Mohamed Bhimji

Reading Time: 5 minutes

Why Your Contact Center Keeps Missing SLAs (And How Forecasting, Occupancy & Shrinkage Hold the Answer)

You’ve hired great people. You’ve invested in the latest CRM. Your team works hard. Yet somehow, you’re still missing SLAs, burning out agents, or overspending on labor costs.

The culprit? Often, it’s not effort or talent — it’s workforce management fundamentals.

After 20+ years leading 24/7 contact center operations across compliance-sensitive environments, I’ve learned that three metrics separate reactive support teams from proactive, high-performing ones: Forecasting, Occupancy, and Shrinkage.

Master these, and you’ll stop firefighting. Ignore them, and you’ll keep cycling through the same problems.

Let’s break them down — with real examples and actionable fixes.

1. Forecasting: Stop Guessing, Start Predicting

   The Problem

   You schedule 10 agents on Monday because “that’s usually enough.” But suddenly, volume spikes 40% and your SLAs tank. Or worse — you overstaff a quiet Tuesday and waste thousands in labor costs.

   What Forecasting Really Means

   Forecasting isn’t crystal-ball gazing. It’s using historical data + business intelligence to predict:

   • Contact volume (calls, chats, tickets) by hour/day/week
   • Average handle time (AHT) trends
   • Seasonal patterns (e.g., post-weekend surges, product launch spikes)
   • Impact of external events (marketing campaigns, outages, regulatory changes)

   Example

   Support volume spiked every Monday morning. Digging deeper, we discovered our client onboarding team batch-processed new restaurant accounts over the weekend. New users hit the system Monday AM → immediate confusion → ticket flood.

   The Fix:

   • Partner with Product/Marketing to get their calendar (launches, promotions, known issues)
   • Analyze 12+ months of historical volume by hour-of-day and day-of-week
   • Build simple models: “If we launch Feature X, expect 3x volume for 72 hours”
   • Review forecasts weekly — adjust for anomalies

   Quick Win: Start tracking volume by 30-minute intervals for two weeks. You’ll spot patterns you never knew existed.

2. Occupancy: The Balancing Act Between Efficiency and Burnout

   The Problem

   Your agents are busy — really busy. Occupancy is at 95%! But CSAT is dropping, attrition is rising, and there’s zero time for coaching. What gives?

   What Occupancy Really Means

   Occupancy = % of logged-in time agents spend actively helping customers (talk time + after-call work) vs. idle but available.

   • <70% occupancy = Underutilized staff, wasted budget
   • 75–85% occupancy = Sweet spot for sustainable performance
   • 90% occupancy = Burnout risk, no coaching time, quality erosion

   Example

   Your team has a 92% occupancy. SLAs looked great on paper. But CSAT has fallen from 88% to 76% in six months, and you are losing your best agents to competitors. Why? They had zero time to breathe, learn, or improve. Every minute was consumed by back-to-back contacts.

   The Fix:

   • Intentionally build coaching time into schedules (even 30 mins/day/agent)
   • Protect “focus blocks” for QA reviews, knowledge base updates, and skill development
   • Monitor occupancy by agent, not just team average — some may be at 98% while others are at 65%
   • Accept that slightly lower occupancy = higher quality + retention

   Quick Win: Audit your team’s occupancy distribution. If anyone is consistently >90%, redistribute workload or hire before you lose them.

3. Shrinkage: The Invisible Thief of Capacity

   The Problem

   You scheduled 20 agents. But only 14 are taking calls. Where did the rest go? Meetings, breaks, sick calls, training — welcome to shrinkage.

   What Shrinkage Really Means
   Shrinkage = % of scheduled time agents are unavailable to handle contacts.

   Two Types:

   Planned Shrinkage: Vacations, holidays; Training, coaching; Team meetings, 1:1s; Breaks, lunches
   Unplanned Shrinkage: Sick calls, no-shows; System outages; Emergency absences; Unscheduled personal time

   Typical shrinkage: 25–35% for healthy teams.

   Example

   You scheduled based on “pure volume need” — ignoring shrinkage. Result? Chronic understaffing, missed SLAs, and frustrated agents covering for absent colleagues. Once you start planning for 30% shrinkage (15% planned + 15% unplanned), you are finally staffed realistically and hit our targets consistently.

   The Fix:

   • Track shrinkage by category — know your baseline (e.g., “We average 5% sick rate, 10% meeting time, 15% breaks/lunch”)
   • Reduce unplanned shrinkage through engagement, flexibility, and wellness support
   • Protect planned shrinkage — don’t let “urgent” work cannibalize coaching/training time
   • Schedule above raw volume need: If you need 100 hours of handle time at 80% occupancy and 30% shrinkage, schedule ~179 hours, not 100

   Quick Win: Calculate your actual shrinkage rate for last month. If it’s >35%, investigate root causes.

   If it’s <20%, you’re probably under-investing in development.

How These Three Work Together

These aren’t isolated metrics — they’re a system:

Step 1: Forecast Volume
→ “We expect 500 hours of handle time next week”

Step 2: Adjust for Target Occupancy
→ At 80% occupancy: 500 ÷ 0.8 = 625 logged-in hours needed

Step 3: Adjust for Expected Shrinkage
→ At 30% shrinkage: 625 ÷ 0.7 = 893 scheduled hours

Step 4: Build the Schedule
→ ~22 FTEs to deliver 500 hours of actual support

Get one wrong, and the whole system breaks.

Your Action Plan

This Week:

• Pull 90 days of volume data by hour/day
• Calculate current occupancy and shrinkage rates
• Identify one pattern you can act on (e.g., “Mondays need 2 extra agents”)

This Month:

• Meet with Product/Marketing to get their 90-day calendar
• Build a simple forecasting spreadsheet (or upgrade your WFM tool settings)
• Audit your team’s occupancy distribution — who’s burning out?

This Quarter:

• Reduce unplanned shrinkage by 20% through engagement initiatives
• Protect 2–3 hours/week/agent for coaching and development
• Review forecast accuracy monthly — adjust your model

The Bottom Line

Forecasting, occupancy, and shrinkage aren’t just WFM jargon. They’re the operational backbone of every high-performing support team.

Master them, and you’ll:

✅ Hit SLAs consistently without burning out your team
✅ Reduce labor costs while improving quality
✅ Build a sustainable, scalable operation
✅ Earn credibility with Finance and Executive leadership

Ignore them, and you’ll stay stuck in reactive mode — no matter how hard your team works.

The good news? You don’t need expensive tools or a PhD in data science. You need discipline, curiosity, and a willingness to plan ahead.

Start small. Track one metric. Have one conversation with Product. Adjust one schedule.

Your future self — and your team — will thank you.

BONUS: How AI Can Supercharge WFM (Without Replacing the Human Element)

1. AI-Powered Forecasting

What it does: Machine learning models ingest years of historical volume, handle time, weather, marketing calendars, product release dates, and even social sentiment to generate predictive, self-correcting forecasts. Your move:

• Stop relying on “last year + 10%” spreadsheets. Modern AI tools (like Calabrio WFM, NICE IEX, or even integrated CRM AI) auto-adjust for anomalies and seasonality.
• Quick win: Use AI to run “what-if” scenarios before big changes: “If we change our onboarding flow, how will ticket volume shift over 30 days?”

2. AI for Occupancy & Intelligent Routing

What it does: Real-time AI analyzes agent skill sets, current workload, sentiment, and even burnout signals to route contacts to the best available human, not just the next available agent. Your move:

• Deploy AI-assisted routing that matches complex issues to senior agents and simple queries to newer team members—automatically balancing occupancy across skill tiers.
• Use AI real-time assist (like agent copilot tools) to auto-suggest responses, pull KB articles, and draft wrap-up notes. This cuts after-call work by 20–40%, directly boosting productive occupancy.

3. AI for Shrinkage Reduction & QA

What it does: AI automates the “invisible” work that eats into agent capacity: scheduling conflicts, manual QA scoring, routine training, and tier-1 deflection. Your move:

• Tier-1 deflection: Deploy AI chatbots or interactive voice response that resolves password resets, billing checks, and order status without human touch. Every deflection = one less scheduled hour.
• Automated QA scoring: AI evaluates 100% of interactions for compliance, empathy, and resolution accuracy—freeing supervisors from manual scoring so they can focus on 1:1 coaching.
• Predictive absence modeling: AI flags burnout or attrition risk before agents call in sick, letting you intervene proactively and protect planned capacity.

⚠️ The Leadership Trap to Avoid

AI is a multiplier, not a magic wand. The teams that fail with AI:

• Deploy it without clean data or clear SOPs
• Expect instant 50% deflection rates
• Forget to train agents on how to use AI alongside customers

The winning playbook:

1. Start small (one channel, one use case)
2. Measure baseline → implement → measure again
3. Keep humans in the loop for empathy, escalation, and continuous tuning
4. Celebrate wins publicly so your team sees AI as a tool, not a threat

About The Author