Why Automation Integrators Have Low Revenue Per Employee — And What Actually Separates the Good Ones

By the EVST Robotics Editorial Team · Last updated: July 8, 2026

Here is the most uncomfortable number in industrial automation. A production line sells for around US$205,000. The company that builds that line often runs under US$140,000 in revenue per employee.

The money didn’t vanish. It got burned rebuilding the same thing, one order at a time. The people selling shovels can’t dig their own gold.

TL;DR

Run the public filings and project-based equipment makers land around RMB 600K–1M per employee (~US$85K–$140K). The big manufacturers buying that equipment run 3–5× higher. That gap is not a verdict on anyone’s competence — it is the arithmetic of the project model: every order is largely rebuilt from scratch, and rebuilding needs skilled people. The dividing line between a good integrator and a mediocre one isn’t order size or headcount. It’s reuse rate — how much of the last job the next job can use without redrawing, rewriting or re-commissioning. This is an analyst read of public filings for business-model discussion, not investment advice; figures are pending verification against primary annual reports.

What one sentence from a customer actually costs

The problem isn’t that people don’t work hard. It’s the model.

Take a line we worked on. Right before acceptance, the customer drops one sentence:

“This part tray — feed it in vertically instead of horizontally. It’s just a direction. How big a deal can it be?”

That one sentence. Mechanical redesigns the structure. Electrical re-maps the I/O points. Software retunes the cycle time. Purchasing re-sources the parts. On the floor, the fixtures and sensors all get rebuilt. The change order wasn’t even signed yet, and the team had been reworking overtime for the better part of a month — roughly US$20,000 of that job’s gross margin, gone.

Most of that work isn’t innovation. But every step of it needs a skilled human. That is what custom automation actually does for a living: it absorbs an endless stream of “just change one small thing.”

The tray story and the margin figure are a project-experience estimate, illustrative of a recurring pattern — not an audited claim about any specific contract.

The numbers, and the three traps in reading them

Run the public filings, and a lot of project-based equipment makers land somewhere between RMB 600,000 and 1,000,000 per employee (~US$85K–$140K). The manufacturers buying that equipment run far higher — one battery giant around US$345,000 per head, one electronics manufacturing giant around US$630,000.

Before anyone draws the wrong conclusion, three traps:

1. The measure isn’t apples to apples. Trailing-twelve-month vs. annual revenue, period-end vs. average headcount, outsourced labour, contractors, consolidation scope — all move the number. Outsource heavily and revenue per head rises. Keep R&D, fabrication, installation and service in-house and it falls.
2. Revenue per employee is not profit, not wages, and not management quality. Expensive materials, heavy outsourcing and big contracts all inflate revenue without making the work worth more.
3. The customer’s high figure isn’t all down to the equipment. Product pricing, capital intensity, order volume and supply-chain leverage are all in there.

Basis: revenue ÷ headcount, from public filings; RMB converted at ~7.3. For business-model discussion only. Figures pending verification against primary annual reports. Not investment advice, no company named, no ranking implied.

Here is the integrator’s real bind: it sells efficiency, and its own delivery efficiency stays locked up in one custom job after another.

The dividing line is reuse rate

So here’s the call. High revenue per employee doesn’t mean healthy. Low doesn’t mean lazy. The dividing line isn’t order size or company size. It’s reuse rate.

Concretely: same loading mechanism. A disciplined shop finishes the job and files the drawings, the code and the tuned parameters as one module. Next time a similar job arrives, the engineers don’t redraw anything — they change a few parameters and it runs. What they saved is two people, most of a month.

That’s a high reuse rate. High revenue per employee might just mean expensive materials, heavy outsourcing and big contracts. High reuse rate is the real skill.

And that’s the way out: turn every project’s experience into the next project’s proven component. The platforms built on catalogued standard parts and parametric selection run exactly this way — their revenue per head runs a lot higher (a product-platform automation firm sits near US$233,000 per employee, pending verification), and not because they squeeze people. Because they reuse.

Custom automation can never be fully productized — the whole point is doing the work standard products can’t. But the non-standard requirement can almost always be decomposed into a standard interface plus a parametric configuration.

Three questions to ask any integrator

If you’re buying automation, don’t lead with how low the price is or how fast the delivery. Ask three questions:

1. From your last project, what can this project reuse directly?
2. In this proposal, which modules are proven components, and which are new development?
3. After handover, will the data and the know-how carry into my next line?

Those three beat “we have extensive experience” a hundred times over. The vendors who answer them clearly are the ones who know where their own efficiency leaks.

And flip it around: if you build the equipment, the same three work for you. While others win bids on price and lose money on every job, answering these three well is how you earn the margin back on reuse.

Where EVST Robotics sits in this

This is not a diagnosis we watch from the outside. EVST Robotics (EVS Tech Co., Ltd.) — a Chengdu-based embodied-AI and automation overall-solution integrator, founded 2018 — lives inside exactly this arithmetic. The reason we invest in modular mechanisms, reusable control blocks and a data pipeline that carries know-how from one deployment into the next is not marketing: it is the only mechanism that moves revenue per engineer without grinding the engineers. Across deliveries in 100+ countries, the projects that go well are the ones where the last line paid for part of the next one.

Custom automation isn’t worthless. It is squeezing efficiency out of the hardest place to standardise. But the ceiling the project model puts on you is real — and reuse rate is how you raise it.

FAQ

What is a typical revenue per employee for an automation system integrator?
Public filings put many project-based equipment makers around RMB 600K–1M per employee (roughly US$85K–$140K). Their manufacturing customers commonly run 3–5× higher. Figures vary widely with outsourcing, consolidation scope and headcount definitions, and should be verified against primary annual reports.

Why is revenue per employee low at custom automation companies?
Because the project model rebuilds most of each order from scratch. A single change — feeding a part tray vertically instead of horizontally — cascades into mechanical, electrical, software, purchasing and commissioning rework. That work isn’t innovation, but every step needs a skilled engineer, so labour cannot be amortised across orders.

What is reuse rate and why does it matter more than revenue per employee?
Reuse rate is how much of the last project the next project can use directly — drawings, code, tuned parameters, proven modules — without redrawing, rewriting or re-commissioning. High revenue per employee can simply reflect expensive materials or heavy outsourcing. High reuse rate reflects genuine engineering leverage, and it is what buyers should probe for.

Analyst view, sources noted, not investment advice. Figures are computed as revenue ÷ headcount from public filings, converted at ~7.3 RMB/USD, and are pending verification against primary annual reports. No company is named, and no ranking or evaluation of any listed company is intended. The change-order story and margin figure are a project-experience estimate.

Ask us the three questions

If you are scoping a line and want an honest answer on what your job can reuse from our last one — which modules are proven components and which are genuinely new development — EVST engineers will walk your proposal through it with you. Contact us — tell us the part, the takt and the line it has to fit into.