Platform Analytics

Pandora provides comprehensive analytics for platform administrators and organization admins to monitor marketplace health, worker performance, and financial outcomes. This page covers the key metrics available and how to use them effectively.

Worker Tier Distribution

Understanding how workers are distributed across tiers helps assess marketplace health:

MetricWhat It ShowsHealthy Range
Tier 1 (New) countWorkers still in onboarding/early phase20-40% of total
Tier 2 (Proven) countWorkers with basic reliability established20-30%
Tier 3 (Reliable) countSolid mid-tier workers15-25%
Tier 4 (Expert) countHigh-performing workers10-15%
Tier 5 (Elite) countTop performers3-8%

A healthy distribution looks like a pyramid — many workers at the bottom, fewer at the top. If too many workers are clustering at Tier 1 with no advancement, consider adjusting the minimum job requirements or scoring weights.

Job Posting Metrics

Track how the marketplace is being used:

MetricDescription
Total jobs postedCumulative Pandora job postings
Active postingsCurrently open and accepting claims
Average time to first claimHow quickly posted jobs attract workers
Average time to fill all slotsHow long until all slots are claimed
Fill ratePercentage of postings that get fully claimed
Cancellation ratePercentage of postings cancelled before any claim
Escalation ratePercentage of internal-first postings that escalate to external

Time to Claim Analysis

Time to first claim is one of the most important marketplace health indicators:

Average TimeAssessmentAction
Under 1 hourExcellent — high worker demandConsider expanding job volume
1-4 hoursGood — healthy marketplaceMonitor, no action needed
4-24 hoursFair — adequate but room for improvementReview payout values, worker base
Over 24 hoursConcerning — low worker engagementInvestigate: too few workers? Payouts too low?

Worker Performance Metrics

Aggregate performance data across the worker base:

MetricDescriptionTarget
Average completion rateCompleted jobs / accepted jobs across all workersAbove 90%
Average on-time rateOn-time completions / total completionsAbove 85%
No-show rateNo-shows / total accepted jobsBelow 3%
Put-back ratePut-backs / total accepted jobsBelow 10%
Average satisfaction scoreMean org rating across all completionsAbove 4.0/5.0
Average response timeMedian time from job visible to claimUnder 2 hours

Score Event Analysis

The PandoraScoreEvent table provides granular data on score changes:

Event TypeWhat It Tracks
job_completeScore change from on-time completion
job_lateScore change from late completion
no_showScore change from no-show (always negative)
putbackontimeScore change from on-time put-back
putbacklateScore change from late put-back
rating_receivedScore change from a new org rating
manual_overrideAdmin-initiated tier change
scheduled_recalcDaily scheduled recalculation

Analyzing the distribution of event types helps identify systemic issues. A spike in no-shows, for example, might indicate scheduling problems or payout dissatisfaction.

Financial Metrics

For Organizations

MetricDescription
Total Pandora spendSum of all worker payouts
Average payout per jobMean worker payout across completed jobs
Blended payout rateAverage tier percentage across all claims
Tier savingsDifference between Pandora value and actual payouts (savings from sub-100% tiers)
Margin per jobJob value minus worker payout

For the Platform

MetricDescription
Total marketplace volumeSum of all Pandora values across postings
Gross merchandise valueSum of all job values (client payments)
Worker earningsTotal payouts to workers
Active worker countWorkers who completed at least one job in the last 30 days
Revenue per active workerAverage earnings per active worker

Visibility Pipeline Analytics

Track how the visibility pipeline affects job distribution:

MetricDescription
Jobs filtered by tierHow many jobs each tier level cannot see
Jobs filtered by lead timeJobs hidden due to lead time restrictions
Jobs filtered by value capJobs hidden due to value caps
Jobs filtered by supply tagsJobs hidden due to supply mismatches
Average visible jobs per workerHow many jobs each worker typically sees

This helps identify if the visibility configuration is too restrictive (workers see too few jobs) or too loose (workers are overwhelmed with irrelevant listings).

Monitoring Dashboards

Real-Time Dashboard

A live view showing:

  • Active job postings count
  • Pending claims in progress
  • Recent completions (last 24 hours)
  • Active worker count (currently browsing or on jobs)
  • Escalation timer countdowns for internal-first postings

Weekly Summary

An automated report including:

  • New workers onboarded
  • Tier promotions and demotions
  • Total jobs completed
  • Total payouts issued
  • Top performing workers (by score delta)
  • Flagged issues (high no-show rate, escalation failures)

Using Analytics Effectively

Identifying Issues

SignalPossible IssueInvestigation
Dropping fill ratePayouts too low or worker shortageCompare payout values to market rates; check active worker count
Rising no-show rateWorkers overcommittingCheck average claims per worker; consider adding schedule conflict warnings
Slow time to claimWorker base not large enough or jobs not attractiveReview tier distribution and lead time configuration
Clustering at Tier 1Advancement too difficultReview minimum job counts and scoring weights
Many escalationsInternal workers not claimingCheck if internal workers are active; consider adjusting escalation timing

Making Data-Driven Adjustments

  1. Review analytics weekly — Look for trends, not single data points
  2. Compare before and after — When you change tier configuration or scoring weights, compare metrics from the week before and after
  3. Segment by tier — Performance metrics are most useful when broken down by tier level
  4. Monitor seasonality — Job volume and worker availability may vary by season
  5. Track cohorts — Follow groups of workers who joined at the same time to understand advancement patterns