Short answer: smart manufacturing is production that measures itself, feeds that back, and adjusts. That is the whole idea, and it is genuinely valuable. What makes the topic exhausting is that it arrives wrapped in a vocabulary built for conference keynotes: Industry 4.0, smart factory, IIoT, digital twin, dark factory, predictive everything. Underneath those words sits a ladder, and almost everyone benefits from the bottom two rungs while almost nobody is ready for the top one. This page sorts the terms, gives the ladder, and says plainly where the money is.
The vocabulary, sorted
They are not synonyms, though they get used as if they were.
- Industry 4.0. German industrial-policy framing: the fourth industrial revolution, after steam, electricity and computing. Broad, political, and useful mostly as an umbrella.
- Smart manufacturing. The more operational term for the same shift. Production that senses, records, analyses and adapts.
- Smart factory. One connected site where those capabilities are actually in place. A place, not a philosophy.
- IIoT (Industrial Internet of Things). The plumbing: sensors, gateways and connectivity that get machine data out of the machine.
- Digital twin. A live virtual model of a machine, line or product, fed by real data, used to simulate before you commit in the real world.
- Predictive maintenance. Using machine data to intervene before something breaks, rather than on a schedule or after a failure.
- Dark factory. Production that runs unmanned. Real in a handful of industries, a thought experiment in most.
Notice that only the last three are technologies. The first four are framings, and framings do not have a price or a payback.
The maturity ladder
The order matters more than the technology, because each rung is only worth what the rung below it supports.
- One system for the backbone. Sales, purchasing, inventory, production and finance on one database. Unglamorous, and the precondition for everything else: without it you are analysing fragments.
- Digital reporting on the floor. Operators record operations, quantities, scrap and quality as it happens instead of on paper. This is the MES layer, and it is where most mid-sized manufacturers get their largest single jump in insight.
- Measurement. Now that reporting is reliable, real metrics become possible: OEE, downtime reasons, actual versus planned times feeding back into job costing and planning.
- Machine connectivity. Sensors and machine data where it pays: high-value machines, high-frequency processes, or measurements people cannot capture reliably by hand.
- Prediction and self-adjustment. Predictive maintenance, digital twins, automated parameter adjustment. Genuine, and genuinely demanding in data and expertise.
Most mid-sized manufacturers we meet are somewhere between rung one and two. That is not a criticism; it is where the value is. The gap between “we find out next week” and “we know now” is worth more than the gap between “we know now” and “we predicted it”.
Measurement comes before intelligence: the OEE check
Rung three is the pivot, and OEE is the metric everyone reaches for. It is availability x performance x quality, and its real value is not the percentage but which of the three is dragging you down.
Time
One shift, one machine.
Output
What came off in that time.
How this was calculated
- Availability = run time / planned time
- Performance = (ideal cycle x total count) / run time
- Quality = good units / total units
- OEE = availability x performance x quality
One machine, one shift. The honest caveat: OEE is only as good as the data behind it. If downtime is only logged when someone remembers, or planned time excludes the hours things went wrong, the percentage looks great and means nothing.
Two things worth noticing. First, the three factors multiply, so 90% on each gives 73%, not 90%. That is why OEE feels harsh the first time you calculate it, and why the widely cited 85% is genuinely world class rather than a target you should expect to hit next quarter.
Second, if you set an ideal cycle time that is slower than the machine can actually run, performance climbs over 100% and your OEE flatters you. That is the single most common way these numbers get quietly wrong, which is a good preview of the honest part.
Our opinion: most Industry 4.0 spend is a measurement problem wearing a technology costume
Our position, stated plainly, and consistent with what we say about MES and scheduling: the pattern repeats because the underlying mistake repeats.
Sensors are easy to buy and dashboards demo beautifully. Neither fixes the thing that actually limits most factories, which is that nobody reliably records what happened. Put a real-time dashboard on top of a floor where downtime gets logged when someone remembers, and you have built an expensive instrument for displaying guesses. Precisely, in colour, on a big screen.
Four positions we will defend:
- A dashboard nobody acts on is a screensaver. Before buying visibility, decide who will look at the number, how often, and what they are authorised to change. If there is no answer, the project is decoration.
- Predictive maintenance is oversold to companies that have not tried preventive maintenance. Prediction needs machine data over long periods and enough failures to learn from. Planned preventive maintenance with honest logging captures most of the benefit at a fraction of the cost. Start there; let the data tell you whether prediction is worth adding.
- Digital twins are real, and rarely your next step. They pay off in high-volume or high-risk process industries where simulating beats experimenting. For a jobbing shop with 200 different products a year, the modelling effort exceeds the benefit.
- The most valuable Industry 4.0 project is usually boring. Getting operators to report every operation the moment it happens, in two taps, is worth more than any sensor. It is also harder, because it is a people problem rather than a purchase.
And the one that costs us work: if someone shows you an Industry 4.0 roadmap before asking how you currently record downtime, they are selling a category, not a solution. We would rather implement rung one and two well and have you come back for rung four in two years than sell you all five now.
What actually pays off, in order
If you want a practical shortlist for a mid-sized manufacturer:
- Highest return: digital reporting per operation, downtime reasons captured with a reason code, actual times flowing back into costing and planning.
- Good return once the above works: OEE per bottleneck machine (not per machine, just the constraint), quality checks recorded in-process rather than at the end.
- Situational: machine connectivity on high-value or high-frequency equipment, energy monitoring where energy is a real cost line, automated storage integration where picking volume justifies it.
- Later, if ever: predictive maintenance, digital twins, autonomous scheduling.
That list is deliberately unexciting. It is also the sequence we have watched work.
Where Odoo fits, and where it does not
In Odoo, rungs one to three are covered directly: one database for the backbone, Shop Floor for digital reporting, Quality for in-process checks, Maintenance for preventive maintenance, and the IoT box for connecting scanners, scales, printers and simple machine signals.
Where it stops, said plainly: Odoo is not an Industry 4.0 platform. OEE dashboards, high-frequency machine data historians, predictive models and digital twins are not standard functionality, and anyone telling you otherwise is describing a roadmap as if it were a feature. Those live in specialist tools that you connect to Odoo, with Odoo as the system of record.
We think that is the right division. Rungs one to three are where the returns are, and they are exactly the rungs that benefit from sitting on one database with your orders, stock and costs. Rungs four and five are specialist work, and pretending otherwise is how implementations disappoint. The categories and where each fits are mapped in manufacturing software.
Frequently asked questions
What is smart manufacturing? Production that measures itself, feeds the data back and adjusts. A spectrum from reliable shop-floor capture through machine connectivity to prediction, not a single state.
What is the difference between Industry 4.0 and smart manufacturing? Largely the same shift: Industry 4.0 is the German policy framing, smart manufacturing the operational term. Smart factory is a connected site, IIoT the sensor layer, digital twin a live virtual model.
Where should a mid-sized manufacturer start? With reliable digital reporting on the floor, not with sensors. Backbone, then reporting, then measurement, then connectivity, then prediction.
Is predictive maintenance worth it for a smaller manufacturer? Usually not first, sometimes not at all. Preventive maintenance with honest logging captures most of the benefit far more cheaply.
What is OEE and why does it matter? Availability times performance times quality. It is usually the first genuinely useful factory metric, and it exposes which factor costs you most, provided the underlying data is honest.
Want to know which rung you are actually on? Book a free Odoo scan - we look at how you record production today and tell you honestly which step pays off next, and which ones you can safely ignore for now.
Read more: Manufacturing software: the categories · What is a MES? · MES software: how to choose · Production planning and scheduling · MRP system explained · Odoo for manufacturers