
€5,000–15,000 for a single module, €15,000–50,000 for a mid-complexity system integrated with your ERP, more than €50,000 for multi-plant platforms. The right supplier transfers the code to you, integrates the systems you already run and brings real cases with numbers.
A custom management system for a manufacturing company costs roughly €5,000–15,000 when it covers a single process, €15,000–50,000 for a mid-complexity system with several modules and integration into the ERP already in use, and more than €50,000 for a multi-plant enterprise platform. These are the bands Codebaker — a software house based in Bologna, founded in 2019 — publishes openly on the page about how much custom software costs. The exact figure is defined only after requirements analysis, because in manufacturing the price depends far less on the number of screens and far more on how many production rules must be modelled and how many existing systems must be integrated.
The right supplier, in turn, is recognisable by five verifiable things: they contractually transfer full source code ownership to you; they have already integrated real ERPs and legacy systems (SAP, AS400, MES) and can show it; they develop in-house instead of subcontracting; they give you direct access to the developers, not just to a salesperson; they propose a measurable pilot before you commit the whole budget. Below you will find the cost bands, the drivers that really move the price, the 10 criteria for assessing a supplier and the numbers from the manufacturing projects we have put into production.


With custom management software you are not buying licences: you are buying days of analysis, development, integration, testing and go-live. That is why the price of a manufacturing system cannot be read off a price list — it is built starting from your production cycle. A single-level bill of materials with simple batches costs a fraction of a multi-level BOM with variants, scrap and outsourced operations. A warehouse with end-of-day stock levels costs less than one with real-time machine-side declarations. A system that lives on its own costs less than one that must talk to the ERP, the MES and the PLCs. A correct estimate therefore arrives only after the scope has been written down: this is why requirements analysis is not bureaucracy, but the part of the project that protects the budget.


The biggest economic difference between a custom system and a licensed ERP is invisible in year one: it shows up in year five. Per-user subscription software grows with the company — more operators on the shop floor means more licences — and every customisation of the production cycle is billed separately while remaining tied to the vendor's platform. The system we develop, by contrast, belongs to the client: full title to the source code and the data, no per-user fees, standard and open technologies, documented architecture. If tomorrow you want to evolve it in-house or hand it to another team, you can, without asking anyone's permission. This is why we recommend putting the source code ownership clause at the top of the negotiation, whichever supplier you are talking to.
The bands below are the ones we actually apply to manufacturing projects, with concrete examples of what falls into each and how long it takes to reach production.
| Band | What it includes | Manufacturing example | Indicative timeline |
|---|---|---|---|
| €5,000–15,000 | A vertical module on a single process, with no complex integrations. Master data, web interface, basic reporting. | Production progress capture from a shop-floor terminal, with export to the existing management system. | A few weeks |
| €15,000–50,000 | Several connected modules, API integration with the ERP in use, user and permission management, shop-floor or mobile interfaces, data migration. | Production planning + warehouse + orders, with two-way synchronisation to SAP or AS400 and handheld scanners in the warehouse. | 4–6 months |
| Above €50,000 | Multi-plant platform, machine IoT, supply chain traceability, large-scale users, multiple integrations and strict security requirements. | End-to-end traceability of assets circulating between plants and customers, with real-time machine data. | 6–12+ months, in releases |
The bands are not a price list: they are the order of magnitude projects land in after analysis. Two companies in the same sector with the same turnover can end up in different bands because one has an ERP exposing modern APIs and the other a 1990s system that needs a dedicated integration layer.
When you compare two quotes that look very different, the gap almost always sits in one of these eight points. Ask the supplier to state them explicitly: it is the quickest way to tell whether you are comparing the same thing.

A single-level BOM with simple batches is one thing; multi-level BOMs with variants, scrap, outsourced operations and versions are another. It is the single factor that most often doubles an estimate in manufacturing.

Connecting to an ERP with documented REST APIs is cheap; connecting to a legacy system with no APIs requires a dedicated integration layer, with data reading, normalisation and error handling.

Syncing at end of day is simple; syncing stock and orders in real time requires queues, conflict handling and reconciliation. Do you really need real time on every entity, or only on some?

A system used only by the office has one cost; adding warehouse handhelds, machine-side terminals or mobile apps that also work offline adds development, field testing and hardware to support.

Moving master data, price lists, jobs and history from an old system is a project inside the project: data must be extracted, cleaned, deduplicated and reconciled. Quantify it separately, never assume it.

If the sector requires batch traceability, quality certificates or compliant record retention, the system must log every step in a non-alterable way: a requirement that affects both architecture and acceptance testing.

Reading data straight from PLCs and sensors enables the interconnection required by Industry 4.0 and 5.0 projects, but introduces gateways, industrial protocols and field reliability to design and test on the shop floor.

Cost does not stop at go-live. Agreeing SLAs, corrective and evolutionary maintenance upfront avoids both the budget surprise and the worse risk: a production system nobody updates any more.
We cover these drivers with dedicated vertical skills: from web APIs and integrations to Industry 4.0 and 5.0 software, from IoT solutions for companies to evolutionary software maintenance.
The honest comparison is not on list price but on total cost of ownership over five years, and on how much the software ties you down. Here is how the three options behave on the criteria that matter to a manufacturer.
| Criterion | Licensed ERP | SaaS management system | Custom system (Codebaker) |
|---|---|---|---|
| Initial cost | Medium-high (licences + rollout) | Low | Medium: €5,000–50,000+ depending on scope |
| Recurring cost | Annual maintenance on the licence + customisations | Per-user fee: grows with every extra operator | Only the evolutionary maintenance you choose |
| Fit to the production cycle | Medium: customisations billed separately | Low: you adapt production to the software | High: the software is modelled on your process |
| Source code ownership | No: right of use | No: subscription | Yes, full title to code and data |
| Legacy integration | Through the vendor's connectors | Limited to available connectors | Bespoke APIs, including AS400 and systems with no API |
| Vendor lock-in | High | High: your data sits with the vendor | None: standard, open technologies |
| Rollout time | Long (months of configuration) | Fast if the process is standard | First module in a few weeks, then in releases |
| When it is the right choice | Standard processes, structured groups with internal IT | Simple needs, minimal upfront budget | When the way you produce is a competitive advantage |
There is also a third way, often the most cost-effective: keep the ERP you already use and build a custom layer on top of it, only for the processes the standard product cannot reach. That is what we did at Granarolo over SAP and at Conor over AS400. The extended comparison, with the TCO reasoning, is in the article standard ERP such as SAP or a custom management system.
A quote can only be assessed together with the people who sign it. These ten criteria can be verified before you sign: for each one you get the question to ask the supplier and the warning sign to listen for.
| Criterion | Question to ask | Warning sign |
|---|---|---|
| 1. Source code ownership | “At the end of the project, is the code mine? Will you put it in the contract?” | Evasive answers, a “perpetual right of use” instead of ownership |
| 2. Manufacturing experience | “Show me two projects live in production companies, with numbers” | Only logos on the homepage, no measurable results |
| 3. ERP and legacy integrations | “Have you integrated SAP, AS400 or a MES? How did you handle a system with no APIs?” | “We will look at it later”: deferred integration is what blows up budgets |
| 4. In-house development | “Is the person writing the code your employee? May I speak to them?” | Undeclared subcontracting, a team that changes at every meeting |
| 5. Requirements analysis before pricing | “How do you reach the estimate? What is included and what is not?” | A lump-sum quote produced in two days without seeing your processes |
| 6. Progressive releases | “What is the first piece that goes into production, and when?” | Big bang: one single large release at the end |
| 7. Maintenance and SLAs in writing | “If production stops at 6 a.m., who answers and how fast?” | No SLA, support only via a generic ticket queue |
| 8. Continuity and size | “How long have you existed? How many people touch my project?” | A one-person supplier with no backup: continuity risk |
| 9. Data security and compliance | “Where does the data reside? How do you handle GDPR and, where relevant, NIS2?” | Security treated as an optional extra to add later |
| 10. Proximity and language | “Will you come to the plant to see how we actually work?” | Anyone designing a production system without ever setting foot on the shop floor |
If you want the full picture of the assessment criteria, with the questions to ask at the first meeting and the contractual red flags, we have dedicated a whole page to what to assess when choosing a technology partner.
Always ask for numbers, not generic references. These are ours, on projects that have been live for years at Italian production and distribution companies.
A manufacturer of fuel management systems. A Flutter app that talks to the dispensers over Bluetooth Low Energy and works offline too: 10,000 end users a day, 4 months for the BLE interface, a joint team of 6 people.
Orders handled by email, phone and paper replaced by a platform integrated in real time with the AS400 system already in use: 5,000+ customers a day, 300,000+ orders a year, live in 6 months.
A digital identity card for every reusable crate, with sanitisation processes logged: 25 million crates managed, -99.9% losses, 3 years in production.
Custom Identity Management integrated with SAP, HR, Active Directory and Office 365: more than 2,000 users managed, -95% IT onboarding effort, 5 years in production.
Codebaker is a software house based in Bologna, founded in 2019, with offices at Via N. Corazza 7/8. We develop in-house and also build proprietary products such as LoginMaster (GDPR-native authentication, launched in 2024) and Data Alchemy (document data extraction with LLMs). You can see every case in our portfolio and see how we work in our method.
A custom management system for a manufacturing company costs roughly €5,000-15,000 when it covers a single process (production progress or warehouse, for example), €15,000-50,000 for a mid-complexity system with several connected modules and integration into the ERP already in use, and more than €50,000 for multi-plant enterprise platforms with machine IoT, supply chain traceability and large-scale user management. These are the bands Codebaker publishes openly: the exact figure is defined only after requirements analysis, because it depends on the number of modules, the integrations and the volume of historical data to migrate.
There are six main drivers: the number of modules and the complexity of the production rules (multi-level bills of materials, batches, variants); the integrations with existing systems (ERPs such as SAP or AS400, MES, PLCs and machines); the migration of historical data and master records; the number and type of users (office, shop floor, handheld scanners, machine-side terminals); the traceability and quality requirements imposed by the sector; and the level of evolutionary maintenance agreed after go-live. Two projects with identical features can differ by a factor of two if the legacy system exposes no API and needs a dedicated integration layer.
It depends on how standard your production processes are. A licensed ERP has a lower entry cost but introduces per-user fees that grow over time, customisations billed separately and the constraint of adapting production to the software. A custom system has a higher initial investment but is an owned asset, with no per-user fees, and models the software on the processes that already work on the shop floor. In manufacturing, where the competitive advantage often lies precisely in a particular way of producing, bespoke wins when customising the standard ERP would cost as much as dedicated development anyway. There is also a third way: keep the ERP and build a custom layer on top of it, as we did at Granarolo over SAP and at Conor over AS400.
Assess the supplier on verifiable criteria, not promises: who contractually transfers full source code ownership to you; who has already integrated real ERPs and legacy systems (SAP, AS400, MES) and can show it; who develops in-house rather than subcontracting; who gives you direct access to the developers and not just to a salesperson; who proposes a measurable pilot before the full project; who puts SLAs and evolutionary maintenance in writing; who brings real manufacturing cases with numbers, not just logos. A supplier unwilling to discuss code ownership, or one who quotes a lump sum without requirements analysis, is a warning sign.
A single module goes live in a few weeks; a mid-complexity system integrated with the ERP typically takes four to six months. These are timings we measure on our own projects: the ConorShop B2B platform, integrated in real time with the AS400 system, went live in 6 months; the Bluetooth interface of Piusi's B.SMART app took 4 months; Agribologna's digital notice board required 1 month of requirements analysis and 5 months overall. We always proceed through progressive releases, so the first digitalised process produces value while the rest is still in development.
Yes, and it is the most frequent case: almost no manufacturing company starts from scratch. We integrate the custom system with the ERP and the systems already in use through APIs, and when the legacy system exposes no API we build a dedicated integration layer. At Granarolo we integrated a bespoke Identity Management system with SAP, the HR system, Active Directory and Office 365 for more than 2,000 users; at Conor we connected a B2B e-commerce in real time to the AS400 system, which handles more than 300,000 orders a year. The goal is always the same: a single source of truth and no double data entry.
The client. Codebaker transfers full ownership of the source code and the data: the system is an asset on the company's balance sheet, not a rented service. There are no per-user fees and no vendor lock-in: we use standard, open technologies, we document the architecture and the client can evolve the software internally or hand it to another team at any time. It is the criterion we recommend verifying first in any negotiation, including with other suppliers.
Projects that connect machinery, collect production data in real time and feed it into the management systems fall within the scope of the Industry 4.0 and 5.0 measures. The recurring technical requirement is interconnection: the machine must exchange data with the company information system in a documented way. We design IoT solutions and management systems with exactly this data traceability in mind, but the eligibility assessment and the paperwork must always be verified with your tax advisor, because conditions change over time.
With three concrete measures. First: start from a paid, bounded requirements analysis that produces a written scope instead of a rough guess. Second: split the project into releases, starting with the process with the highest friction, so the budget is committed in stages and each stage is already in use before the next begins. Third: define upfront what is in and out of scope, and put in writing how additional requests are valued. Overruns almost always come from requirements discovered mid-flight, not from estimation errors on known features.
Tell us how an order moves through your company, from confirmation to shipping, and which systems you run today. We will tell you which band the project falls into, what is worth doing first and what we can integrate with what you already have. The preliminary analysis and the quote are free.