Running an Optimization

This guide walks through the process of running an optimization job to generate caseload recommendations.

Before You Start

Check Your Data

Before running an optimization, verify your data is current:

1. Navigate to Data Sources
2. Check the "Last Synced" timestamp
3. If data is stale, trigger a sync or upload fresh CSV data

Running optimization on outdated data produces recommendations that don't reflect your current reality.

Know Your Goal

Consider what you're trying to achieve:

• Reduce travel time - Focus on geographic efficiency
• Balance workloads - Prioritize even distribution
• Minimize disruption - Emphasize keeping existing assignments
• All of the above - Let the algorithm balance these factors

Your goal influences how you configure the optimization.

Starting an Optimization

Step 1: Navigate to Optimize

Click Optimize in the sidebar to access the optimization page.

If you've run optimizations before, you'll see your history here. Click New Optimization to start a new run.

Step 2: Select Your Scope

Choose what to include in the optimization:

Regions Select which regions to optimize. You can:

• Optimize all regions together
• Focus on specific regions
• Exclude regions not ready for changes

Team Caseloads By default, all team caseloads are included. You can:

• Include all team caseloads in selected regions
• Exclude specific team caseloads that shouldn't change
• Focus on a subset of teams

Step 3: Configure Settings

Adjust how the optimization runs:

Optimization Weights

These sliders control what the algorithm prioritizes:

Travel Efficiency (0-100) Higher values = stronger focus on reducing travel time

• Set higher when driving distance is your main concern
• Set lower if travel is less important than other factors

Workload Balance (0-100) Higher values = stronger focus on equal distribution

• Set higher when some employees are overloaded
• Set lower if current distribution is already reasonable

Continuity (0-100) Higher values = stronger preference for keeping existing assignments

• Set higher to minimize disruption
• Set lower to allow more changes for better results

Tip: For your first optimization, try the default settings. Adjust based on results.

Constraints

Set hard limits the optimization must respect:

Maximum Caseload The maximum number of clients, workers, or visits for a geographic team caseload. This constraint applies at the team level, not to individual workers within the team.

Region Boundaries Whether the optimization can reassign postal codes between adjacent team caseloads.

Step 4: Run the Optimization

Click Run Optimization to start.

You'll see a progress indicator while the optimization runs. Typical run times:

• Small organizations (< 50 employees): 1-2 minutes
• Medium organizations (50-200 employees): 2-5 minutes
• Large organizations (> 200 employees): 5-15 minutes

You can leave the page and return later - the optimization continues in the background.

While Optimization Runs

Monitoring Progress

The progress screen shows:

• Current status (analyzing, optimizing, finalizing)
• Time elapsed
• Estimated time remaining (when available)

Canceling

If you need to stop an optimization:

1. Click Cancel on the progress screen
2. Confirm the cancellation
3. No results are saved from a cancelled run

You might cancel if:

• You realize data needs updating
• Settings need adjustment
• An urgent matter requires attention

When Optimization Completes

Notification

When the optimization finishes:

• The page updates to show results
• You may receive an email notification (if enabled)
• The run appears in your optimization history

Next Steps

After completion, proceed to:

1. Review the results - Understand what's recommended
2. Analyze the changes - Identify specific reassignments
3. Decide on implementation - Choose what to apply

See the Reviewing Results guide for detailed instructions.

Troubleshooting

"Not enough data"

The optimization needs minimum data to run:

• At least 2 employees
• At least 5 clients
• Location data for employees and clients

Verify your data is complete and try again.

"Optimization taking too long"

For very large datasets, optimization may take longer. Options:

• Wait - complex problems take time
• Reduce scope - optimize fewer regions at once
• Contact support if runs consistently exceed 30 minutes

"No improvements found"

If results show minimal improvement:

• Your current assignments may already be efficient
• Try adjusting weights to prioritize different factors
• Verify data is accurate and complete

"Results don't make sense"

If recommendations seem wrong:

• Check data quality, especially locations
• Verify employee capacity settings
• Review constraint settings
• Contact support for assistance