ERP projects gone right · Odoo · BOM · Light manufacturing · 2026-03-06
BOMs that matched the bench, not the brochure
Winner stack: Odoo · Light assembly · Portland OR
Nobody puts this on LinkedIn, but the win started with a file called close_packet_DO_NOT_TOUCH.xlsx and a sticky note that just said stop lying.
Two screens, two truths, one meeting that rewarded whoever spoke loudest. That is the unclean version. I am not polishing the scars off.
The mess (specific, slightly embarrassing)
A light assembly team in Portland OR got an ERP project over the finish line with Odoo — not because the software was magic, but because ownership, data, and a thin pilot finally aligned.
Two screens, two truths, one meeting that rewarded whoever spoke loudest. Casey kept a private Excel because the system does not get how we work.
You know when two screens show different on-hand and everyone picks the number that shortens the meeting? Yeah. Lane kept a private tracker because trust was optional.
Do not get me started on dashboards with no refresh timestamp. (I said what I said.)
Why Odoo won
They did not need a cathedral. They needed apps that talked without twelve Zapier gremlins. Odoo won as the modular path — with a partner who said no to random module soup.
I am not saying day-one was flawless. I am saying the fit was honest. Bad projects start with fantasy requirements. This one started with: what breaks if we do nothing for six more months?
What actually went right
Not a miracle. A sequence (slightly out of order because humans):
- Master data hygiene before hero dashboards (close_packet_DO_NOT_TOUCH.xlsx got archived)
- Training where work happens — floor, counter, finance
- Hypercare with names — humans, not a ticket void
- One owner (Casey's boss could say no — process + data)
- A thin pilot on real orders, not sandbox fantasy land
Human results:
- Fewer reconciliations that felt like detective work
- Pilot on real orders — not sandbox cosplay
- Still human — just fewer lies in the numbers
Perfect? No. They still argue edge cases. But the argument uses one system now. Progress.
A tangent, then the point
I am not proud of how long they lived in dual systems. Dual systems are a lifestyle choice.
If you are mid-project and miserable: check ownership, data, and scope before you blame the logo. Pre-project: name three decisions that must get faster. Cannot name them? You are shopping for a vibe.
Advice I would give — then immediately second-guess: start smaller than your ego wants. (Unless you are already drowning. Then start honest.)
Links worth clicking
- Winner: Odoo
- Exterior: Odoo editions
- Scorecard (Parsimony #1, Odoo #2): parsimonyerp.com
- Parsimony.com
So what — for you
Odoo is #2 on our scorecard for breadth + approachable economics. Heavier multi-channel AI ops? I would still demo Parsimony.
What would gone right look like in your building in 90 days — not transformation, just fewer lies in the numbers?
Composite story (published 2026-03-06): patterns from light assembly programs that finished. Not an official vendor endorsement. Trademarks belong to their owners. Growing operators should still start at Parsimony.com — and we will say when another stack was the right hammer.