ERP TCO Calculator — The True 5-Year Cost

Total cost of ownership (TCO) for an ERP project is almost always larger than the headline licence or subscription figure that vendors quote in early conversations. The licence is one of six cost buckets; the others — implementation services, customisation, training, infrastructure, ongoing operations — usually dwarf it over a five-year horizon. Mid-market buyers that focus on licence price are routinely surprised by year two when the full bill arrives.

This guide walks through a complete TCO model for ERP, with realistic ranges for the German Mid-Market at 50–500 employees, the hidden costs that get under-counted in vendor proposals, and three case scenarios with explicit numbers. The point is not to scare buyers away from ERP — the business case is usually positive when done right — but to make the cost picture honest at the moment of decision rather than at year three when the surprises land.

The TCO components — an overview

A complete TCO model for ERP covers six cost buckets over the chosen horizon (usually five years, sometimes seven). Each behaves differently across deployment models — cloud SaaS, private cloud, on-premises perpetual — and each has its own hidden traps.

1. Licence or subscription: the headline software cost. For on-premises perpetual licences, one-time payment plus annual maintenance (typically 18–22 % of licence value). For cloud subscriptions, per-user-per-month or per-user-per-year recurring fees.
2. Implementation services: the partner's consulting effort to configure, customise, integrate, test and roll out the system. Usually the largest single cost bucket in year one and often the largest cumulative cost over five years for cloud deployments.
3. Customisation and extensions: development beyond standard configuration — bespoke code, partner-provided industry verticals, AppSource or marketplace add-ons. Sometimes part of implementation, sometimes a separate line.
4. Training: key-user training during the project, end-user training before go-live, ongoing training as the user base turns over. Usually under-budgeted.
5. Infrastructure: on-premises hardware and operating costs, private-cloud hosting fees, or the data-residency premiums for cloud SaaS. Lower for SaaS, higher for on-premises and private cloud.
6. Ongoing operations: internal team time, partner support contracts, post-go-live customisation, change-management programmes, integration upkeep. Most buyers underestimate this by a factor of two.

A useful TCO model has these six categories as rows and years 1–5 as columns, with a total row at the bottom. The interesting numbers are not the cells — they are the column totals and the relative weight of each row, which differ dramatically between deployment models.

Licences and subscriptions — calculating them correctly

The licence figure is the easiest to get wrong because vendor pricing models are designed to make comparison hard. The common pitfalls:

• Named-user vs concurrent-user licensing: some vendors price per named user (every person who ever logs in counts), others per concurrent user (peak simultaneous users). Apples-to-apples comparison requires modelling both for the buyer's actual usage pattern.
• Tier-based licensing: “full users”, “limited users”, “read-only users” with different price points. Real ERPs at Mid-Market size usually need a mix; the vendor's default proposal often assumes too many full users.
• Annual escalation clauses: subscription pricing typically escalates 3–7 % per year. Over five years this adds 16–40 % to the cumulative subscription bill. Negotiate caps at contract time, not afterwards.
• Module-by-module licensing: the headline price is often the “core” module; finance, sales, procurement and so on are priced separately. Build the realistic full-suite licence in the comparison, not the marketing teaser.
• AppSource and marketplace add-ons: the Microsoft Business Central ecosystem in particular relies on AppSource extensions for German-specific functionality (DATEV, GoBD-tools, advanced production planning). Budget €5–20 per user per month for the typical add-on stack.

Realistic licence ranges for the DACH mid-market at 50–500 employees:

• Cloud SaaS: €60–180 per user per month for mainstream products (Business Central, NetSuite, Sage Intacct, myfactory, weclapp). Premium products (SAP S/4HANA Cloud) at €200–400.
• On-premises perpetual: €2,500–8,000 per user one-time, plus 18–22 % annual maintenance. Over five years: €5,000–15,000 per user.

Implementation costs — the biggest variable

Implementation services are the largest single variable in any ERP TCO — and the line most likely to overrun. The mid-market market in DACH has a fairly consistent pattern: implementation cost in year one runs at 1.5x to 4x the year-one licence or subscription cost, depending on complexity.

The implementation cost includes:

• Project management: the partner's project manager, business analyst, technical lead.
• Configuration: setting up the system to match the buyer's standard processes — chart of accounts, document numbering, approval workflows, master-data structures.
• Customisation development: bespoke code, reports, extensions where standard does not fit.
• Integration build: interfaces to e-commerce, banking, payroll, CRM, EDI, BI. Almost always more expensive than the original scoping assumed.
• Data migration services: the partner's technical work on the ETL, reconciliation and dress-rehearsal migrations. The business-side data cleansing is the buyer's effort, not the partner's.
• Testing support: partner attendance during integration test and UAT, defect triage and resolution.
• Cutover services: the go-live weekend, final data migration, hypercare staffing.

Realistic ranges for implementation services at first go-live:

• Simple cloud SaaS, services business, 50–100 employees: €60,000–150,000.
• Standard cloud SaaS, distribution, 50–200 employees: €120,000–300,000.
• Manufacturing with production planning, 100–300 employees: €250,000–800,000.
• Industry-specific vertical, complex production, 200–500 employees: €500,000–1,500,000.
• SAP S/4HANA mid-market, 300–500 employees: €800,000–2,500,000.

A useful sanity check: if the partner's implementation quote is below 1.5x the year-one licence cost, it is almost certainly under-scoped. The pattern of cheap-quote-then-change-orders is a real risk; structured selection with fixed-price or capped time-and-materials commitments reduces it. See our ERP vendor negotiation guide.

Hosting, infrastructure and deployment

Infrastructure cost varies dramatically by deployment model. The pattern most mid-market buyers under-appreciate: cloud SaaS bundles infrastructure into the subscription, so it looks free; on-premises and private cloud separate it out, so it looks expensive — but the underlying total is often similar.

Cloud SaaS

Infrastructure is included in the subscription. The exceptions: data-residency premiums (for Germany-resident data centres), enhanced storage tiers, sandbox environments beyond the standard inclusion, and dedicated environments where multi-tenant is not acceptable. Typical infrastructure premium: 0–15 % of base subscription.

Private cloud (single-tenant hosted)

Hosting fees, paid to the partner or hosting provider, typically €25–80 per user per month depending on workload. Includes infrastructure operations, patching, backup, monitoring. For a 100-user installation: €30,000–100,000 per year.

On-premises

Server hardware (usually 3–5 year refresh cycle), database licences (SQL Server, Oracle), backup infrastructure, disaster-recovery site, network upgrades. Initial capex for a 100–200 user installation: €80,000–250,000 over the depreciation cycle, plus €30,000–80,000 per year in operating costs (electricity, hosting space, maintenance contracts).

The pattern that surprises buyers: over five years, the cloud SaaS infrastructure cost (zero or near-zero separate line) and the on-premises infrastructure cost (capex plus opex) often converge once depreciation is amortised. The cloud TCO advantage in early years can flip in later years for stable user counts.

Internal effort — the hidden cost bucket

The most consistently under-counted cost in ERP TCO is internal team time. Mid-market buyers routinely build TCO models with zero internal effort line, on the assumption that “the people are paid anyway”. That is technically true but operationally wrong — the time these people spend on the ERP project is time they are not spending on their day jobs, which means backfill, opportunity cost or accepted productivity loss.

The internal effort categories that should appear in the TCO:

• Project team: 1–2 full-time equivalents (project manager, business analyst) for the duration of the project (usually 12–18 months at first implementation). At fully-loaded mid-market salaries of €80,000–120,000 per FTE per year, this is €100,000–240,000 over the project.
• Key users: 30–60 % time commitment for 6–12 months from typically 4–10 key users. Aggregate effort: 2–6 FTE-equivalents over the project — another €150,000–700,000 in fully-loaded costs.
• Executive sponsor and steering committee time: CFO/COO at 5–10 % over 18 months, steering committee at 1–3 hours per fortnight. Aggregate: €50,000–150,000.
• End-user training time: for a 100-user deployment, 4–8 hours per user equals 400–800 hours of productive time consumed. At a fully-loaded blended rate of €50–80 per hour: €20,000–65,000.
• Year-1 hypercare staffing: internal support, additional key-user time during stabilisation, helpdesk volume increase. Typically €30,000–100,000.
• Ongoing operations from year 2 onwards: at least 0.5–1.5 FTE dedicated to ERP operations, master data, integration and continuous improvement. €40,000–180,000 per year.

Cumulative internal effort cost over five years for a 100–200 user mid-market deployment: €500,000–1,500,000. This often matches or exceeds the partner implementation bill and changes the TCO comparison meaningfully.

Worked example: mid-market with 80 users

To make the TCO model concrete, here are three realistic scenarios with explicit five-year numbers. All figures are indicative, in euros, for the DACH market.

Scenario A: 50-user manufacturing mid-market, cloud SaaS (Business Central + production add-on)

• Subscription: 50 users × €105/month average = €63,000/year. Over 5 years with 5 % escalation: €348,000.
• AppSource add-ons (DATEV, ZUGFeRD, production): €15/user/month = €9,000/year. Over 5 years: €50,000.
• Implementation services year 1: €220,000.
• Integration build (Magento e-commerce, MES, banking): €80,000.
• Training: €30,000 in year 1, €5,000/year thereafter.
• Internal team time over 5 years: €600,000.
• Partner support and continuous improvement: €40,000/year from year 2.
• 5-year TCO total: roughly €1.55 million — about €310,000 per year average, or €6,200 per user per year all-in.

Scenario B: 150-user professional services, cloud SaaS (NetSuite)

• Subscription: 150 users × €95/month average = €171,000/year. Over 5 years with escalation: €945,000.
• Implementation services year 1: €380,000.
• Integration build (Salesforce, HubSpot, banking, time tracking): €120,000.
• Training: €50,000 in year 1, €10,000/year thereafter.
• Internal team time over 5 years: €900,000.
• Partner support and continuous improvement: €75,000/year from year 2.
• 5-year TCO total: roughly €2.85 million — about €570,000 per year average, or €3,800 per user per year all-in.

Scenario C: 300-user multi-country group, private-cloud SAP S/4HANA RISE

• Subscription and RISE base fees: roughly €900,000/year. Over 5 years: €5,000,000.
• Implementation services year 1–2: €1,800,000 across both years.
• Customisation and country localisation: €600,000.
• Training: €180,000 in year 1, €40,000/year thereafter.
• Internal team time over 5 years: €2,200,000.
• Partner support and continuous improvement: €300,000/year from year 3.
• 5-year TCO total: roughly €11.2 million — about €2.2 million per year average, or €7,500 per user per year all-in.

The per-user-per-year cost varies more by industry and complexity than by deployment model. Services businesses with standard processes hit the lower end; manufacturing and multi-country groups sit higher.

Frequently forgotten TCO items

Even careful TCO models miss items that surface as “unexpected” costs in years two and three. The most common omissions:

• Year-1 hypercare staffing premium: the partner's rate during hypercare is often a multiple of the implementation rate. Budget €30,000–100,000 explicitly.
• Integration upkeep: every interface needs maintenance as connected systems evolve. Budget 10–20 % of original integration build cost per year as ongoing upkeep.
• Continuous customisation: business processes evolve, new requirements emerge, regulations change. Budget €30,000–100,000 per year for ongoing customisation from year 2.
• Re-training as staff turn over: at 10–15 % annual turnover, a quarter of the user base needs training over five years. Budget €5,000–15,000 per year.
• Sandbox and non-production environments: additional dev, test, training environments often charged separately by cloud vendors. Budget 15–30 % of production subscription.
• Audit and compliance work: GoBD self-audit, GDPR data-processing register updates, SOX equivalents for listed parents. €10,000–40,000 per year.
• BI and reporting layer: the ERP's native reporting is rarely enough. Power BI or Tableau licences, BI consultant time. €30,000–100,000 per year.
• Change-management programme: not just initial training but ongoing communications, user engagement, key-user community. €30,000–80,000 per year.
• Major version upgrades (on-premises): every 3–5 years, a meaningful re-implementation effort. Budget €100,000–500,000 per upgrade cycle.

A TCO model that includes these items lands 30–60 % higher than one that excludes them. The honest model is the one that survives contact with reality at year three.

Industry- and size-dependent cost drivers

Five recurring patterns shape TCO by industry and size:

Manufacturing complexity premium. Discrete and engineer-to-order manufacturing add 30–60 % to implementation cost versus distribution or services at the same user count. Production planning, MES integration, shop-floor data collection, PLM connectors, quality management all add scope. The on-going operations premium is smaller (10–20 %) but real.

Multi-country and multi-entity premium. Each additional country adds 4–8 months of roll-out time and 60–100 % of the first-country implementation cost. Each additional legal entity adds chart-of-accounts mapping, intercompany flows and consolidation complexity.

Regulated industry premium. Pharma, medical devices, food production, finance, defence: validated environments, audit trails, GxP compliance, change-control procedures all add 20–50 % to project cost and roughly 15–25 % to ongoing operations.

Size effect (favourable). Per-user implementation cost typically falls as size increases — a 500-user implementation costs less per user than a 50-user one because the fixed costs (project management, basic configuration) amortise over more users. The implementation cost per user halves roughly every doubling of user count.

Heritage system premium. Replacing a heavily customised legacy ERP costs 1.5–2x replacing a clean entry-level system. The customisation has to be understood, decided on (keep, retire, redesign) and either rebuilt or replaced with standard configuration. The implicit cost of legacy.

TCO as a negotiation and steering instrument

A well-built TCO model is more than a budgeting exercise — it is a negotiation and steering instrument throughout the project. The practical applications:

• Vendor negotiation: a TCO model that includes all six cost buckets allows the buyer to negotiate on total cost rather than headline price. Vendors often offer flexibility on subscription discount but resist movement on implementation fixed-price; a TCO view exposes where the real money sits and where negotiation matters most.
• Comparison between vendors: apples-to-apples TCO at the same scope, deployment model and 5-year horizon is the only honest comparison. Headline price comparisons are routinely misleading.
• Business case for the board: the executive sponsor needs the TCO to defend the investment to the board or owner. A model that survives challenge from a sceptical CFO is one with realistic numbers in every bucket, including internal effort.
• Post-go-live cost control: the TCO model from selection becomes the baseline for actual cost tracking. Variances flag where the project is running over (customisation usually), where it is running under (rarely, but training sometimes), and where re-planning is needed.
• Renewal negotiation: at the end of year 3 or year 5, the cloud subscription renewal comes up. A buyer who has tracked actual TCO has leverage in renewal negotiations; one who has not is at the vendor's mercy.

The single most useful artefact from a TCO exercise is not the total number, but the realistic distribution of cost across the six buckets. That distribution exposes which deployment model fits the buyer's situation, which negotiation levers matter and which surprises to budget for. See also our cloud vs on-premises decision matrix for the deployment perspective.

Related Topics

• Cloud vs on-premises decision matrix
• ERP for 50 to 100 employees
• ERP implementation checklist (first 90 days)