- 4
- April
What is SLA? — Service Level Agreements Before Signing an ERP Contract
SLA (Service Level Agreement) is a document that defines the service standards between an ERP provider and the client organization. It covers everything from system availability (Uptime), Response Time, Resolution Time, to penalties when targets are not met. This article explains every dimension of SLA that executives must understand before signing an ERP contract.
In Brief: An SLA is a "sub-agreement" within an ERP contract that defines what service level the vendor must deliver and what happens if they fail to meet it.
Organizations without a clear SLA have no leverage when the system goes down.
Why is SLA Critical for ERP Projects?
An ERP system is the heart of business operations. If the system goes down for even one hour, it could mean lost revenue, delays in financial closing, or procurement disruptions. An SLA helps organizations:
- Reduce risk — set minimum standards the vendor must meet, rather than leaving it to their discretion
- Establish accountability — the vendor has specific metrics to maintain, with penalties for non-compliance
- Enable business planning — know exactly how much system availability to expect, allowing clear risk management planning
- Compare vendors — use SLA terms as objective criteria for evaluating proposals from multiple vendors
- Protect budget — establish Penalty/Credit mechanisms when service falls below standards
5 Key SLA Metrics to Specify in an ERP Contract
| # | Metric | Description | Standard Value |
|---|---|---|---|
| 1 | Uptime / Availability | Percentage of time the system is available (excluding Planned Maintenance) | 99.5% - 99.99% |
| 2 | Response Time | Time from issue reported to vendor acknowledgment | 15 min - 4 hours |
| 3 | Resolution Time | Time from issue reported to complete resolution | 1 hour - 5 business days |
| 4 | Data Backup & Recovery (RPO/RTO) | RPO = maximum hours of data loss allowed, RTO = maximum hours to restore the system | RPO ≤ 1 hr, RTO ≤ 4 hrs |
| 5 | Patch & Security Update | Timeframe for applying security patches | Critical: 24 hrs, Normal: 7 days |
Uptime Comparison: A 0.4% Difference Makes a Huge Impact
| Uptime | Downtime per Month | Downtime per Year | Suitable For |
|---|---|---|---|
| 99.5% | ~3.6 hours | ~43.8 hours | Standard organizations, Mon-Fri operations |
| 99.9% | ~43 minutes | ~8.7 hours | Organizations requiring high continuity |
| 99.99% | ~4.3 minutes | ~52 minutes | Mission-critical systems, financial institutions |
Read more about Disaster Recovery for ERP systems to understand system recovery approaches during emergencies.
SLA by Severity Level (Priority Tiers)
A well-structured SLA must classify issue severity into tiers, because not every problem requires the same resolution timeframe:
| Priority | Description | Example | Response Time | Resolution Time |
|---|---|---|---|---|
| P1 - Critical | Entire system is down, widespread business impact | Complete ERP system failure, database corruption | ≤ 15 min | ≤ 4 hours |
| P2 - High | Core function is unavailable, but other modules still work | Accounting module down, unable to issue invoices | ≤ 30 min | ≤ 8 hours |
| P3 - Medium | Secondary function has issues, workaround available | Report showing incorrect data, unable to print documents | ≤ 2 hours | ≤ 2 business days |
| P4 - Low | Minor issue, does not affect core operations | Display text error, UI cosmetic issue, enhancement request | ≤ 4 hours | ≤ 5 business days |
Penalty & Escalation: Enforcement and Escalation Provisions
An SLA without penalties is like a law without enforcement — it has no real effect. Key provisions to include in the contract:
- Service Credit: When uptime falls below the target. For example, Uptime < 99.5% = 10% credit on that month's service fee, < 99.0% = 25% credit, < 98.0% = 50% credit
- Daily penalty: When Resolution Time exceeds the target. For example, 0.1% of contract value per day of delay (capped at 10%)
- Contract termination rights: If SLA targets are missed for 3 consecutive months, the client may terminate the contract without penalty
- Escalation Path: Clearly define who the issue escalates to if not resolved in time (e.g., P1 exceeding 2 hours escalates to VP of Engineering)
Example Escalation Matrix
| Time Elapsed | Escalate To |
| 0 - 1 hour | Support Team L1 (Help Desk) |
| 1 - 2 hours | Support Team L2 (Senior Engineer) |
| 2 - 4 hours | Support Manager + Notify client PM |
| > 4 hours | Senior executives from both sides |
SLA for Government Agencies: What the TOR Must Specify
For government agencies procuring ERP systems per TOR specification guidelines, the SLA should be clearly defined in the TOR document, particularly:
- Minimum Uptime — specify a clear percentage (e.g., no less than 99.5% during business hours)
- Resolution timeframe — broken down by Priority Level as shown in the table above
- Warranty period — per the Government Procurement and Supplies Management Act B.E. 2560, no less than 1 year
- Data backup — must specify RPO/RTO and number of backup copies (at least 2 copies at separate locations)
- SLA reporting — vendor must submit monthly service performance reports against SLA targets
- Penalties — specify penalty rates per Ministry of Finance procurement regulations
10-Point Checklist for SLA Before Signing an ERP Contract
- Uptime Guarantee — specify minimum uptime percentage with measurement method and monitoring tools
- Priority Classification — define P1-P4 priority levels with clear definitions
- Response & Resolution Time — specify acknowledgment time and resolution time per Priority level
- Service Hours — define service hours (8x5, 12x6, 24x7) and after-hours on-call conditions
- Backup & Recovery (RPO/RTO) — specify backup frequency, recovery timeframe, and DR testing procedures
- Patch & Security Update — define Critical Patch update timeframe and regular update schedule
- Penalty & Service Credit — specify penalties or credits when SLA targets are not met
- Escalation Procedure — define the Escalation Matrix with names and contact channels
- Monthly SLA Report — vendor must submit monthly SLA reports with Uptime, Ticket, and Resolution Time statistics
- Review & Renewal — define SLA review cycles (e.g., every 12 months) and conditions for amendments
Common SLA Pitfalls
| Pitfall | Problem Caused | Prevention |
|---|---|---|
| Vague language | "Will resolve promptly" has no numbers and can be interpreted differently | Specify in hours/business days explicitly |
| No penalties | Vendor has no incentive to maintain standards | Define Service Credit / penalties in writing |
| No measurement method | Disputes over how 99.5% Uptime is measured, whether Planned Maintenance is counted | Specify monitoring tools and calculation formula |
| No priority separation | Minor and major issues given the same resolution time, resources stretched thin | Classify P1-P4 with clear conditions |
| No exclusions defined | Vendor claims Force Majeure every time | Define acceptable exclusions and their clear scope |
How to Track and Monitor SLA Performance
An SLA is only valuable if performance is tracked consistently. Recommended approaches:
- Use Monitoring Tools — install Uptime Monitoring systems (e.g., Zabbix, Nagios, UptimeRobot) for automatic uptime measurement
- Ticketing System — use a ticketing system (e.g., Helpdesk, ITSM) to record issue report time, acknowledgment time, and resolution time
- Monthly SLA Reports — require the vendor to submit monthly reports comparing actual performance against SLA targets
- Quarterly Review Meetings — hold SLA Review meetings every 3 months to discuss issues and improvements
- Executive Dashboard — summarize SLA data into a dashboard that executives can view at a glance
Having a solid Disaster Recovery plan helps organizations confidently achieve the RPO/RTO targets defined in their SLA.
Summary
An SLA is not just an appendix to a contract — it is the most important risk management tool in an ERP project. Executives must prioritize SLA from the vendor selection stage: define clear metrics, classify priorities into tiers, enforce real penalties, and track performance consistently. Organizations with a strong SLA can be confident that their ERP system will perform to the expected standard.
"A good SLA is not a document you pull out when there is a problem — it is a standard that both parties uphold together every day."
Related Articles from Knowledge Center
- ERP ROI — How to Measure Return on Investment Clearly Executive
- ERP Project Preparation Checklist Implementation
- ERP Security Basics Every User Must Know End User

