Buyer's guide · 25 June 2026 · 10 min read

How to choose software for a small machine shop

Most software aimed at machine shops is built for someone else — either giant plants with a dedicated IT team, or generic offices that have never seen a traveller or a calibration cert. Choosing well, as a 5-to-30-person UK precision shop, is mostly about filtering out the stuff that doesn't fit. Here are the eight criteria that actually matter, the traps to avoid — and, honestly, the cases where you don't need any of it. We make one of these systems, so take the obvious bias as read; we've tried to write the guide we'd have wanted before we built our own.

First, the honest gut-check: do you even need it?

If you're two or three people and everyone can see the whiteboard, spreadsheets and a shared drive might genuinely be enough — don't let anyone sell you a platform to solve a problem you don't have yet. The pain that justifies software usually shows up around the point where: jobs start falling through cracks between people; you can't answer "are we making money on this part?" without an afternoon's digging; quoting depends on one person who's a single point of failure; or an ISO audit means a week of folder archaeology. If two or more of those are biting, read on.

The eight criteria

1. Built for shops, or retrofitted for them?

Does it actually understand jobs, routings, operations, travellers, tooling, CNC machines and ISO records — or is it a generic CRM/ERP with "manufacturing" bolted on? Generic office tools (the big-name CRMs and accounting suites) don't model shift work or a first-article inspection, and you'll spend your life bending them. Ask to see the actual job and quality screens, not the marketing site.

2. Does it connect the whole job, or is it another silo?

The biggest hidden cost in a shop isn't the software — it's the re-keying between tools that don't talk. A quote that becomes a job that becomes an inspection that becomes an invoice should carry its data through automatically. If you'd still be copying numbers from one screen into another, you're buying a fourth silo, not a solution. Map your own enquiry-to-invoice flow and ask where the system makes you re-type.

3. Right-sized — not enterprise bloat, not a toy

Enterprise MRP/ERP (the six-figure implementations) is built for plants ten times your size; it'll take a year and a consultant to deploy and you'll use 10% of it. The cheap single-purpose apps solve one slice and leave the rest disconnected. For a small shop the sweet spot is something that covers the whole job, that you can be live on in weeks not quarters, and that you can actually run yourself.

4. Is quality (ISO 9001) built in or bolted on?

If you're certified or heading there, quality can't be a separate spreadsheet. Look for NCRs, CAPAs, FAI, calibration and document control that are linked to the jobs — so audit prep is a few clicks, not a week. A system where quality records live apart from production records will quietly cost you every surveillance audit. (Our clause-by-clause ISO guide covers what to look for here.)

5. The pricing model — watch for the traps

Per-user pricing punishes you for putting the system on the shop floor (where it's most useful). Per-quote or per-transaction pricing makes your costs lurch with your busy months. A flat, whole-shop price is the easiest to budget and the one least likely to discourage you from actually using it. Read what's excluded, too — setup, support tiers, API costs.

6. Data ownership and lock-in

Your jobs, customers and quality records are your business. Before you commit, ask the unglamorous question: "if I leave in two years, how do I get all my data out, and what does it cost?" The right answer is a one-click export to open formats (CSV) at any time, with no fee. If the answer is vague, that vagueness is the product. (More on data ownership and no lock-in.)

7. Who actually supports it?

When something breaks at 4pm with a delivery due, do you reach a human who understands shops, or a ticket queue and a chatbot? Smaller vendors often win here — you get the people who built it. Larger ones have more resources but more distance. Decide which matters more to you, and test it before you buy by asking a real question and seeing who answers.

8. Does the vendor tell you what it's not?

This one's a tell. A vendor who'll tell you the honest limitations — what's still rough, what they're not certified for, who they're not a fit for — is a vendor you can trust on the rest. One who only has strengths and a fictional ROI table is managing you. Precision people know the difference.

Where DMOS fits — and where it doesn't

To practise what we preach: DMOS is built for exactly this gap — 5-to-30-person UK precision and CNC shops that want the whole job connected, quality included, at a flat whole-shop price, run by a working shop owner. It scores well on these criteria because they're the criteria we built it against, in our own shop.

Where it's not the answer: if you're a solo maker, it's probably more than you need. If you're a 200-person plant with complex MRP needs, you likely want a heavier enterprise system. And we'll say it plainly — we're a small company, not yet ISO 27001 or SOC 2 certified, and our AI estimator is still maturing. If any of those is a dealbreaker for you, we'd rather you knew now.