← Insights

Industry, production and logistics

Industry, production and logistics: where margin leaks and how to recover it

Half of companies already have an ERP. Time, margin and capital are lost in the orders, invoices and stock that circulate around it.

Published 16 November 2026Updated 16 November 202616 min read

Portuguese manufacturing productivity vs the EU[8]
49%
Manufacturers with a net loss in 2025[6]
33.5%
Manufacturing inventory in days of cost of goods and materials consumed (2025)[6]
≈104 days
Wholesalers using AI, Portugal vs EU[15]
8.6% vs 23.4%

Executive summary

  1. Portuguese manufacturing produces 49% of the European GVA per person, and the gap repeats in every size class [8],[9].
  2. Manufacturing EBITDA margin fell from 11.3% (2023) to 10.3% (2025): staff and external services absorbed the lower weight of raw materials [6].
  3. 33.5% of manufacturers and 49.5% of transport and storage companies reported a loss in 2025 [6].
  4. Manufacturing inventory rose to about 104 days of cost of goods and materials (Vektrum calculation on BdP), more than the 58-day collection period [6].
  5. Portuguese SMEs use ERP as much as European ones, but in 2022 only 24.5% of companies with 10 or more people sent structured invoices, against 38.7% in the EU [16],[17].
  6. In Vektrum's base scenarios, a 30-person manufacturer gains €13,760 a year (+0.45 pp of margin) and releases €38,400 of capital; the assumptions are shown.

01

Conclusions

Executive summary

Portuguese manufacturing accounts for 4.4% of non-financial enterprises but 15.1% of employment and 19.1% of gross value added [3]. Wholesale distribution and logistics run on the same physical flow of orders, stock and invoices. This report analyses the three subsectors using official data from INE (Statistics Portugal), Banco de Portugal and Eurostat, and turns the figures into concrete decisions for an SME with 20 to 50 people.

The central conclusion is counter-intuitive: the Portuguese gap is not a lack of management systems but what happens around them. Half of companies with 10 or more people already have an ERP (48.3%, against 46.5% in the EU) [16], yet orders, supplier invoices and price lists still arrive as PDF, paper or email and are retyped by hand. That is where time, margin and capital are lost.

Portuguese manufacturing productivity relative to the EU average (GVA per person, 2024)[8]
49%
Manufacturing EBITDA margin in 2025, down from 11.3% in 2023[6]
10.3%
Manufacturing inventories in days of cost of goods and materials consumed (2025)[6]
≈104 days
Portuguese companies with 10 or more people sending structured e-invoices in 2022, against 38.7% in the EU[17]
24.5%

Seven conclusions for management

  1. Productivity is about half the European level, and no size class reaches two thirds. Portuguese manufacturing generates €40.9k of GVA per person against €83.8k in the EU (2024); per hour, €24.1 against €54.7 [8]. A Portuguese manufacturer with 250 or more people produces less per person (€60.3k) than the average European firm with 50 to 249 (€67.4k) [9].
  2. Margins are being squeezed by staff and external services. Between 2022 and 2025, staff costs rose from 12.3% to 15.5% of manufacturing sales and external supplies and services from 16.7% to 18.2%; 2025 EBITDA was 8.7% below 2023 [6].
  3. One in three manufacturers loses money. In 2025, 33.5% of manufacturing corporations reported a net loss; in transport and storage, 49.5% [6].
  4. Capital is locked in stock, not only in customers. Manufacturers collect in 58 days and pay in 64, but hold about 104 days of inventory, up from 94 in 2022 (Vektrum calculation on BdP data) [6]. At aggregate level, each day of stock removed releases about €199m.
  5. The ERP is not the problem. Small Portuguese companies use ERP as much as European ones (42.2% against 41.1%) [16], but in 2022 only 21.8% sent structured invoices, against 37.0% in the EU, and 83% of companies with 10 or more people still issued paper invoices [17].
  6. Among the subsectors analysed in this report, wholesale distribution has the widest AI gap. 8.6% of Portuguese wholesalers with 10 or more people use AI, against 23.4% in the EU; Portuguese transport, by contrast, is ahead (14.8% against 11.2%) [15].
  7. Administrative automation pays off, but the margin effect is modest and measurable. In our base scenarios, a 30-person manufacturer gains €13,760 net per year from invoice reading and stock alerts (+0.45 pp of EBITDA margin) and releases €38,400 of capital once; a 20-person distributor gains €16,056 from email orders and accounts payable. These are Vektrum estimates, with the assumptions shown in chapter 6.

02

Structure

The sector: four business models

In 2024, manufacturing had 69,876 enterprises, 750,259 people and €125,636m of turnover [8]. SMEs are, in practice, the industry: 99.4% of firms and 74.1% of industrial employment, but only 46.1% of turnover and 58.8% of GVA (2023) [4]. The largest division is food, with 15.2% of industrial sales, and the five largest divisions account for 48.2% of turnover [3].

Talking about "industry" as one block hides business models with very different operational problems. For this report we distinguish four.

  • Make-to-order (metalworking, moulds, machinery): each order is a small project. The bottleneck is usually technical quoting, planning and the hand-off of information between sales, estimating and production.
  • Series production and agri-food (food, textiles, footwear): repetitive volumes, tight margins, shelf lives and traceability requirements. The bottleneck is planning, stock and receiving orders from large customers.
  • Wholesale distribution: thousands of SKUs, dozens of suppliers with their own price lists, and small, frequent orders. The bottleneck is the catalogue, pricing and order entry.
  • Logistics and transport: high aggregate operating margins, but many fragile firms, driver shortages and customers who demand real-time visibility. The bottleneck is route planning, documentation and customer service.
Industrial subsegments in Portugal (2024)
SubsegmentEnterprisesPeopleTurnover (€m)Productivity (€k/person)Gross operating rate (%)
Food (C10)9,58098,01320,12236.87.9
Textiles (C13)3,67545,8374,21329.39.2
Clothing (C14)8,29072,6323,99819.96.6
Leather and footwear (C15)2,52140,4962,56822.26
Metal products (C25)12,391102,26710,39736.711.9
Tools, incl. moulds (C25.73)82012,2271,15039.310.6
Machinery and equipment (C28)1,48028,9683,81843.611.5
Wholesale trade (G46)60,492258,89895,30047.15.3
Land transport (H49)45,917155,92711,852––
Warehousing and support (H52)3,41438,51911,130––

Eurostat SBS. Moulds have no NACE code of their own; class C25.73 is the usual proxy and includes other tools. Gross operating rate = gross operating surplus ÷ turnover. Sources: [8]

Wholesale distribution is almost as large as manufacturing by turnover: 60,492 enterprises, 258,898 people and €95,300m in 2024 [8]. In 2025, wholesale and retail trade made the largest contribution to the growth of business turnover (+1.6 pp out of +4.5%) [1]. In transport and storage, the number of firms grew 14.6% in 2025, to 75,231 [1]. Land transport is highly fragmented (about 3.4 people per firm) and warehousing more concentrated (about 11) (Vektrum calculation on Eurostat) [8].

The productivity gap is not only about scale

Portuguese manufacturing productivity as % of the EU, by size class (2023)
0–9 people65%10–19 people59%20–49 people63%50–249 people59%250+ people55%Total49%
Portuguese manufacturing productivity as % of the EU, by size class (2023)
0–9 people65%
10–19 people59%
20–49 people63%
50–249 people59%
250+ people55%
Total49%

GVA per person employed, Portugal ÷ EU27. Vektrum calculation on Eurostat sbs_sc_ovw. Sources: [9]

In every size class, Portuguese manufacturing sits between 55% and 65% of its European equivalent [9]. If the problem were only size, Portuguese firms in each class would be close to European firms of the same class. They are not. Part of the difference is price and product positioning; part is process, and that is the part an SME controls.

There is an offset: wage-adjusted productivity (GVA ÷ staff costs) in manufacturing is 166.9% in Portugal against 161.9% in the EU (2023) [8]. Per euro of wages, value added is close to the EU level. Vektrum reading: with the minimum wage up 5.7% in 2026 [18] and average pay across all corporations (all sectors) up 6.3% in 2025 against 3.6% productivity growth [1], that cushion is likely to narrow, although the narrowing in manufacturing has not yet been measured.

Exports: a difficult year

The 33,742 exporting corporations are 6.3% of all corporations but generate 34.1% of turnover (2024) [2]. In 2025, goods exports grew 0.5% to €79,312m; sales to the US fell 13.4%, to 5.8% of the total, under the flat 15% tariff on European goods, and sales to Germany rose 14.5% [22]. Metals and metalworking set a new record (€24,169m, +3%, 33% of manufacturing, 75% to the EU) [23]; footwear rose 0.8% to €1,718m, with the US down 12.3% [24]; textiles and clothing fell 0.8% to €5,499m [25]. Moulds exported €505.4m in 2022, the latest year published, and Portugal is Europe's third-largest producer [26]. The food and beverage industry generated about €4,870m of GVA in 2025, according to FIPA's projection [27].

Activity slowed: manufacturing output fell 0.8% in 2025 while the EU's rose 1.8% [10], and the industrial production index fell 2.5% year on year in the third quarter of 2026 [30]. In October 2026, manufacturing confidence rose again [31]. Under the Recovery and Resilience Plan, the 51 Mobilising Agendas had €3.2bn of contracted incentives and 53% financial execution at 30/06/2026 [28].

03

Economics

Margins, costs and capital

The ratios in this chapter are Vektrum calculations on Banco de Portugal Central Balance Sheet aggregates (item ÷ sales and services) [6]. They represent the sector weighted by size, dominated by large firms, not the median SME. They are meant to read trends, not to benchmark a specific company.

EBITDA margin and cost structure, % of sales
Ratio2022202320242025
Manufacturing: EBITDA10.011.310.710.3
Manufacturing: goods and materials consumed64.060.859.358.3
Manufacturing: external supplies and services16.716.817.618.2
Manufacturing: staff costs12.314.014.815.5
Trade (wholesale and retail): EBITDA6.46.36.26.2
Trade: staff costs7.68.08.48.5
Transport and storage: EBITDA16.816.816.016.8
Transport: staff costs18.720.521.222.0

Vektrum calculation on BdP data. Trade = section G (includes retail and motor trade). Sources: [6]

Manufacturing: EBITDA margin bridge 2022 → 2025 (pp of sales)
EBITDA margin 202210 ppLower share of goods and materials+5.7 ppHigher share of external services−1.5 ppHigher share of staff costs−3.2 ppOther income and costs (residual)−0.7 ppEBITDA margin 202510.3 pp
Manufacturing: EBITDA margin bridge 2022 → 2025 (pp of sales)
EBITDA margin 202210 pp
Lower share of goods and materials+5.7 pp
Higher share of external services−1.5 pp
Higher share of staff costs−3.2 pp
Other income and costs (residual)−0.7 pp
EBITDA margin 202510.3 pp

Vektrum calculation on BdP data. Each step is the change in the item's share of sales between 2022 and 2025, signed by its effect on the margin. The residual groups changes in production, other income and other operating costs. Sources: [6]

Profitability and fragility

Manufacturers with a net loss (2025; 31.8% in 2022)[6]
33.5%
Trade companies with a net loss (2025)[6]
36.0%
Transport and storage companies with a net loss (2025)[6]
49.5%
Manufacturers with financing costs above EBITDA (2025)[6]
15.0%

Manufacturing return on assets fell from 11.5% (2023) to 9.7% (2025) and return on equity from 11.2% (2022) to 7.8% (2025) [6]. Transport and storage has the highest aggregate EBITDA margin of the three subsectors (16.8%), yet almost half of its firms lose money. Vektrum reading, not measured: the average, dominated by large firms, hides many small loss-making operators, in a sector where staff costs grew 44.3% between 2022 and 2025 [6].

Against Europe, the gross operating rate in manufacturing is 9.8% in Portugal against 10.5% in the EU; in wholesale, 5.3% in both; in transport and storage, 14.1% against 12.8% (2024) [8].

Working capital: where the cash is

Average terms and inventory (2025)
SubsectorCollection (days)Payment (days)Inventory (days of COGS)
Manufacturing5864104
Wholesale and retail trade385061
Transport and storage4156–

Terms: BdP. Inventory days: Vektrum calculation (inventories ÷ cost of goods sold and materials consumed × 365), an approximate indicator mixing raw materials, work in progress and finished goods. Trade includes retail. Sources: [6]

The average collection period in industry (sections B+C, quarterly series) has stayed at 60 to 61 days since 2024 [7]: there is no sign of spontaneous improvement. The more relevant figure, however, is stock. Manufacturing inventories total €20,773m and rose from about 94 to 104 days between 2022 and 2025 (Vektrum calculation on BdP data) [6]. They equal 15.8% of manufacturing assets and 19.6% of assets in trade. Manufacturers finance their customers for almost two months and their own stock for more than three.

Energy and labour

Non-household prices, €/kWh, excluding VAT and recoverable taxes
2021-S12022-S22023-S22024-S22025-S22026-S1
Electricity PT0.10660.13590.11610.13320.13290.1273
Electricity EU270.12900.21070.20030.19410.18370.1800
Gas PT0.02450.10060.05990.05210.04810.0512
Gas EU270.03020.08670.06910.06400.06050.0616

Eurostat. Electricity: 500–1,999 MWh band; gas: 10,000–99,999 GJ band. Values include non-recoverable charges. Sources: [11],[12]

Energy is no longer the Portuguese disadvantage: industrial electricity costs about 71% of the EU average and gas is about 17% lower (2026-S1) [11],[12]. But gas prices multiplied by 4.1 between 2021 and 2022, and electricity is still 19% above 2021 (Vektrum calculations on Eurostat). Only 11% of companies surveyed by AIP name energy as their main constraint; 25% name labour legislation and workforce, the most cited factor (n=156, not representative by sector) [29].

The hourly labour cost in manufacturing is €16.5 in Portugal against €35.0 in the EU (2025), 13% more than in 2023 [13]. The minimum wage rose from €870 to €920 in 2026 [18]. Across Europe, 13% of heavy-goods driver positions were unfilled in 2025, about 502,000 vacancies; the figure for Portugal is not published in the study's preview [37]. Labour scarcity makes each administrative hour saved worth more than its wage cost suggests: it is capacity that does not have to be hired.

04

Diagnosis

Where time and money are lost

There are no official Portuguese data on the time spent entering orders, preparing quotes or answering status requests. That is itself a finding: most SMEs do not measure these costs, because they are spread across the daily work of several people. What follows combines indirect official indicators, international studies with their scope stated, and what we see in the field.

Portugal vs EU: where the digital gap is (companies with 10+ people, %)
ERP, Portugal (2025)48.3%ERP, EU (2025)46.5%Structured invoice, Portugal (2023)24.5%Structured invoice, EU (2023)38.7%Electronic supply-chain sharing,Portugal (2023)19.2%Electronic supply-chain sharing, EU(2023)24.6%
Portugal vs EU: where the digital gap is (companies with 10+ people, %)
ERP, Portugal (2025)48.3%
ERP, EU (2025)46.5%
Structured invoice, Portugal (2023)24.5%
Structured invoice, EU (2023)38.7%
Electronic supply-chain sharing, Portugal (2023)19.2%
Electronic supply-chain sharing, EU (2023)24.6%

Eurostat isoc_eb_iip and isoc_eb_ics. Invoicing data refer to the 2023 survey, covering 2022. Sources: [16],[17]

The pattern is clear: the core system exists, the link with customers and suppliers does not. In 2022, 83.0% of Portuguese companies with 10 or more people still issued paper invoices, against 70.2% in the EU (2023 survey) [17]. Vektrum reading: the indicator measures sending, not receiving, but if suppliers send paper or PDF, the data reach customers in a format that has to be read or keyed in by hand.

Eight loss points

  1. Manual order entry. Customer orders arrive by email, PDF or spreadsheet and are retyped into the ERP. We estimate that 40% to 60% of a distributor's or make-to-order manufacturer's orders arrive this way and that each takes 8 to 15 minutes (Vektrum estimates, to be validated against the mailbox history). A transcription error can lead to a return, a credit note or a repeat shipment.
  2. Technical quotes. In make-to-order manufacturing, a quote takes 1 to 4 hours depending on drawings, materials and routing (Vektrum estimate). The information needed already exists in previous quotes but tends to be scattered across folders and the memory of one or two people. Vektrum reading: response speed weighs on the customer's decision.
  3. Production planning in Excel. Many plants plan in a spreadsheet fed by hand from the ERP. A typical OEE of 55% to 65% means 35% to 45% of planned time produces no good output, and the gap between OEE as estimated by management and as measured is 10 to 15 points; these figures come from an OEE software vendor and should be confirmed in each plant [33]. At the world's 500 largest companies, unplanned downtime costs 11% of revenue, and the rise in cost per hour of downtime was driven mainly by energy [32]; this is a large-company benchmark, not a Portuguese SME one.
  4. Accounts payable. Each supplier invoice is received, posted, matched to the purchase order and delivery note, and approved. In a mostly North American survey, the average cost per invoice is USD 9.84, the cycle 8.2 days and 18.4% of invoices raise exceptions; best-in-class teams have 79% lower costs and 79% faster cycles, and teams spend on average 21.9% of their time dealing with suppliers [34]. APQC shows top performers process payables at one fifth of the cost of the bottom quartile [35]. There is no public Portuguese benchmark for cost per invoice.
  5. Catalogue and pricing at distributors. An SME distributor manages 2,000 to 20,000 SKUs and receives 2 to 10 supplier price lists per month (Vektrum estimates, to be validated in the ERP or online store). Each list has its own format. Until it is loaded, the company may be selling below the new cost.
  6. Stock. In both manufacturing and trade, inventory lasts more days than the collection period (chapter 3) [6]: stock is the largest use of operating capital. In global retail, inventory distortion (stock-outs plus overstocks) is estimated at USD 1.73 trillion a year [36]; it is a retail figure, useful only as an order of magnitude for distributors.
  7. "Where is my order" requests. We estimate 20 to 60 such requests per week at an industrial or distribution SME (Vektrum estimate, no public source). Each one interrupts someone who must check the ERP, the warehouse and the carrier to give an answer the customer could look up alone.
  8. Invoicing compliance. The law in force accepts PDF as an electronic invoice only until 31/12/2026 [19]; the 2027 budget bill extends that deadline by one year, but it is not yet law [21],[20]. A company that only issues and receives PDFs must know, in time, what changes in its invoicing software, with its suppliers and in its accounting.

05

Levers

Opportunities by subsegment

AI adoption in Portuguese manufacturing is 9.9% among companies with 10 or more people, against 17.3% in the EU [15]. The other indicators exist only for all activities combined: in small companies (10 to 49 people), 9.4% against 17.0% [14]; among Portuguese companies already using AI, the leading application is organising administrative or management processes (40.8%) [5], not production: AI use in production processes is 3.5% in Portugal and 4.1% in the EU (all activities, 10+ people) [14]. The barriers most cited by non-users (all activities) are lack of skills (74.4%) and cost (60.4%) [5].

The concrete opportunity lies in the document and data flows that already pass through the company every day. We organise it into four levers: revenue (respond faster and sell more with the same team), margin (prices and costs updated on time), cost (fewer hours of repetitive work) and capital (less stock and faster collection).

Opportunities by subsegment
SubsegmentRevenueMarginCostCapital
Make-to-orderTechnical quotes in hours, not daysQuote with actual costs from past ordersOrders and invoices read automaticallyInvoice on the day of shipment
Series production and agri-foodAutomatic intake of large-customer ordersOEE and waste measured from existing dataPlanning without manual retyping into ExcelStock-out and overstock alerts per SKU
Wholesale distributionOnline catalogue always up to datePrices reviewed when the supplier changes its listEmail orders posted to the ERPReplenishment based on actual turnover
Logistics and transportDelivery status visible to the customerCost per service and per customerTransport documents and invoices reconciledFaster invoicing after delivery

Vektrum synthesis. Each opportunity is quantified in chapter 6.

Distributors: the widest gap among the subsectors analysed, the most immediate opportunity. With 8.6% AI use against 23.4% in the EU [15] and productivity of €47.1k per person against €79.1k [8], Portuguese wholesalers work with EBITDA margins of about 6% [6]. In this model, one percentage point of margin lost to a late price list is worth more than many administrative hours.

Manufacturers: capital before hours. For a manufacturer, the largest financial lever is stock. On the aggregate ratio, each day of inventory removed releases about €199m across manufacturing (Vektrum calculation on BdP data) [6]; in an SME, the same logic applies at its own scale.

Transport: already ahead, but fragile. Transport and storage saw the largest rise in AI use in Portugal, from 4.9% to 14.8% between 2024 and 2025 [5]. With half of firms loss-making [6], the priority is knowing the real cost per service and invoicing without delay.

06

Quantification

Impact scenarios

The following scenarios apply the official data to two typical profiles and clearly separate data (with a source) from Vektrum assumptions. Every result is calculated from the assumptions shown. The aim is not to promise a return but to show the order of magnitude and the point at which each automation pays off.

On time per invoice: Ardent's USD 9.84 per invoice [34], at the US private-sector hourly cost (USD 46.89/h [38], about twice Portugal's €19.4/h [13]), equals about 12.6 minutes of work. Part of that cost is not labour, so 6 to 10 minutes is a prudent range.

Company profiles used in the scenarios
IndicatorMake-to-order manufacturerWholesale distributor
People3020
Estimated turnover€3,080,000€7,360,000
Reference EBITDA margin10.3%6.2%
Reference EBITDA€317,240€456,320
Estimated average inventory€512,500–

Vektrum calculation. Manufacturer turnover = turnover per person of manufacturing SMEs (INE, 2023) × 30; distributor = turnover per person in wholesale (Eurostat, 2024) × 20, an average influenced by large firms. Margins: BdP 2025 (manufacturing; trade as section G). Inventory = sales × COGS share × inventory days ÷ 365 (BdP 2025). Sources: [4],[8],[6]

Profile A: make-to-order manufacturer, 30 people

Manufacturer, 30 people: supplier-invoice reading and accounts payable

Assumptions

  • Supplier invoices per month300 / 450 / 600Vektrum estimate
  • Manual minutes per invoice[34],[38]6 / 8 / 10Vektrum estimate
  • Reduction of manual time[34]60% / 70% / 80%Vektrum estimate
  • Loaded cost of one administrative hour[13]15 €/h / 18 €/h / 22 €/hVektrum estimate
  • Upfront investment€10,000 / €8,000 / €6,000Vektrum estimate
  • Annual running cost (hosting, licences, support)€2,400 / €1,800 / €1,200Vektrum estimate
  • Reference EBITDA margin[6]10.3%Vektrum calculation
ResultCautiousBaseFavourable
Hours freed per year216504960
Annual value of the hours€3,240€9,072€21,120
Annual running cost€2,400€1,800€1,200
Net annual gain€840€7,272€19,920
Upfront investment€10,000€8,000€6,000
Effect on EBITDA margin+0.03 pp+0.24 pp+0.65 pp
Increase in current EBITDA0.3%2.3%6.3%
First-year ROI−92%−9%232%
Three-year ROI−75%173%896%
PaybackOver 60 months13.2 months3.6 months

The three scenarios combine the low, central and high values of each assumption, in the order shown. The value of hours assumes the time freed is redeployed to useful work; if it is not, the cash effect is smaller. First-year ROI includes the full investment; three-year ROI is undiscounted. Replace the assumptions with your company's figures.

Manufacturer, 30 people: stock-out and overstock alerts from the ERP

Assumptions

  • Estimated average inventory[6],[4]€512,500Vektrum calculation
  • Inventory reduction5.0% / 7.5% / 10.0%Vektrum estimate
  • Annual inventory holding cost20%Vektrum estimate
  • Upfront investment€9,000 / €7,000 / €5,000Vektrum estimate
  • Annual running cost (hosting, licences, support)€1,800 / €1,200 / €900Vektrum estimate
ResultCautiousBaseFavourable
Capital released (once)€25,600€38,400€51,300
Annual holding-cost saving€5,125€7,688€10,250
Annual running cost€1,800€1,200€900
Net annual gain€3,325€6,488€9,350
Upfront investment€9,000€7,000€5,000
Effect on EBITDA margin+0.11 pp+0.21 pp+0.30 pp
Increase in current EBITDA1.0%2.0%2.9%
First-year ROI−63%−7%87%
Three-year ROI11%178%461%
Payback32.5 months12.9 months6.4 months

The starting inventory applies the sector's aggregate ratio (Vektrum calculation on BdP data), dominated by large firms; use your own balance-sheet figure. ROI and payback consider only the annual holding-cost saving; the capital released enters cash once and is not included in ROI.

Profile B: wholesale distributor, 20 people

Distributor, 20 people: email or PDF orders posted to the ERP

Assumptions

  • Customer orders per month400 / 700 / 1,000Vektrum estimate
  • Share received by email or PDF40% / 50% / 60%Vektrum estimate
  • Manual minutes per order (10 to 30 lines)8 / 12 / 15Vektrum estimate
  • Reduction of manual time[34]60% / 70% / 80%Vektrum estimate
  • Loaded cost of one administrative hour[13]15 €/h / 18 €/h / 22 €/hVektrum estimate
  • Upfront investment€12,000 / €10,000 / €8,000Vektrum estimate
  • Annual running cost (hosting, licences, support)€2,400 / €1,800 / €1,200Vektrum estimate
  • Reference EBITDA margin (trade, section G)[6]6.2%Vektrum calculation
ResultCautiousBaseFavourable
Hours freed per year1545881,440
Annual value of the hours€2,304€10,584€31,680
Annual running cost€2,400€1,800€1,200
Net annual gain−€96€8,784€30,480
Upfront investment€12,000€10,000€8,000
Effect on EBITDA margin−0.001 pp+0.12 pp+0.41 pp
Increase in current EBITDA−0.02%1.9%6.7%
First-year ROIDoes not pay back−12%281%
Three-year ROIDoes not pay back164%1,043%
PaybackDoes not pay back13.7 months3.1 months

Excludes the value of fewer order errors and faster customer response, for which there is no reliable public source. The three scenarios combine the low, central and high values of each assumption, in the order shown. The value of hours assumes the time freed is redeployed to useful work; if it is not, the cash effect is smaller. First-year ROI includes the full investment; three-year ROI is undiscounted. Replace the assumptions with your company's figures.

Distributor, 20 people: supplier-invoice reading and accounts payable

Assumptions

  • Supplier invoices per month300 / 450 / 600Vektrum estimate
  • Manual minutes per invoice[34],[38]6 / 8 / 10Vektrum estimate
  • Reduction of manual time[34]60% / 70% / 80%Vektrum estimate
  • Loaded cost of one administrative hour[13]15 €/h / 18 €/h / 22 €/hVektrum estimate
  • Upfront investment€10,000 / €8,000 / €6,000Vektrum estimate
  • Annual running cost (hosting, licences, support)€2,400 / €1,800 / €1,200Vektrum estimate
  • Reference EBITDA margin (trade, section G)[6]6.2%Vektrum calculation
ResultCautiousBaseFavourable
Hours freed per year216504960
Annual value of the hours€3,240€9,072€21,120
Annual running cost€2,400€1,800€1,200
Net annual gain€840€7,272€19,920
Upfront investment€10,000€8,000€6,000
Effect on EBITDA margin+0.01 pp+0.10 pp+0.27 pp
Increase in current EBITDA0.2%1.6%4.4%
First-year ROI−92%−9%232%
Three-year ROI−75%173%896%
PaybackOver 60 months13.2 months3.6 months

The invoice range was estimated for an industrial SME with 20 to 50 people; validate it against your company's purchase SAF-T. The three scenarios combine the low, central and high values of each assumption, in the order shown. The value of hours assumes the time freed is redeployed to useful work; if it is not, the cash effect is smaller. First-year ROI includes the full investment; three-year ROI is undiscounted. Replace the assumptions with your company's figures.

Summary of the base scenarios
ProfileNet annual gainEffect on EBITDA marginIncrease in current EBITDACapital released (once)
Manufacturer, 30 people€13,760+0.45 pp4.3%€38,400
Distributor, 20 people€16,056+0.22 pp3.5%–

Vektrum estimate: sum of each profile's two base scenarios.

07

Experience

What Vektrum has built

Here we show what we have already built, with the status of each piece of work stated. We do not publish client names or project results: the impact figures in this report come from the cited sources and the scenarios in the previous chapter, not from Vektrum cases.

Vektrum case

Delivered

Supplier catalogue sync for a distributor with an online store

A distributor receives its supplier's catalogue and prices as Excel or CSV and has to reflect them in its online store (WooCommerce).

What was built

  • Ingestion of the supplier's Excel/CSV file by stateless services, in the format the supplier already sends
  • Mapping of supplier SKUs to the store's catalogue
  • Store changes executed by an n8n workflow; the sync service never writes to the store directly
  • AI-assisted pricing service that checks competitor prices and suggests a sale price, subject to human approval

What we learned

  • Separating data ingestion from writing to the target system makes each step verifiable and reversible.
  • Prices always keep human approval: the AI suggests, the company decides.
  • Starting from the file the supplier already sends avoids negotiating new formats and speeds up the start.

Vektrum case

Transferable capability

Supplier-invoice reading and accounts payable

A system that reads supplier invoices, prepares posting and organises payment, while keeping control of the process with the finance team.

What was built

  • Reads the AT (Tax Authority) QR code first, with local OCR as fallback; no invoice is sent to external AI services
  • Cost-centre suggestion for each invoice
  • Validation queues for invoices with doubts or discrepancies
  • Payment batches prepared from approved invoices
  • Credit-note handling
  • Supplier IBAN check before payment
  • Audit log of every change

Developed and delivered for a condominium management company, not an industrial company. It is presented as a capability transferable to manufacturing and distribution; it is not a case from this sector.

Vektrum case

Transferable capability

Dashboards and natural-language queries over the company's own database

Connection to the structured data the company already holds (ERP, invoicing, orders) to build management dashboards and allow plain-language questions about that data.

What was built

  • Management dashboards fed directly from the company's database
  • Natural-language queries, with each question translated into a verifiable query
  • Read-only access to source data

Transferable capability; no outcome metrics are presented.

Sector-specific ideas

  • Email and PDF order intake into the ERP: the order is read, customer references are converted into the company's own, and a draft lands in the ERP for validation.
  • Technical quote assistant: searches past quotes and orders for similar jobs and proposes materials, times and a starting price for the estimator to review.
  • Stock-out and overstock alerts: from actual turnover in the ERP, a daily list of at-risk SKUs, without changing systems.
  • OEE reporting from existing data: production and downtime records already captured, consolidated into a weekly report per machine and shift.
  • Order-status self-service: customers check where their order and delivery stand without calling.
  • E-invoicing readiness: inventory of who sends and receives PDFs, confirmation of the timeline with the invoicing-software vendor and the accountant, and automatic reading of whatever keeps arriving as PDF.

08

Priorities

Action plan

The matrix ranks 11 actions by impact, effort and horizon. The rule is to start with what can be measured in weeks using systems the company already has, and only then move to projects that depend on clean historical data.

Priority matrix

Impact

EffortLow → High

Sort by

0 to 30 days

  1. 1

    Measure the baseline

    Count and time invoices, email orders, quotes and status requests for two weeks.

    Impact 2/3 · Effort 1/3 · 0 to 30 days

  2. 2

    Prepare for e-invoicing

    Under the law in force, PDF is accepted until 31/12/2026; confirm with the software vendor and the accountant what changes, without relying on the 2027 budget bill before it becomes law.

    Impact 2/3 · Effort 1/3 · 0 to 30 days

  3. 3

    Invoice on the day of shipment

    Generate the invoice from the delivery note and send automatic reminders before the due date.

    Impact 2/3 · Effort 1/3 · 0 to 30 days

30 to 90 days

  1. 4

    Supplier-invoice reading

    Read, match to the purchase order and prepare posting and payment, with human validation.

    Impact 3/3 · Effort 2/3 · 30 to 90 days

  2. 5

    Email orders into the ERP

    Automatic draft order in the ERP from email or PDF, for validation.

    Impact 3/3 · Effort 2/3 · 30 to 90 days

  3. 6

    Stock alerts

    Daily list of SKUs at risk of stock-out or overstock, from the ERP.

    Impact 3/3 · Effort 2/3 · 30 to 90 days

  4. 7

    Price-list synchronisation

    For distributors: load supplier lists in the format they arrive in, with human approval of prices.

    Impact 2/3 · Effort 2/3 · 30 to 90 days

  5. 8

    Order-status self-service

    Automatic reply or portal showing order and delivery status.

    Impact 2/3 · Effort 2/3 · 30 to 90 days

90 to 180 days

  1. 9

    Technical quote assistant

    Depends on an organised history of quotes and orders; start with one product family.

    Impact 3/3 · Effort 3/3 · 90 to 180 days

  2. 10

    OEE reporting

    Consolidate existing production and downtime records before investing in sensors.

    Impact 2/3 · Effort 2/3 · 90 to 180 days

  3. 11

    Margin dashboard and natural-language queries

    Margin by customer, product and order, from the ERP and invoicing.

    Impact 2/3 · Effort 2/3 · 90 to 180 days

09

Method

Assumptions and limitations

Every figure in this report comes from a research dossier verified on 16/11/2026, in which each number was checked against its primary source. Numbers that could not be confirmed were excluded. Ratios marked as Vektrum calculations are arithmetic on cited official series; Vektrum estimates are our own assumptions, stated and identified in each scenario.

  • Series break. INE moved to CAE Rev. 4, the Portuguese activity classification, in its 2025 results [1]. INE's "Industry and energy" includes electricity and gas; for manufacturing alone we used BdP and Eurostat.
  • BdP ratios. These are sector aggregates weighted by size and dominated by large firms [6]. Trade data refer to section G (wholesale, retail and motor trade); wholesale alone is available only from Eurostat.
  • International benchmarks. Accounts-payable benchmarks [34],[35] come mostly from larger North American companies; OEE benchmarks [33] come from a software vendor; downtime figures [32] refer to the world's 500 largest companies; inventory figures [36] refer to retail. None describes the Portuguese SME.
  • Energy. Prices exclude VAT and recoverable taxes but include non-recoverable charges, which were negative for Portuguese electricity in 2022–2023 [11].
  • E-invoicing. At publication, the 2027 budget is a bill, with the final overall vote scheduled for 24/11/2026 [20]. This report will be revised once the law is published.
  • Assumptions without a public source. Invoice and order volumes, minutes per task, stock reduction, inventory holding cost, investment and running cost are Vektrum estimates. They were chosen prudently and should be replaced with your company's data.

Sources

Every figure in the report points to a numbered source. Vektrum calculations and estimates are labelled as such.

  1. [1]Enterprises in Portugal, Provisional Results 2025 (30/10/2026)INE · 2026 · Official source
  2. [2]Enterprises in Portugal, Final Results 2024 (11/12/2025)INE · 2025 · Official source
  3. [3]Industrial Production Statistics 2024 (15/12/2025)INE · 2025 · Official source
  4. [4]International Day of Micro, Small and Medium Enterprises (27/06/2025)INE · 2025 · Official source
  5. [5]Survey on ICT Usage in Enterprises 2025 (21/11/2025)INE · 2025 · Official source
  6. [6]BPstat, annual indicators and items of non-financial corporations by sector (Central Balance Sheet Database)Banco de Portugal · 2026 · Official source
  7. [7]BPstat, series 12587225: quarterly average collection period in industryBanco de Portugal · 2026 · Official source
  8. [8]Structural business statistics, sbs_ovw_actEurostat · 2026 · Official source
  9. [9]Structural business statistics by size class, sbs_sc_ovwEurostat · 2026 · Official source
  10. [10]Annual industrial production, sts_inpr_aEurostat · 2026 · Official source
  11. [11]Electricity prices for non-household consumers, nrg_pc_205Eurostat · 2026 · Official source
  12. [12]Gas prices for non-household consumers, nrg_pc_203Eurostat · 2026 · Official source
  13. [13]Hourly labour costs, lc_lci_levEurostat · 2026 · Official source
  14. [14]Use of artificial intelligence in enterprises by size class, isoc_eb_aiEurostat · 2026 · Official source
  15. [15]Use of artificial intelligence in enterprises by activity, isoc_eb_ain2Eurostat · 2026 · Official source
  16. [16]Use of ERP software, isoc_eb_iipEurostat · 2026 · Official source
  17. [17]E-invoicing and supply-chain information sharing, isoc_eb_icsEurostat · 2024 · Official source
  18. [18]National minimum wage for 2026 set at €920 (Decree-Law 139/2025)APCMC · 2025 · Official source
  19. [19]Law 73-A/2025 of 30 December (2026 State Budget), article 95Diário da República · 2025 · Official source
  20. [20]2027 State Budget: Bill 110/XVII/2, voting calendarAssembleia da República · 2026 · Official source
  21. [21]Government again postpones the obligation to submit the accounting SAF-T file (08/10/2026)ECO · 2026 · Press
  22. [22]Exports grew in 2025; sales to the US fell 13.4% (09/02/2026)ECO · 2026 · Press
  23. [23]Metal industry sets a new export record (10/02/2026)ECO · 2026 · Press
  24. [24]Portuguese footwear exports rise 0.8% to €1,718m in 2025 (11/02/2026)Observador · 2026 · Press
  25. [25]Textile and clothing exports fall 0.8% to €5,499m in 2025Jornal Económico · 2026 · Press
  26. [26]Barometer, mould industry dataCEFAMOL · 2023 · Study
  27. [27]Agri-food sector aims to exceed €5,000m GVA by 2030 – FIPA (14/06/2026)Agroportal · 2026 · Press
  28. [28]IAPMEI claims success for the 51 mobilising agendas but admits problemsTSF · 2026 · Press
  29. [29]AIP survey finds the labour market is the main obstacle to competitiveness (23/02/2026)ECO · 2026 · Press
  30. [30]Industrial production falls 2.7% in SeptemberDinheiro Vivo · 2026 · Press
  31. [31]Consumer confidence falls for the first time in four months; economic climate also declines (29/10/2026)ECO · 2026 · Press
  32. [32]The True Cost of Downtime 2024Siemens / Senseye · 2024 · Vendor
  33. [33]OEE Benchmark Report 2026 (OEE software vendor)TEEPTRAK · 2026 · Vendor
  34. [34]State of ePayables (Part Nine): AP Benchmarks and Best-in-Class Performance (01/2026)Ardent Partners · 2026 · Study
  35. [35]4 KPIs set good accounts payable organizations apart (22/05/2018)APQC · 2018 · Study
  36. [36]Inventory Distortion infographic (10/2025)IHL Group / Board · 2025 · Study
  37. [37]Global Driver Shortage Report 2025 (preview)IRU · 2026 · Study
  38. [38]Compensation costs for private industry workers averaged $46.89 per hour worked in June 2026U.S. Bureau of Labor Statistics · 2026 · Official source

Methodology

Sources: INE, Banco de Portugal (Central Balance Sheet Database, via BPstat) and Eurostat for Portuguese and EU data; the Official Journal and Parliament for the legal framework [19],[20]; sector and international studies identified as such, with their scope stated. All sources were accessed on 16/11/2026.

Verification: every number in the underlying dossier was checked against its primary source; derived ratios were recalculated from the original series. Values that could not be confirmed were excluded from the report.

Vektrum calculations: ratios (item ÷ sales), inventory days (inventories ÷ cost of goods sold and materials consumed × 365) and Portugal/EU comparisons are arithmetic on published series, with the formula stated.

Scenarios: hours freed = monthly volume × 12 × minutes ÷ 60 × reduction; value = hours × hourly cost; margin effect = (value − running cost) ÷ turnover; stock = inventory × reduction × holding cost. All results are calculated in code from the published assumptions; assumptions without a source are marked as Vektrum estimates.

Limitations: sector aggregates are dominated by large firms; international benchmarks do not describe the Portuguese SME; the legal position on e-invoicing may change after 24/11/2026.

Other reports

We like meeting new companies, understanding how they work and finding where AI can make a difference.

Talk to us