The importance of a centralising ERP for SMEs

Friday, 5:30 pm. The finance manager of a garment manufacturer in Famalicão with 75 employees is closing the month. He has four Excel files open, the invoicing program in a separate window, and an email from the external accountant with a different version of the figures. No one is lying — the data is all correct, it just doesn't talk to each other. This guide shows, step by step, how to assess whether your SME needs a centralising ERP and what to check before proceeding.

What you need before speaking to any vendor

Without this preliminary work, any demonstration of ERP MULTI or another platform will be theatre. You'll see attractive features without knowing whether they solve your real problem.

• List every active tool: who uses it, what data it produces, how often.
• Identify the manual bridges: where files are exported and re-imported between systems.
• Quantify the cost of reconciliation: hours per week spent cross-checking data between departments.
• Confirm who decides: CEO, CFO and IT manager must be aligned from day one — not from the contract stage onwards.
• Check the active regulatory obligations: DL 28/2019 (ATCUD, AT certification) and Portaria 195/2020 (monthly SAF-T) are not optional. The ERP must comply with both, without subsequent adaptation.
• Define the three-year growth horizon: new markets, new product lines, internationalisation, opening a remote warehouse.
• Estimate the monthly transaction volume: orders, invoices, production orders, warehouse movements. This figure defines the class of platform you need.

Step 1 — Diagnose the current fragmentation

Go through each operational area and answer one simple question: where does the data I use to make decisions today come from? Don't answer from memory. Sit down with each area manager for 30 minutes and document what you see.

• List all data sources by area: finance, purchasing, production, sales, warehouse, HR.
• Flag those that don't communicate with each other in real time.
• Identify where there is duplication of data entry — the same data entered into two different systems.
• Record the recurring errors caused by that duplication over the last 90 days.
• Calculate the time lost per week in each area reconciling information.

In a typical Northern garment factory with 60 employees, it is common to find five or six systems without a direct link: the invoicing program, the production Excel sheet, the external accountant's software, the manual attendance record, the customer's order file and the email with the purchase orders to the supplier. The invisible cost of this model is not in the software — it's in the hours of middle management building bridges that a machine should be building. When that time is measured, it rarely comes to less than 15 to 20 hours per week in companies of this size.

Document everything on a simple sheet. Three columns are enough: area, data source, reconciliation hours per week. That document will be the strongest argument in the meeting with the CEO and the CFO.

• ☐ Data sources per area identified and listed
• ☐ Manual bridges quantified in hours per week
• ☐ Recurring errors documented with estimated financial impact

Step 2 — Define the minimum functional perimeter

A centralising ERP doesn't need to do everything on day 1. It needs to cover the core that eliminates the most expensive bridges. Defining that core before seeing any demonstration saves you months of evaluation and prevents the vendor from selling you modules you don't need yet.

• Identify the three information flows with the highest cost of fragmentation — the ones you found in Step 1.
• Separate what is mandatory (AT-certified invoicing, SAF-T) from what is desirable (BI, sales mobility, B2B portal).
• Check whether the sector has specific requirements: lot traceability in textiles, multi-axis SKU management in footwear, warehouse picking in distribution.
• Document the exceptional processes the ERP has to support. Don't ignore them because "they're rare" — those are precisely the ones that cause crises at go-live.
• Define what falls outside the initial perimeter and when it comes in: phase 2, phase 3, with indicative dates.

The right ERP for an SME is not the most complete one — it's the one that covers the financial-operational flow without requiring the company to change what already works well.

This exercise has a useful side effect: it forces the CEO-CFO-IT triad to align priorities before entering negotiations with vendors. When that alignment doesn't exist, the project starts with different expectations on each side of the table — and ends badly.

• ☐ Three priority flows identified by cost of fragmentation
• ☐ Mandatory requirements separated from desirable ones
• ☐ Exceptional processes documented
• ☐ Phase 1 perimeter defined and approved by the CEO-CFO-IT triad

Step 3 — Assess the vertical fit, not just the functional fit

A generalist ERP covers 80% of any company's needs. The other 20% is exactly where your operation lives. In footwear, it's the colour-size-lining axes — a collection of 800 to 1,200 SKUs with three axes of variation that a generic ERP models with workarounds. In textiles, it's lot traceability for compliance with the EU Strategy for Sustainable and Circular Textiles. In distribution, it's the WMS integrated with picking and cross-docking, not a basic stock module renamed.

• Check whether the ERP has native modules for your sector — not generic configurations built on top of a standard stock module.
• Ask the vendor for references from clients in the same sector and of a similar size. Contact them before signing.
• Test the data model for your most complex cases before signing the contract. If the vendor hesitates, that's a sign.
• Confirm whether the vendor's team has knowledge of the sector — not just of the software. The difference shows in the first two hours of conversation.

We see in INFOS projects that companies which skip this step reach go-live with a list of "exceptions" that the ERP doesn't support natively. Each exception is an adaptation. Each adaptation has a cost and a lead time. To explore this point further, read the guide Industrial ERP Portugal: how to assess vertical fit before signing a contract.

• ☐ Sector modules identified and tested in a real demonstration
• ☐ References in the sector requested and contacted
• ☐ Complex cases tested in a demonstration environment with the company's real data

Step 4 — Analyse the integration architecture

Centralising doesn't mean replacing everything. It means there is a single source of truth. Some peripheral systems — the store POS, the shop-floor terminal, the B2B portal — can and should continue to exist, provided they feed and receive data from the ERP in real time. The problem is not having several systems. The problem is that they don't communicate.

• Map the systems that will integrate with the ERP and those that will be replaced. These are different decisions with different impacts on the team.
• Check whether the ERP has documented APIs for the systems that stay. "We can do it" without technical documentation is a promise, not a guarantee.
• Confirm how synchronisation between branches or remote warehouses works — especially if you have operations in locations with unstable connectivity.
• Assess whether you need real-time production capture and whether the ERP supports industrial shop-floor terminals. For this layer, see what KORA Productivity adds to the ERP in terms of OEE and operational efficiency.
• Check the management of operational KPIs: the ERP produces the data, but it needs a BI layer to make it actionable in good time for management.

In a distribution operation with an 8,000 m² warehouse, the ERP centralises finance and purchasing, but warehouse management requires a dedicated integrated KORA Inventory Suite module — with picking, packing list and location control. An Excel sheet running parallel to the ERP in this context is a step backwards disguised as a transitional solution.

Document the integration map before finalising the tender specification. A simple diagram with boxes and arrows — current systems, new systems, direction of data flow, synchronisation frequency — avoids surprises during the implementation phase.

• ☐ Integration map between systems drawn up and validated
• ☐ ERP APIs verified for each peripheral system that stays
• ☐ Synchronisation between locations tested in a proof of concept
• ☐ BI need identified and included in the perimeter from the outset

Step 5 — Plan data migration and training

Most ERP projects that go wrong don't fail on the software. They fail on the historical data and on the people. Treat both with the same rigour you apply to choosing the platform — and with the same reserved budget.

• Audit the quality of the current data before migrating: duplicates, empty fields, inconsistent encodings, items with duplicate references. This work takes longer than expected.
• Define what migrates — typically active data from the last two to three years — and what stays in a consultation archive.
• Plan training by user profile. The warehouse manager needs different training from the accountant, who needs different training from the sales director. Generic training for everyone is effective training for no one.
• Appoint an "internal champion" per department: someone who learns first, validates the processes in the new system and supports colleagues in the first weeks after go-live.
• Define a phased go-live plan. Launching finance, production, warehouse and sales at the same time in an industrial SME is the highest-risk scenario. If you can phase it, phase it.
• Reserve a minimum buffer of 20% in the schedule for migration contingencies. It's not pessimism — it's the norm in projects with complex historical data.

The warehouse chief who works with the system every day will be the project's greatest ally or greatest obstacle. It's not a question of resistance to change — it's a question of trust. If the system fails in the first weeks because the data was dirty, you lose that trust and it takes months to recover it. Clean the data first. Migrate afterwards.

• ☐ Data quality audit completed before starting migration
• ☐ Migration perimeter defined and approved
• ☐ Training plan by profile documented with hours and responsible persons
• ☐ Internal champions identified by department
• ☐ Phased go-live plan approved by management with dates and validation criteria

Decision matrix: centralised ERP vs. maintaining the current model

Common mistakes — and what causes them

• Choosing on the licence price, ignoring the total cost. Add implementation, training, migration, annual maintenance and the cost of sector adaptations. The licence price is rarely the largest item on the final invoice. In medium-sized industrial projects, implementation and training typically represent double or triple the value of the licence.
• Configuring the ERP to imitate the current Excel. If the process was wrong before, the ERP will automate the error with more speed and scale. Review the processes before configuring — not during and not after.
• Ignoring the production module because "it's complicated". That is exactly where the biggest gain in visibility lies in an industrial SME. Postpone the financial module if you must — not the production one. Finance already works, even if badly. Production without visibility is money going out with no record.
• Not involving the warehouse manager in the project. The strongest resistance comes from those who work with the system daily — not from those who approved it in a meeting. Involve them from the requirements phase, not in go-live week.
• Going live in production with dirty data. Migrating inconsistent data to a new ERP is the fastest way to lose the team's trust in the system. When the stock in the ERP doesn't match the physical stock in the first week, the warehouse chief goes back to Excel — and is right to do so. Clean first, migrate afterwards.
• Underestimating the impact of change on operational-profile users. The finance director adapts in days. The warehouse operator who has worked with a system for 12 years needs real support, not a PDF manual.

A note on funding

ERP implementation projects in Portuguese industrial SMEs may be eligible for co-funding via PT2030, COMPETE 2030 or Norte 2030, depending on the location, size and nature of the investment. IAPMEI and the regional CCDRs are the correct entry points for assessing eligibility. The application requires a detailed technical report — which reinforces the need to document the Step 1 diagnosis well from the outset. A well-done diagnosis serves simultaneously as an internal argument and as the basis for the application's descriptive memorandum.

For the textile and garment sector, ATP provides information on specific sector instruments. For footwear, APICCAPS has its own channels for supporting digitalisation.

The next step

Complete the fragmentation diagnosis from Step 1 before speaking to any vendor. Take the results to the meeting with the CEO, CFO and IT. An A4 sheet with hours lost per week, recurring errors and estimated financial impact is worth more than any generic RFP sent to six vendors at once.

To understand how a vertical ERP differs from a generalist solution in the Portuguese industrial context, read Industrial ERP Portugal: how to assess vertical fit before signing a contract. If your sector is textiles or garments, the guide Textile production management: a guide for industrial directors directly complements this journey.

Sources

• Decree-Law No. 28/2019, of 15 February — Diário da República, 1st series, No. 32 (electronic invoicing obligations and software certification by the AT)
• Portaria No. 195/2020, of 13 August — monthly SAF-T communication to the Tax Authority
• Law No. 93/2021, of 20 December — general regime for the protection of whistleblowers, mandatory channel for organisations with 50 or more workers
• Regulation (EU) 2024/1689 — AI Act, risk classification of artificial intelligence systems
• European Commission Communication COM(2022) 141 — EU Strategy for Sustainable and Circular Textiles