On-Call Compensation Models

The Problem with Unpaid On-Call

You ask engineers to carry the pager. They adjust their plans, stay near computers, and maintain availability outside work hours. Then Friday night, production breaks—and they spend three hours fixing it while their friends are at dinner.

The next morning, you ask: “How was your weekend?”

Too many organizations treat on-call as an implicit job requirement, something engineers should accept as part of being professional. But unpaid on-call creates predictable problems: resentment builds, talented engineers leave for competitors with better policies, and response quality degrades as exhausted teams stop caring about alerts.

Fair compensation isn’t generosity—it’s operational necessity. When engineers know their availability has value, they respond faster, stay longer, and maintain higher alerting standards. The question isn’t whether to compensate on-call. It’s which model matches your operational needs and budget constraints.

Why On-Call Deserves Compensation

Being on-call isn’t the same as working, but it’s not the same as being off-duty either. The distinction matters.

Availability has cost. Engineers on-call can’t fully disconnect. They avoid situations where they can’t respond quickly—no movies in theaters, no alcohol, no travel far from reliable internet. Even when alerts don’t fire, the possibility constrains behavior.

Interruptions fragment life. A 2 AM page doesn’t just take the hour spent fixing the issue. It destroys sleep quality, ruins the next morning, and creates stress that lingers for days. One weekend incident can eliminate recovery time that engineers need to sustain long-term performance.

Responsibility creates pressure. Knowing that production depends on your response generates mental load. Some engineers describe constant low-level anxiety during on-call shifts, checking their phone obsessively even when alerts aren’t firing.

The best organizations recognize these costs explicitly and compensate accordingly. Approximately 40% of companies now provide on-call compensation, with larger organizations leading the way. The trend is clear: as engineering labor markets tighten, compensation becomes table stakes for competitive talent.

Common Compensation Models

Different organizations need different approaches based on incident frequency, severity, team size, and budget. Here are the proven models, with trade-offs for each.

Fixed Payment Per Shift

The most common approach: pay a flat amount for each on-call period regardless of whether incidents occur.

How it works: Engineers receive $350–$1,000 per week on-call, paid regardless of alert volume. Some companies pay daily ($50–$150 per day), others monthly for regular rotation participation.

Examples from industry:

• incident.io pays approximately $350 per week
• European companies often pay €800–€1,050 per week
• Large tech companies typically pay $500–$1,000 per weekly shift

When this works: Best for teams with predictable incident rates and stable budgets. The fixed amount simplifies payroll and sets clear expectations. Engineers appreciate knowing compensation in advance.

Trade-offs: Doesn’t scale with actual workload. An engineer who handles six incidents gets the same pay as one who handles zero. This can create perceived unfairness, especially when some shifts consistently generate more work than others.

Hourly On-Call Rate

Pay for every hour of availability at a reduced rate, typically a percentage of base hourly compensation.

How it works: Engineers earn 10–35% of their standard hourly rate for each hour on-call. Google, for example, pays Tier-1 SREs (five-minute response) 66% of base rate for on-call hours, and Tier-2 (30-minute response) 33% of base rate.

Calculation example: An engineer with $150,000 annual salary ($72/hour) who works 25% on-call rate would earn $18/hour for availability. A 24-hour weekend shift pays $432.

When this works: Ideal for organizations that want compensation proportional to availability. Works well when on-call hours vary significantly week to week. Provides fairest distribution when engineers have different shift lengths or frequencies.

Trade-offs: More complex payroll calculations, especially across timezones. Doesn’t directly account for incident response time—an engineer who works five hours during a shift might earn the same availability pay as one who never got paged.

Pay Per Incident Response

Compensate only for time actively spent working on incidents, not for passive availability.

How it works: Engineers receive hourly or flat rates when they actually respond to pages. Typically 1.5x–2x standard hourly rate for incident work outside business hours.

Calculation example: An engineer paged at 2 AM spends 90 minutes resolving an incident. At $72/hour base rate, they earn $108–$144 for that response (1.5x–2x multiplier).

When this works: Budget-conscious organizations with infrequent incidents. Reduces costs when alerts rarely fire. Appeals to engineers who prefer being compensated only for active work rather than passive availability.

Trade-offs: Doesn’t account for the constraint that availability creates. An engineer who goes weeks without pages still adjusted their lifestyle but receives nothing. Can incentivize keeping incidents open longer to maximize compensation. Doesn’t compensate for the sleep disruption and recovery cost of middle-of-night pages.

Salary Inclusion Model

Build on-call expectations and compensation directly into base salary rather than paying separately.

How it works: Job contracts explicitly state on-call requirements (e.g., “one week per month”) and salary reflects that expectation. Some companies add 10–20% to base salary for roles with regular on-call responsibilities.

Example structure: A senior engineer position advertised at $180,000 with explicit expectation of quarterly on-call rotation, compared to $165,000 for similar role without on-call.

When this works: Organizations that want predictable costs and simplified payroll. Works best when on-call frequency is consistent and known during hiring. Transparent approach that sets expectations before employment begins.

Trade-offs: Less flexibility when on-call burden changes. Difficult to adjust when incident rates increase or decrease. Engineers can’t opt out without salary renegotiation. New hires comparing offers may not recognize the on-call premium if not clearly documented.

Hybrid Models

Combine multiple approaches to balance availability compensation with incident workload.

How it works: Fixed weekly stipend ($200–$500) plus hourly pay for actual incident response (1.5x–2x rate), plus compensatory time off after severe incidents.

Example structure:

• $300 base payment per weekly on-call shift
• $100/hour for active incident response outside business hours
• Half-day PTO granted for any Saturday/Sunday page requiring more than one hour work

When this works: Organizations that want fairness across variable incident severity. Recognizes both availability cost and response work. Provides flexibility to adjust as incident patterns change.

Trade-offs: Most complex to administer. Requires careful tracking of incident hours and severity. Engineers need clear documentation about which components apply when.

Additional Compensation Components

Beyond direct payment, other forms of compensation address the full impact of on-call.

Compensatory Time Off

Grant paid time off to recover from on-call shifts, especially after incidents.

Common approaches:

• One PTO day per on-call week
• Half-day PTO for weekend pages
• Full day off after critical incidents requiring overnight work

AWS, for example, provides half-day PTO for every Saturday or night-time page in addition to hourly compensation.

Why it matters: Money doesn’t restore sleep or eliminate exhaustion. Time off recognizes that incident response creates recovery debt that must be repaid.

Incident Severity Multipliers

Adjust compensation based on incident impact and response complexity.

Structure example:

• SEV-1 (critical outage): 2x hourly rate
• SEV-2 (major degradation): 1.5x hourly rate
• SEV-3 (minor issue): 1x hourly rate

Rationale: A five-hour overnight incident that takes down your entire platform deserves higher compensation than a five-minute fix for a single customer’s issue.

Holiday and Weekend Premiums

Recognize that availability during high-value personal time costs more.

Common multipliers:

• Weekend shifts: 1.25x–1.5x standard on-call rate
• Major holidays: 2x–3x standard rate
• New Year’s Eve, Thanksgiving, Christmas: 3x–4x rates

Retention impact: Engineers with families especially value recognition that holiday on-call prevents presence at important personal events.

Benchmarking: What Companies Actually Pay

Compensation varies widely by company size, location, and industry. Here’s what organizations actually pay in 2025:

Large Tech Companies (10,000+ employees):

• Google: 33–66% of base hourly rate for on-call hours
• Amazon/AWS: 25% of base pay per on-call hour + time off
• Microsoft: $500–$800 per weekly shift
• Meta: $600–$1,000 per week + incident response pay

Mid-Market Companies (500–5,000 employees):

• European tech companies: €800–€1,050 per week (Zalando, HelloFresh)
• U.S. SaaS companies: $350–$600 per week
• Financial services: $400–$800 per week + incident bonuses

Startups (under 500 employees):

• Early stage (under 50): Often no compensation or equity-only
• Growth stage (50–200): $200–$400 per week becoming standard
• Late stage (200–500): Approaching mid-market rates

Geographic variations:

• San Francisco Bay Area: 20–30% above national averages
• New York / Boston: 15–25% above national averages
• Seattle / Austin: Aligned with national averages
• Remote-first companies: Typically use national averages regardless of employee location

The data shows clear patterns: larger companies pay more, compensation is rising industry-wide, and organizations without competitive pay struggle to fill rotations.

Building Your Compensation Model

Creating fair on-call pay requires balancing multiple factors: operational needs, budget constraints, team preferences, and market competition.

Assess Your Incident Profile

Start with data about your actual on-call burden:

Incident frequency: How many alerts fire per shift? One per week creates different cost than five per day.

Incident severity: Are most issues five-minute fixes or multi-hour investigations? Severity distribution should influence your model choice.

Time distribution: Do incidents concentrate during business hours or spread across nights and weekends? After-hours incident rates justify higher compensation.

Recovery time: How long does it take engineers to return to normal productivity after incidents? Some teams report full-day impact from severe overnight pages.

Review the past three months of incident data before designing compensation. Your model should match reality, not aspirations.

Calculate True Costs

Understand what on-call actually costs your organization when not compensated:

Turnover cost: When engineers leave due to on-call burnout, replacement costs include recruiting ($20,000–$40,000), onboarding time (3–6 months reduced productivity), and institutional knowledge loss (immeasurable but significant).

Response degradation: Uncompensated on-call leads to slower acknowledgment, lower effort investigations, and reduced alerting quality as teams tune out noise.

Opportunity cost: Engineers who refuse on-call participation reduce your coverage pool, forcing remaining engineers into more frequent rotations.

Compare these hidden costs against direct compensation. Most organizations discover that paying fairly costs far less than the alternatives.

Match Model to Culture

Different compensation approaches align with different organizational values:

Fixed stipends communicate: “We value your availability and constrained lifestyle.”

Hourly rates communicate: “We recognize your time has quantifiable value.”

Incident-only pay communicates: “We pay for actual work, not passive waiting.”

Salary inclusion communicates: “On-call is a core responsibility we factor into compensation.”

Choose the model that matches your culture and the message you want to send about how you value engineer time.

Start Transparent and Iterate

When implementing on-call compensation, transparency prevents misunderstandings:

Document everything: Write clear policies covering how pay is calculated, when it’s distributed, what scenarios qualify, and how engineers can raise questions.

Communicate early: Announce compensation plans before the next rotation starts. Retroactive pay decisions create confusion and perceived unfairness.

Gather feedback: After 2–3 months, survey on-call participants about whether compensation feels fair and what adjustments would improve the model.

Adjust based on data: Track metrics like incident frequency, response times, rotation participation rates, and turnover. Adjust compensation if these indicators decline.

The best organizations treat compensation models as living policies, not static decisions. Review annually and adjust when incident patterns change or market rates shift.

Common Mistakes to Avoid

Several compensation approaches predictably fail. Learn from others’ mistakes:

Paying Nothing

The most common mistake: treating on-call as an implicit job expectation without compensation.

Why this fails: Engineers notice immediately when competitors pay for on-call. The best performers leave first, creating coverage gaps that force remaining engineers into more frequent rotations, accelerating the exodus.

Exception: Very early-stage startups (under 20 employees) where equity significantly exceeds cash compensation and everyone shares operational burden equally. Even then, establish compensation before Series A.

Paying Too Little

Offering token compensation that doesn’t reflect actual burden.

Example: $50 per weekly shift when engineers spend 10+ hours on incidents creates resentment worse than no pay. Engineers perceive this as checking a box rather than genuine recognition.

Minimum viable compensation: At least $200–$300 per weekly shift for typical incident rates, scaled up based on severity and frequency.

Inconsistent Application

Paying some teams or individuals while leaving others uncompensated.

Why this fails: Creates internal equity issues. Backend engineers feel undervalued when they learn frontend engineers receive on-call pay while they don’t, even if incident rates differ.

Solution: If compensation varies by team, document clear criteria that explain differences. Better yet, apply consistent models across the organization.

Ignoring Non-Monetary Costs

Focusing solely on cash while ignoring time off, schedule flexibility, and recovery needs.

Example: Paying well but requiring engineers to work normal hours the day after an overnight incident. The money doesn’t compensate for exhaustion.

Solution: Combine monetary compensation with time-off policies, flexible scheduling, and explicit permission to prioritize recovery.

Complex, Opaque Calculations

Creating compensation formulas so intricate that engineers can’t predict their pay.

Example: “Base rate incident severity time-of-day multiplier response time modifier weekend premium” becomes impossible to calculate mentally.

Solution: Simplicity wins. Engineers should know within 10% what their on-call shift will pay without spreadsheet calculations.

Legal Considerations

On-call compensation intersects with employment law in ways many organizations overlook:

FLSA Requirements (U.S.): Non-exempt employees must be paid for all hours worked. If on-call engineers are “engaged to wait” (unable to use time freely), those hours may require minimum wage payment. Consult employment counsel about your specific circumstances.

Overtime calculations: On-call hours may count toward weekly totals that trigger overtime thresholds. Track carefully for non-exempt employees.

Contractual obligations: If employment contracts specify compensation, on-call pay must align with those terms. Document on-call expectations during hiring.

International variations: European labor laws often mandate higher on-call compensation than U.S. standards. Companies with global teams need location-specific policies.

This isn’t legal advice—consult qualified employment counsel about your jurisdiction’s requirements. But ignorance of legal obligations doesn’t protect you from penalties.

Measuring Compensation Effectiveness

How do you know if your on-call compensation is working? Track these indicators:

Rotation participation rate: Percentage of eligible engineers who accept on-call shifts. Rates below 70% suggest compensation or rotation frequency issues.

Mean time to acknowledge (MTTA): How quickly engineers respond to pages. Increasing MTTA may indicate growing resentment or burnout despite compensation.

Turnover rate: Compare attrition for engineers in on-call rotations versus those not on-call. Significant differences suggest on-call burden outweighs compensation.

Incident quality: Track false positive rates and alert refinement activity. Teams that feel fairly compensated invest effort in improving alerts rather than tolerating noise.

Direct feedback: Anonymous surveys asking “Do you feel fairly compensated for on-call?” Simple but revealing.

Effective compensation doesn’t eliminate on-call burden, but it prevents the burden from creating retention problems and response degradation.

Upstat and On-Call Management

Technology can’t solve compensation questions, but the right tools make fair models easier to implement. Platforms like Upstat support multiple rotation strategies—weekly rotation, sequential assignment, fair distribution algorithms—with automatic handling of holidays, user exclusions, and timezone complexity.

When you need to track who was on-call when for compensation calculations, automated rotation management provides clear audit trails. When engineers need flexibility to swap shifts or override assignments, tools that handle these changes without manual coordination reduce administrative overhead.

Fair compensation works best when paired with sustainable rotation design, clear escalation paths, and tools that respect engineers’ time outside on-call shifts.

Conclusion

On-call compensation isn’t overhead—it’s operational investment. Engineers who feel fairly compensated respond faster, stay longer, and maintain higher standards. The organizations that recognize availability as costly and compensate accordingly build better retention, better response, and better reliability.

Start with your incident data. Choose a model that matches your operational reality. Document it clearly. Implement it consistently. Measure its effectiveness. Adjust as needed.

The engineers carrying your pager deserve more than gratitude. They deserve compensation that reflects the real cost of availability. Get this right, and your on-call rotation becomes sustainable. Get it wrong, and your best engineers will find competitors who value their time appropriately.

Your on-call rotation should protect systems and respect people. Fair compensation is how you do both.