Intro: ERP selection is a "process decision" not a "software decision"#
Treating ERP as software is a mistake. ERP is the company's operational backbone: accounting, stock, sales, procurement, manufacturing, HR — all unified. "SAP or Logo" is a surface question; the real one is "do I bend my processes to the software, or the software to my processes?".
In this post we examine ERP selection under 8 headings: 3 main options (SaaS, on-prem, custom), TCO, process-fit, regulation, scale, team capability, ROI, exit strategy.
2026 reference bands: small business (SMB, 10-50 users) Logo/Mikro stock $7-35K setup + $1.7-7K/year. Mid-market (50-500 users) Microsoft Dynamics / SAP Business One $35-170K + $10-35K/year. Enterprise (500+ users) SAP S/4HANA / Oracle Fusion $170K-1.7M + $35-340K/year. Custom ERP $35-500K setup + $7-70K/year maintenance.
1. 3 main options: SaaS, on-prem, custom#
SaaS ERP (NetSuite, Microsoft Dynamics 365, SAP Business ByDesign, Odoo Cloud): per-user monthly subscription ($50-300/user/month), fast deploy (3-6 months), limited customization. Cloud + maintenance with vendor.
On-premise ERP (SAP S/4HANA, Oracle, Logo Tiger 3 Enterprise, Mikro Jump): on your server + license (per user $170-1.7K one-off + 20-25% annual maintenance). 6-18 months implementation. Customization possible but expensive.
Custom ERP (Next.js + PostgreSQL + custom backend or Odoo + heavy customization): tailored to company; 9-18 months build. Full process fit; no license but maintenance team needed.
Hybrid: off-the-shelf ERP core + custom modules (CRM, manufacturing planning, reporting). Most common approach; "complete custom" is rarely economical.
2. Which question signals custom ERP?#
1. Process difference: Does the off-the-shelf standard flow change 30%+? Is your "X first, then Y, then Z" flow different from sector standard?
2. Competitive advantage: Are your processes the company's competitive moat? (e.g. specialized manufacturing flow, supplier scoring algorithm, dynamic pricing). Yes → standard ERP "flattens your moat".
3. Integration complexity: Will 10+ systems integrate? (e-commerce + marketplace + WMS + 3PL + banking + Salesforce). Off-the-shelf ERP integration limits strain at this level.
4. Data volume + scale: 100K+ transactions/day? Most off-the-shelf ERPs have limited performance optimization.
2+ "yes" → custom ERP economical. 1 "yes" → hybrid. No "yes" → off-the-shelf is enough.
3. TCO (Total Cost of Ownership): real 5-year cost#
SaaS (mid-market): Year 1: $7K (implementation) + $20K (license 100 users × $500/month). Year 2-5: $20K/year × 4 = $80K. 5-year TCO ~$110K.
On-premise (mid-market): Year 1: $27K (implementation) + $35K (license 100 users × $340) + $7K (servers). Year 2-5: $9K/year maintenance × 4 = $36K. 5-year TCO ~$110K.
Custom ERP (mid-market): Year 1: $100K (build). Year 2-5: $14K/year maintenance × 4 = $56K. 5-year TCO ~$160K — looks more expensive but no license dependency.
Hidden lines (every option): user training ($1.7-7K), data migration ($3.5-17K), integration ($3.5-10K per system), business process re-engineering consulting ($7-27K). These add 1.5-2x of the original quote.
4. Process-fit: the "software vs process" fight#
Off-the-shelf ERP philosophy: "adapt your processes to the software" (best practice). Pro: standard process + easy maintenance. Con: lose company-specific efficiencies.
Custom ERP philosophy: "adapt the software to your processes". Pro: full fit + efficiencies preserved. Con: bad processes get encoded too; "chaos automation" trap.
Right approach: first business process re-engineering — redesign processes from scratch. Then choose ERP. "Doing the same bad process faster" creates no value.
Process-Software Fit Score: compute "software fit" score per process (1-10). Total <70 → custom ERP. 70-90 → hybrid. >90 → off-the-shelf ERP.
5. Regulation + local compliance: Turkey-specific lines#
Turkey standard integrations: e-invoice (GİB), e-archive, e-waybill, e-ledger, e-receipt. Built into off-the-shelf ERPs (Logo, Mikro, SAP TR); custom ERP $3.5-10K extra integration.
Tax regulation tracking: VAT, special consumption tax, banking insurance tax rates change yearly. New regulation → ERP update mandatory. Off-the-shelf = vendor's responsibility; custom = your team's.
KEP + KEP-integrated transactions: official correspondence, e-notification. KEP integration available in off-the-shelf ERPs; extra build in custom.
SGK + payroll: Turkey HR payroll regulation (minimum wage, family allowance, severance, notice) is highly dynamic. Off-the-shelf HR modules mandatory; custom builds wrestle with regulation patches every 6 months.
6. Module-based incremental release: "big bang" vs step-by-step#
Big bang ERP migration (all modules at once): 12-18 month duration, 50-70% failure rate (Standish/IDC data). Reasons: too many variables at once + insufficient team training + unclear requirements.
Incremental release: 1-2 core modules first (accounting + stock), 3 months later sales + CRM, 6 months later procurement + manufacturing, 9 months later HR + HR self-service. Each module separate test + go-live.
Advantages: early value (first module in production after 8 weeks), risk distribution, step-by-step team training, rollback possible.
Recommended sequence: finance (accounting + budget) → supply chain (stock + procurement) → sales (CRM + orders) → manufacturing → HR → BI/reporting. This sequence rests on "money flow" + "product flow" foundation.
7. Team + change management: the "human side"#
Implementation team: ERP consultant (1 senior + 2-3 junior), business analyst (1-2), key user (1 person per department, 50% on project), ERP manager (founder/CIO).
Training: 16-40 hours per module per user. Online video + classroom + simulation + handbook. Avg $100-250/user.
Resistance management: the strongest resistance to ERP comes from "middle managers" (loss of control fear). Communication plan + early adopter formation + executive sponsorship critical.
Adoption metric: active user ratio should reach 80%+ within 6 months. Less → "system exists but nobody uses it" scenario (40% of ERP projects in Turkey live in this half-sunken state).
8. ROI + exit strategy#
ROI metrics: stock turnover increase (10-30%), procurement time reduction (20-50%), invoicing time reduction (30-70%), reporting prep time reduction (50-90%).
Typical payback: off-the-shelf ERP 18-30 months. Custom ERP 24-36 months. Hybrid 18-24 months.
Vendor lock-in: with off-the-shelf ERP, data export formats + integration APIs are vendor-bound. With custom, full ownership is yours.
Exit strategy: ERP replacement plan in 5-10 years should be designed from day one. Data migration + integration renewal + team retraining adds $70-700K over 6-18 months. "ERP isn't immortal" — average lifespan is 7-10 years.
Conclusion: not "which ERP" but "what's my company's maturity"#
ERP selection is not just "SAP or Logo"; it's analyzing company maturity + process clarity + team capacity + scale + budget + 5-year vision together.
Healthy approach: first business process re-engineering (3 months) → option evaluation (6 weeks) → pilot project (3 months) → incremental release (12-18 months) → continuous improvement (permanent).
For ERP selection + implementation + customization strategy, reach out via our web software page; we'll prepare a hybrid ERP plan tailored to your sector + scale.
Related services
Related articles
Other articles that support the same decision
Decision Guide
Corporate Website vs E-commerce 2026: Detailed Decision Guide
Sales model, volume, content, architecture, SEO, cost, hybrid + sector scenarios. 8-heading data-driven decision matrix.
Decision Guide
iOS or Android First in 2026: Budget + Market + Strategy Decision
Which platform first? Market share + user value + budget + monetization + review speed + native need. 8-heading data-driven guide.
AI Tools
What is Claude AI? A Comprehensive 2026 Developer Guide
Comprehensive guide to Claude AI's 2026 capabilities — Claude Code, MCP, API integration, KVKK compliance, and enterprise case studies for software developers.
Next step
If you are planning a similar project, we can clarify the scope and shape the right proposal flow together.
Start a project request