cruise speed, air density, and lift coefficient, you can derive the weight. Aerospace-specific.
The displacement-via-barge answer is the “clever” one and was traditionally rewarded. In practice, an airport-scale answer is more useful and a senior candidate might note both.
What the question actually tested
When the question worked in the early 2000s, it tested four things:
- Decomposition reflex. Did the candidate immediately split the plane into components?
- Defensible numbers. Could they attach reasonable estimates to each component without panicking about precision?
- Multiple approaches. A polished candidate would offer parts decomposition, then volumetric, then proxy reasoning, and triangulate.
- Physical measurement creativity. The displacement answer specifically rewarded clever physics.
This was a more legitimate test than the absurdist brainteasers because the underlying skills were actually relevant for engineering work. Estimating quantities under uncertainty is something a senior engineer does constantly. The question was retired not because it tested nothing — it tested something — but because the same skill could be tested more efficiently with role-relevant problems (design capacity, compute cost estimation, etc.) without using a transparently leaked brainteaser.
Where it survives
- PM market sizing. “Estimate the total revenue of the airline industry” uses the same structured-decomposition skill.
- Senior+ system design. “Estimate the storage cost of running this at 100x scale” is a Fermi problem.
- Capacity planning at infrastructure roles. Sizing fleets, network capacity, storage all use the same reasoning.
- Aerospace engineering interviews. The actual aerospace industry uses these problems for real reasons.
- Quant firm “estimate this market” questions. Trading interviews include problems like “how big is the daily volume of US treasury futures” with the same Fermi structure.
Is the literal question still asked in 2026?
In tech engineering, no. The pure 747 question is extinct. Asking it earnestly in 2026 would signal a process that has not been updated since the early 2000s. The Fermi-estimation skill it tests is alive and useful, but better questions for that skill have replaced it.
Frequently Asked Questions
What is the actual weight of a 747?
About 178,000 kg empty, up to 412,000 kg at maximum takeoff weight, depending on the variant. The exact number is less important than producing an estimate in the right order of magnitude.
Why is the displacement answer “clever”?
Because it draws on Archimedes’ principle (a 2,000-year-old physics result) and applies it to a problem the candidate has likely never encountered in that framing. The displacement method is also genuinely the most practical method for weighing very heavy objects you cannot put on a scale.
Should I do parts decomposition or physical measurement?
Both, in a strong answer. State the parts decomposition first (gives you a defensible estimate), then offer the displacement method as a measurement approach if the question is asking for both estimation and method.
Is this question still useful?
For tech engineering, no. For PM, consulting, senior+ system design, or aerospace, the underlying skill is still relevant, but better questions exist for those contexts.
Where can I practice this kind of estimation?
For PM-style questions, Decode and Conquer (Lewis Lin) is a common reference. For physics estimation, university-level physics textbooks have Fermi-problem chapters. For tech-relevant capacity estimation, Designing Data-Intensive Applications and Alex Xu’s System Design books have worked examples.