How much does a custom ERP cost for a small manufacturer? (2026, USD / GBP / AUD)

If you run a small manufacturer and you've typed this question into a search box, you've probably already seen the two unhelpful answers. Vendors say "it depends" and ask for a call. Listicles quote "$50,000 to $500,000", which is true in the way "a vehicle costs $500 to $5 million" is true.

Here is a more useful answer, from a studio that builds these systems and publishes its prices.

The short version

For a manufacturer with 10 to 100 staff, a custom ERP built in phases costs $10,000 to $45,000 for the modules most shops actually need, delivered over two to five months. In GBP that is roughly £8,000 to £36,000; in AUD roughly A$15,000 to A$68,000. Running costs after launch are $300 to $800 a month for hosting, monitoring and small improvements.

That is for a system built around your existing accounting software, not replacing it. Replacing accounting too adds a lot of cost and almost never pays back for a business this size.

What you are actually paying for

"ERP" is a word that hides five or six separate tools. Price them separately and the fog lifts.

Module	What it replaces	Typical cost (USD)	Build time
Quoting and RFQ intake	Email inbox, Excel quote template	$3,000 – $7,000	2–4 weeks
Job and production tracking	Whiteboard, job cards, "where is it?" phone calls	$4,000 – $9,000	3–5 weeks
Inventory and stock movements	The stock spreadsheet nobody trusts	$4,000 – $10,000	3–6 weeks
Purchasing and suppliers	Reorder by memory, PO PDFs in a folder	$2,500 – $6,000	2–4 weeks
Job costing (quoted vs actual)	Nothing. Most shops don't know their real margin per job	$3,000 – $6,000	2–3 weeks
Dispatch and delivery notes	Word templates, signatures on paper	$2,000 – $5,000	2–3 weeks
Dashboards and reports	Monthly Excel exercise	$1,500 – $4,000	1–2 weeks

Add the modules you need. Most shops start with two or three. A quoting tool plus a job board, the pair that removes the most daily pain, lands at $7,000 to $16,000 and is usable within six weeks.

What pushes the price up

Four things, in order of how often we see them:

Integrations. Pushing invoices to Xero or QuickBooks is cheap and well-trodden. Two-way sync with a legacy on-premise system from 2009 is not. Each integration is $1,000 to $5,000 depending on how friendly the other side is.
Data migration. Twelve years of part numbers in six spreadsheets with three naming conventions takes real time to clean. Budget $1,500 to $5,000 and expect to make decisions about what to leave behind.
Shop-floor hardware. Barcode scanners, tablets at machines, label printers. The software is simple; procurement and setup are where time goes.
Approval and compliance workflows. ISO 9001 traceability, batch records for food, material certs for aerospace. Doable, but each adds structure that has to be designed carefully.

Things that do not push the price up as much as people expect: number of users (custom software has no per-seat fee), number of products, and "we want it to look nice" (it should, and that is included).

When off-the-shelf is the better buy

Be honest with yourself here, because a custom build is a commitment.

Buy off-the-shelf (MRPeasy, Katana, JobBOSS, Fulcrum, Odoo and the like) if:

Your processes are standard and you are willing to change them to match the software.
You have under ten staff and nobody who will own the system internally.
You need it next week.

Build custom if:

You have tried two off-the-shelf tools and your team abandoned both.
Your quoting logic, routing or costing is genuinely specific to your shop and is where your margin comes from.
Per-seat pricing is punishing a team that is growing.
You need the system to fit how the floor actually works, not how a product manager in another country imagined it.

The honest middle ground, and the one we recommend most often: keep your accounting package, buy nothing else, and build the two or three modules that are specific to you.

How to buy it without betting the business

Never sign a $40,000 contract for a system you haven't seen. The buying pattern that works:

A free audit. Someone maps one workflow with you and shows you a clickable mockup. This takes a short call and a few days. If a vendor won't do this, be suspicious.
A fixed-fee pilot. One module, two weeks, in daily use. At Telkast this starts at $750 and is credited in full toward the bigger build. You learn more about a vendor in two weeks of working together than in any sales process.
Phases with a fixed price each. Each module is quoted, built, demoed weekly and launched before the next begins. You can stop after any phase and keep everything built so far.
Your code, your accounts. The repo, the hosting and the database live in accounts you own. If the relationship ends, nothing ends with it.

A worked example

An 18-person precision machine shop (a composite of the shops we talk to, not a client) quotes from an inbox and tracks jobs on a whiteboard. Two people spend roughly eight hours a week each re-keying and chasing status.

Phase 1, quoting and RFQ intake: $5,500, four weeks. Quotes go out in hours instead of days.
Phase 2, job board with customer status page: $6,500, four weeks. The phone stops ringing with "where is my part?".
Phase 3, job costing: $4,000, three weeks. For the first time the owner sees margin per job.

Total $16,000 over eleven weeks, plus about $400 a month to run. Sixteen hours a week of admin recovered is worth more than that in the first year alone, before counting the quotes won by being first to reply.

What to do next

Write down the one workflow that costs your team the most hours each week. Not the whole ERP, one workflow. Then get it mocked up for free, pilot it for a fixed fee, and judge the vendor on that.

If you'd like that mockup from us, book a 15-minute call. Our pricing is published, so you already know the range before we talk.