Article
Systems Engineering Isn't Just for Aircraft Carriers

On this page
Say "systems engineering" and most people picture a defense prime with a thousand engineers, a wall of binders and a review board for every bolt. It's easy to conclude it isn't for a 12-person company building industrial sensors, a startup with its first medical device, or a supplier making one subsystem for someone else's product.
That conclusion is expensive. Small teams need systems engineering at least as much as large ones; they just need less ceremony around it.
The problems are the same size
A small company building a complex product faces the same failures as a large one:
- A requirement nobody wrote down, discovered at the customer's acceptance test.
- Two subsystems that each work but don't fit, because the interface lived in two people's heads.
- A design change that quietly breaks something three steps away.
- A test campaign that proves the product works, without showing it does what the customer asked for.
Large organizations have process, specialists and margin to absorb these. A small company often has none of the three. One late surprise can eat a quarter's budget or a key customer's trust.
Late fixes cost more, and small teams feel it most
NASA's study of error cost escalation across the project life cycle found that a requirements error caught during operations can cost hundreds of times more to fix than the same error caught while the requirements were being written. The exact multiplier varies by project. The direction never does.
For a large program, that's a line item. For a small company, it can be the difference between shipping and not.
Systems engineering is, at heart, the discipline of finding those problems early, while they are still sentences on a page rather than parts on a bench.
Knowledge in heads doesn't scale, even to 15 people
In a small team, the system often lives in the heads of two or three people. That works until one of them is on vacation during a customer escalation, leaves the company, or the team doubles. Then the questions start:
- Why did we choose this sensor?
- Which requirement does this test verify?
- What breaks if we change this voltage?
Writing down requirements, interfaces, decisions and the links between them isn't bureaucracy. It's how a small company keeps working as it grows.
Customers and regulators expect it
Small companies increasingly sell into markets that assume systems engineering:
- Aerospace and defense primes flow requirements down to their suppliers and expect traceability back.
- Medical device rules expect design inputs, design outputs and verification that ties them together.
- Automotive, rail and energy customers ask for the same evidence in their own vocabulary.
Being able to show "here is the requirement, here is the design element that meets it, here is the test that proves it" is often what separates a supplier that wins the contract from one that doesn't.
Right-sized, not watered down
The standards themselves agree that one size doesn't fit all. ISO/IEC/IEEE 15288, the international standard for system life cycle processes, explicitly expects organizations to tailor its processes to the project in front of them.
For a small team, a practical core looks like this:
- Requirements you can verify. Each one is measurable and has a way to prove it's met.
- Interfaces written down. What passes between subsystems, in what form and within what limits.
- Decisions recorded with their reasons. Trade studies don't need to be long; they need to exist.
- Traceability. Links from needs to requirements to design to tests, so a change shows its ripple effects.
- One source of truth. Not five versions of a spreadsheet in five inboxes.
None of that needs a department. It needs a habit and a tool that makes the habit cheap.
Models make it practical
This is where model-based systems engineering (MBSE) earns its place. Instead of keeping requirements in a document, the architecture in slides and tests in a spreadsheet, a model holds them all as connected items. Ask "what does this requirement affect?" and the answer is a click away, not an afternoon of searching.
Languages like the Lifecycle Modeling Language (LML) were designed to be approachable for exactly this reason: a small team can start modeling without first becoming specialists in modeling.
The bottom line
Systems engineering isn't overhead that small companies can't afford. It's insurance they can't afford to skip. Start small:
- Write requirements you can verify.
- Write down your interfaces.
- Link the two.
The first late surprise it prevents will pay for the effort many times over.
References
- J. M. Stecklein et al., "Error Cost Escalation Through the Project Life Cycle," NASA Johnson Space Center, 2004.
- ISO/IEC/IEEE, "ISO/IEC/IEEE 15288:2023, Systems and software engineering — System life cycle processes."
- Lifecycle Modeling Organization, "Lifecycle Modeling Language (LML) specification."
- Object Management Group, "SysML v2 specification."
Holarch is a model-based systems engineering tool built on LML and SysML v2, for teams of any size.