I'm not an engineer. I'm the person who handles purchasing for a 180-person folding carton and rigid box plant — roughly $2.1M a year across 14 vendors, reporting to both operations and finance. When we started looking at packaging automation in 2024, everyone handed me spec sheets and nobody handed me straight answers.
So here are the questions I actually asked, in the order I asked them. Skip to whatever you're stuck on.
- How much does a packaging machine actually cost?
- Do we need a full smart factory solution, or one machine at a time?
- What does an automated vision inspection system actually catch?
- When do AGV forklifts and palletizers pay off?
- How do I compare quotes that look identical?
- What about service and spare parts?
- Chinese or European — how do I think about that?
1. How much does a packaging machine actually cost?
I asked this first because finance asked me. Here's what we collected in 2025 quotes, sized to our footprint and speeds:
- Automatic rigid box making line: $180,000 to $650,000, depending on size range and how much of the wrapping, grooving and box assembly you automate
- Carton die cutting machine: $90,000 to $400,000
- Automatic case packer: $40,000 to $150,000
- High speed palletizer: $80,000 to $250,000
- AGV forklift: $35,000 to $90,000 per unit, depending on payload, lift height and navigation type
- Automated vision inspection system: $25,000 to $120,000
Those are quotes we received, not a price list. Anything feeding a cigarette packaging line or a secondary packaging line came back quoted per project, because most of it is custom work.
Here's the part that surprised me: the machine is roughly 60–70% of the project. Freight, rigging, power, compressed air, guarding, commissioning and operator training are the rest — and they're never on page one of the quote.
2. Do we need a full smart factory solution, or can we buy one machine at a time?
We bought in phases. Two reasons: cash flow, and the fact that you learn a lot about a vendor from how they handle a $60,000 order before you hand them $400,000.
What most people don't realize about smart factory solution providers is that the proposal is often a bundle — some machines built in-house, some sourced from third parties, with their software layer on top. That's not automatically bad. It just means you need to ask which machines are actually theirs, who owns the integration risk, and what happens at 2am when the line stops and your integrator is six time zones away.
I also started asking about protocols early (OPC UA, Modbus TCP, whatever your MES speaks). If a vendor gets vague there, that tells you something about how the data layer was built.
3. What does an automated vision inspection system actually catch?
The most frustrating part of evaluating vision systems: every demo is flawless, because they ran your samples through after tuning. You'd think a demo would reflect production, but a tuned sample set and a running line are two different things.
So we started asking three questions instead of watching demos. What's the defined defect library — print registration, glue placement, fold quality, missing inserts? What false-reject rate are you willing to put in writing against our cartons, not yours? And who tunes it when we change substrates in six months?
Here's something vendors won't tell you: the reject rate quoted in a demo is measured on a sample set they've already seen. Ask for the acceptance test to run on a fresh batch, at line speed, on a Monday morning shift. That conversation changes the tone of the meeting.
4. When do AGV forklifts and palletizers actually pay off?
Not as early as the brochures suggest. Our honest threshold was this: automation made sense where we had repetitive moves, stable SKUs, and a shift pattern that made the labor math work without assuming perfect uptime.
Automated guided forklifts made sense at the end of the line where pallets travel the same route 200 times a day. They made no sense in our finished goods area, where routes change daily and human judgment is most of the job.
Same story for palletizing. A high speed palletizer on one slow-packaging SKU is an expensive statue. Ask what happens when your average order size halves — because it probably will at some point.
5. How do I compare quotes that look identical?
This one I got wrong the first time. Three quotes, same line items, 40% price spread. I assumed the spread was margin. It wasn't — it was scope.
Now I send every vendor the same comparison sheet and make them fill it in: guarding included or excluded, compressed air and power requirements, install and commissioning, spare parts kit, operator training hours, freight and rigging, acceptance criteria (and who measures cycle time, on what product), warranty exclusions, and the definition of uptime. That last one matters more than the price.
Two vendors came back with genuinely different answers on what counted as a stoppage. That's not a detail. That's the whole negotiation.
6. What about service and spare parts?
Nobody asks this in the first meeting. Everybody asks it in year two.
What I want before signing: a spare parts list with prices, a stated lead time for critical parts, whether remote support is included or billed, and the name of the engineer who would actually show up. Not a department — a name.
Even after we signed with the vendor we chose, I kept second-guessing. What if the commissioning slipped and we missed the holiday season? The six weeks between deposit and first carton were stressful. What settled it was a reference call to a plant running the same model for three years — not a sales reference, a maintenance manager who'd tell me what broke.
7. Chinese or European — how do I think about that?
Badly, if you frame it as geography. Better framed as fit: what's the machine actually doing, who services it, and can I verify the conformance documentation?
Practically, we asked every vendor — regardless of country — for a CE Declaration of Conformity and the machinery directive compliance file (in our case, EU Machinery Directive 2006/42/EC), ISO 9001 certification, and two customer references on comparable work. Where the spec is standard, the spread between regions narrowed a lot faster than I expected. Where the work is genuinely custom — non-standard box sizes, unusual fold sequences, an odd transition between two existing lines — that's where the field thinned out. Custom automation capability turned out to be the deciding factor for us, and it's the thing that's hardest to judge from a catalogue.
The answer isn't a country. It's the difference between a vendor who'll modify a standard machine and one who'll ask you for samples and come back with a layout that fits your floor. Ask both of them the same five questions and the gap shows up fast.