SAP Total Cost of Ownership: What to Include and How SAP Inflates Its TCO Models

What Belongs in a Genuine SAP TCO Model: The Eight Essential Cost Categories

A defensible TCO model has eight cost categories that must be included in equal measure for both on-premise and cloud scenarios. If SAP's calculator omits or minimizes any of these for the cloud scenario, it's biased.

1. License Fees (Perpetual or Subscription)

• On-premise: Annual SAP software maintenance (typically 22% of licence value) + licence amortization if you calculate cap-ex spread over 5 years
• Cloud (RISE with SAP): Monthly or annual RISE subscription per user/instance, documented in Order Form
• Why SAP hides this: They include vague "hyperscaler infrastructure" bundled into RISE that's actually separately charged; they don't break down what's included

2. Infrastructure (Hardware, Colocation, Cloud Compute)

• On-premise: Server hardware (servers, storage, networking), refresh cycle (5–7 years), colocation/data center costs (power, cooling, space), backup systems
• Cloud (RISE): Hyperscaler compute (Azure, AWS hourly rates scaled to your instance size), storage (per GB/month), data egress (often €0.10–0.25 per GB, a shocking hidden cost), network bandwidth
• Why SAP hides this: They bundle hyperscaler costs into a flat RISE fee, then "surprise" you in year 2 with egress charges and overprovisioning adjustments

3. Staffing (Basis, Finance, Data Management)

• On-premise SAP team: Basis engineers (patching, upgrades, 24/7 support), Finance team (GL, AP/AR, consolidation), Data warehouse/BI team, security/compliance, capacity planning
• Cloud (RISE) team: Reduced Basis work (SAP handles patching, 90% of operational tasks), but added roles: cloud optimization, FinOps, egress management, cloud security
• Staffing shift: On-premise teams may shrink 20–30%, but cloud requires different skillsets that often cost more to hire. Net savings are rarely >15%

4. Support and Maintenance

• On-premise: SAP Enterprise Support (22% annually of licence value), optional third-party support (10–15% alternative), incident response, patching windows
• Cloud (RISE): Support is bundled; SAP sets availability SLAs and response times. No alternative vendors. You're locked in
• Real cost: RISE support is not cheaper; it's just included in the subscription

5. Upgrades and Patching

• On-premise: Major version upgrades (S/4HANA migration, ECC → S/4HANA) every 5–10 years; quarterly or monthly patch cycles. Costs: testing, downtime, consulting
• Cloud (RISE): SAP forces quarterly upgrades and mandatory patching (you don't choose). Hidden cost: regression testing, change management, business disruption from SAP-mandated changes
• Why cloud is more expensive: You lose the ability to defer patches and plan upgrade cycles around your business; you react to SAP's schedule

6. Training and Change Management

• On-premise: Periodic training as systems are upgraded or configured changes occur
• Cloud (RISE): Quarterly SAP-mandated UI/functionality changes, continuous adaptation. Training costs are front-loaded in years 1–2, then ongoing every quarter
• Real cost: Business users spend 50–100 hours/year on cloud-mandated training vs 10–20 hours on-premise

7. Migration Costs (One-Time)

• On-premise to cloud: Data migration, cutover testing, parallel run period, consulting. Typically €1–5M depending on complexity. Occurs once
• On-premise refresh (5-year cycle): Infrastructure refresh costs, some software optimization. Much lower than migration
• Why SAP's TCO hides this: They bury migration costs in "Year 1" or spread them over 5 years, making cloud appear cheaper than it is in the critical first 2 years

8. Contingency and Overrun Risk

• On-premise: Well-understood risks; enterprises have 10–15 years of operational cost history
• Cloud (RISE): Uncertain risks: egress charges surprise you, hyperscaler price increases, SAP compliance changes, lock-in costs if you want to exit. Budget 15–25% contingency
• Why SAP omits this: They assume perfect execution; they don't budget for failure modes

How SAP's TCO Calculator Systematically Undercounts On-Premise Costs

SAP's TCO models start with a bias: they assume on-premise systems are run poorly and inefficiently. They overstate on-premise staffing, hardware, and support costs to make the baseline look bad.

Hardware Inflation

SAP's models often assume on-premise customers run on legacy hardware, aging data centers, and inefficient infrastructure. In reality:

• Most large enterprises refresh servers every 5–7 years, not 3 years
• Colocation and data center costs are often lower than SAP estimates (enterprise rates: €1–3 per kWh vs SAP's assumption of €4–5)
• SAP factors in 100% redundancy and HA costs; many on-premise deployments use tiered HA (cold standby for dev, hot standby for production)

Real on-premise hardware cost is often 30–40% lower than SAP's models show.

Staffing Inflation (On-Premise)

SAP's TCO models often assume:

• 4–6 full-time Basis engineers managing SAP (vs real average: 2–3)
• Separate infrastructure teams, database administrators, security staff (in reality, many companies share these across multiple systems)
• 24/7 support coverage (expensive; most on-premise shops run business hours + on-call)

By inflating on-premise headcount, SAP makes the cost savings from "cloud automation" look larger. But if you're running lean on-premise (as most are), cloud headcount savings evaporate.

Support Cost Inflation

SAP shows on-premise support at 22% of licence cost (standard SAP Enterprise Support). But many enterprises negotiate support contracts at 18–20%, use third-party support (€80–120K/year instead of €300K+), or use a hybrid approach. SAP's models don't account for these negotiated rates.

The "Productivity Gains" Sleight of Hand

This is where SAP's TCO becomes outright fictional. SAP claims cloud delivers:

• "Reduced time to value": Implies users are more productive, but provides no quantification. Reality: users spend first 12 months learning new cloud UI and business processes
• "Faster reporting and analytics": Assumes cloud query performance is better than on-premise. Truth: cloud performance is equivalent for most queries; egress costs make large-scale analytics more expensive
• "Business agility": Implies the ability to configure systems faster, but SAP controls all changes; you're less agile, not more

Real Example: What SAP's Calculator Gets Wrong

Fortune 500 financial services firm, current on-premise state:

• 2 full-time Basis engineers (not 5 as SAP assumes) = €320K/year
• SAP Enterprise Support at negotiated 20% (not SAP's standard 22%) = €880K/year (not €1.1M)
• Hardware refresh on 6-year cycle, €2M every 6 years = €333K/year (SAP assumes 5-year cycle and higher cost)

Real on-premise cost: €1.5M/year. SAP's calculator shows: €2.1M/year. By understating the on-premise baseline 30%, cloud looks comparatively cheaper.

Cloud Benefits and Hidden Costs: RISE with SAP TCO Assumptions That Don't Hold

SAP's cloud pitch promises cost savings, operational simplicity, and continuous innovation. But most benefits either don't materialize or come with hidden costs.

The "Operational Efficiency" Promise

SAP claims RISE eliminates:

• Patching burden (SAP handles it)
• Hardware management (hyperscaler owns it)
• 24/7 operations (SAP guarantees uptime)

Reality:

• Patching burden shifts, not disappears. You still regression-test every patch; you just don't control the schedule
• Hardware management is still your problem via FinOps. You manage cloud spend, right-sizing, and resource optimization—a job most enterprises underestimate by 50%
• RISE SLAs are 99.5% (not 99.99% for premium on-premise deployments). You trade guaranteed uptime for shared hyperscaler responsibility

Hyperscaler Infrastructure Costs and Egress Charges

This is where RISE gets expensive fast. SAP bundles hyperscaler compute into a flat monthly fee, but additional costs emerge:

• Data egress: SAP charges ~€0.12–0.25 per GB for data moving out of the RISE environment to your on-premise systems or other cloud providers. A large analytics workload (100 GB/month of data export) costs €1,200–3,000/month or €14K–36K/year. SAP doesn't mention this upfront
• Compute overprovisioning: SAP sizes your instance conservatively (to be safe), then charges if you need more. Year 2–3 overprovisioning costs can add 20–30% to your monthly fee
• Storage overage: Each additional TB beyond your baseline costs €100–200/month. Accumulate 20 TB of additional data in years 2–3, and you're paying €2,400–4,800/month extra

The "Continuous Innovation" Sleight of Hand

SAP claims cloud gives you access to continuous innovation (quarterly feature releases, new modules). But:

• Forced adoption: You can't opt out of SAP's quarterly releases. Many new features break existing custom code, require regression testing, or introduce UI changes that confuse users
• Customization costs: Every upgrade cycle, custom enhancements require testing and often rework. Budget 100–200 hours/quarter for testing and remediation
• Real cost: "Continuous innovation" is really "continuous change management cost" masked as a benefit

Egress Lock-In

The most insidious cloud cost: egress charges create a lock-in barrier. If you want to leave RISE in year 3 because costs are too high, you face:

• €500K–5M to migrate data back to on-premise or another cloud provider
• Months of downtime and testing
• Potential egress charges for the final data export

This lock-in is intentional. SAP knows egress costs will make leaving too expensive to justify.

RISE with SAP TCO: What SAP Hides in "Hyperscaler Infrastructure"

RISE with SAP is SAP's cloud subscription for S/4HANA. Understanding the real cost structure is critical, because SAP deliberately obscures it.

What RISE Claims to Include

• S/4HANA software licence
• SAP Cloud Platform (BTP) services
• Hyperscaler infrastructure (Azure, AWS, or GCP)
• 24/7 SAP support
• Quarterly updates and patches

What RISE Actually Costs (Reality)

Realistic RISE Year 1 TCO (Large Enterprise)

• Base RISE subscription: €300K/year
• Data egress (estimated 50 GB/month): €36K/year
• Compute overage (Year 1 typically low): €12K/year
• Professional services (optimization, governance): €30K/year
• Internal FinOps and cloud optimization staff: €200K/year (1 FTE)
• Total Year 1: €578K

Realistic RISE Year 3 TCO (Same Enterprise)

• Base RISE subscription (2% annual increase): €324K/year
• Data egress (growth to 150 GB/month): €54K/year
• Compute overage (expanded analytics workload): €60K/year
• Storage expansion (added 40 TB): €48K/year
• Professional services (SAP FinOps consulting): €50K/year
• Internal FinOps staff (1.5 FTE due to complexity): €300K/year
• Total Year 3: €836K/year

Most enterprises' RISE costs increase 30–40% from Year 1 to Year 3 due to data growth, compute expansion, and egress charges.

Support Cost Modelling: SAP Enterprise Support vs Third-Party vs RISE

Support costs are a hidden variable in any SAP TCO model. Understanding your options is critical.

On-Premise Support Options

• SAP Enterprise Support (Standard): 22% of licence value annually. Includes patches, updates, and SAP support. Average cost: €300K–500K/year for large enterprise
• SAP Premier Support: 32% of licence value. Higher priority, dedicated support team. Cost: €450K–750K/year. Rarely justified unless you have mission-critical SLAs
• Third-party Maintenance (e.g., Rimini Street, TCS, Gexperts): 10–15% of SAP licence cost. Includes patches and support. Cost: €150K–250K/year. Savings vs SAP: €150K–300K/year. Risk: SAP blacklists you for audits and won't work with you on licensing disputes

Cloud Support (RISE)

• Support is bundled; you don't choose. SAP provides 24/7 support included in the RISE fee
• No alternative vendors. You cannot use third-party support with RISE
• SLA guarantees: 99.5% availability. Compare to on-premise guarantees (often 99.99% with active-active HA)
• Hidden cost: Cloud support is more expensive over time because you lose the ability to negotiate pricing or switch vendors

TCO Implication

On-premise support is typically negotiable and can be optimized through third-party vendors. Cloud support is fixed and locked in. For a €2M on-premise deployment, support cost differences can be €100K–300K/year. When aggregated over 5 years, this is €500K–1.5M—material enough to swing a TCO decision.

Migration Cost: Why SAP's Migration Estimates Are Consistently 40-60% Below Actual

Migrating to RISE with SAP is expensive and disruptive. SAP's estimates almost always undercount the true cost.

SAP's Typical Migration Estimate

• Data migration: €200K–500K
• Cutover testing: €100K–200K
• Consulting support: €300K–800K
• Total: €600K–1.5M

What SAP Misses (Real Costs)

• Data cleanup: Most enterprises discover decades of data quality issues during cloud migration. Fixing master data, GL reconciliations, and custom tables can consume €200K–800K and 3–6 months of effort
• Custom code remediation: Average 60–80% of on-premise custom enhancements don't work in cloud. Rewriting, testing, and validating custom code adds €300K–1.5M
• Change management and training: Users are disrupted for 6–12 months during cloud transition. Training, communication, and support costs: €100K–500K. Lost productivity: €500K–2M
• Parallel run period: Most enterprises maintain on-premise systems for 3–6 months parallel to cloud to reduce risk. This adds infrastructure costs, staffing costs, and testing overhead: €200K–500K
• Testing and validation: Comprehensive end-to-end testing (often overlooked by SAP) requires 500–1,500 hours: €150K–500K
• Contingency for overruns: Most cloud migrations encounter go-live delays (1–3 months). Extending the migration timeline adds €100K–300K in consulting and staffing costs

Realistic Total Migration Cost

• SAP's estimate: €600K–1.5M
• Real cost for complex enterprise: €2M–4M
• Adjustment: +150–200% vs SAP's estimate

Where Costs Hide

Most of the hidden migration costs fall into two buckets:

• Data and customization work (40%): Cleaning data, rewriting custom code, decommissioning legacy systems
• Internal labour and change management (40%): Your staff effort in testing, training, cutover support
• Contingency and risk (20%): Delays, rework, failures

SAP's estimates focus narrowly on external consulting (the 20% contingency piece) and miss the 80% that's your labour.

The 5-Year Horizon Problem: Why SAP Uses 3-Year Models and You Need 5-Year Data

SAP's standard TCO models look at 3 years. Your board should be looking at 5 years. This timing mismatch hides true cloud economics.

Why SAP Loves 3-Year Models

• Cloud migration happens in Year 1 (costs upfront)
• Years 2–3 show cloud cost advantages (lower run-rate costs)
• Year 4+ is ignored (cloud costs compound)
• A 3-year model makes cloud look cheaper because you're not seeing egress and overprovisioning costs accelerate

What 5-Year Models Reveal

• Year 1: High migration costs (€1.5–4M), RISE subscription (€300–500K), on-premise decommissioning
• Year 2: RISE cost stabilizes; cloud teams mature; egress charges appear (€30K–50K/month)
• Year 3: Data growth compounds; egress charges increase 50%; compute overprovisioning costs rise. RISE now costs €600K–800K/year all-in
• Year 4–5: Continuous growth in cloud costs; egress becomes material (€100K–200K/year). Cumulative 5-year cost is 20–40% higher than Year 3 projection

Real Example: 5-Year Financial Model

Manufacturing enterprise, on-premise S/4HANA with €2M/year software and infrastructure costs, considering RISE migration:

In this scenario, SAP's 3-year model would show cloud as slightly more expensive (€9.7M vs €6.0M cumulative on-premise), but the 5-year view reveals cloud costs escalate significantly.

Building Your Own Defensible TCO Model: The Framework

Here's the structure you need to build a TCO model that withstands board scrutiny and SAP pushback:

Cost Category 1: Licence Fees

• On-premise: Annual maintenance (22% of licence value) × 5 years
• Cloud (RISE): Monthly subscription × 60 months, plus escalators (typically 2–3%/year)

Cost Category 2: Infrastructure

• On-premise: Server hardware (€300K–800K initial, €150K refresh every 5–7 years), storage (€50K–200K), networking, colocation (€50K–150K/year)
• Cloud (RISE): Hyperscaler compute bundled into RISE + egress (€0.12–0.25 per GB) + storage overages (€100–200 per TB)

Cost Category 3: Staffing

• On-premise: 2–4 Basis engineers (€250K–500K/year), 1 Finance team member (€100K), 1 BI/data person (€100K) = €450K–700K/year
• Cloud: 1–2 cloud architects (€200K/year) + 1 FinOps specialist (€120K/year) + reduced Basis burden (save €150K/year) = €170K–320K net new cost/year

Cost Category 4: Support

• On-premise: 22% of licence annually, or €200K–400K via third-party
• Cloud: Bundled in RISE subscription

Cost Category 5: Upgrades and Patching

• On-premise: €100K–300K per major upgrade (once every 5–10 years), quarterly patch testing (€20K–50K/year)
• Cloud: Continuous patching (included), but testing and change management (€50K–100K/year)

Cost Category 6: Training and Change Management

• On-premise: €30K–50K/year
• Cloud: €80K–150K/year (front-loaded in Years 1–2 due to UI changes and cloud adoption)

Cost Category 7: Migration (One-Time)

• On-premise to cloud: €2M–4M (Year 1 only)
• On-premise refresh: €500K–1.5M every 5–7 years

Cost Category 8: Contingency

• On-premise: 10% of total costs
• Cloud: 20% of total costs (higher risk due to lock-in and unknowns)

Your TCO Model Checklist

Before presenting any TCO analysis to your board, verify:

• All 8 cost categories are included for both on-premise and cloud scenarios
• Assumptions are documented and sourced (not SAP's estimates)
• 5-year horizon is modelled (not 3-year SAP standard)
• Migration costs include data cleanup, custom code, and testing (not just external consulting)
• Egress and storage overages are budgeted conservatively (€500K–2M over 5 years)
• Staffing changes are realistic (cloud requires different skillets, not just fewer people)
• Contingency buffer is 20% for cloud, 10% for on-premise
• Sensitivity analysis shows outcomes if assumptions change by ±20% (cloud costs, data growth, staffing availability)

This checklist ensures your model withstands scrutiny and SAP won't be able to argue you've undersold cloud benefits.