Skip to main content

​
What is VoicePact Analytics?

VoicePact Analytics provides business intelligence on voice-based contracts by extracting, transforming, and loading (ETL) operational data into a dedicated analytics database. This enables stakeholders to gain insights into contract performance, payment patterns, and user behavior without impacting the production system.

​
Key Capabilities

​
Business Metrics

The analytics system tracks critical business metrics including:
  • Contract Performance: Time to completion, completion rates, and contract statuses
  • Payment Analytics: Total amounts released, payment velocities, and escrow patterns
  • User Activity: Buyer and seller participation, contract types, and product categories
  • Temporal Patterns: Contract creation trends, completion timelines, and seasonal patterns

​
Data Model

The analytics database uses a dimensional model centered around a fact table that consolidates contract, party, and payment information:

​
Fact Table: fct_contracts_summary

This is the primary analytics table containing:
ColumnTypeDescription
contract_idStringUnique contract identifier
contract_typeStringType of contract (agricultural_supply, service_agreement, etc.)
contract_statusStringCurrent status (pending, active, completed, etc.)
contract_total_amountDecimalTotal contract value
currencyStringCurrency code (e.g., KES)
productStringProduct/service being contracted
created_atDatetimeContract creation timestamp
completed_atDatetimeContract completion timestamp
time_to_completion_daysIntegerDays from creation to completion
buyer_phoneStringBuyer’s phone number
seller_phoneStringSeller’s phone number
total_paid_releasedDecimalTotal amount released from escrow

​
Source Data

The ETL pipeline extracts data from three operational tables:
  1. contracts: Core contract information including terms, amounts, and status
  2. contract_parties: Buyer, seller, mediator, and witness information
  3. payments: Payment transactions and escrow releases

​
Example Queries

Once the ETL pipeline has run, you can query the analytics database for insights:

​
Total Contracts by Type

SELECT 
    contract_type,
    COUNT(*) as total_contracts,
    SUM(contract_total_amount) as total_value
FROM fct_contracts_summary
GROUP BY contract_type
ORDER BY total_value DESC;

​
Average Time to Completion

SELECT 
    contract_type,
    AVG(time_to_completion_days) as avg_days_to_complete,
    COUNT(*) as completed_contracts
FROM fct_contracts_summary
WHERE contract_status = 'completed'
  AND time_to_completion_days IS NOT NULL
GROUP BY contract_type;

​
Payment Release Rate

SELECT 
    contract_status,
    COUNT(*) as contract_count,
    SUM(contract_total_amount) as total_contracted,
    SUM(total_paid_released) as total_released,
    ROUND(100.0 * SUM(total_paid_released) / NULLIF(SUM(contract_total_amount), 0), 2) as release_rate_pct
FROM fct_contracts_summary
GROUP BY contract_status;

​
Top Products by Volume

SELECT 
    product,
    COUNT(*) as contract_count,
    SUM(contract_total_amount) as total_value,
    AVG(contract_total_amount) as avg_contract_value
FROM fct_contracts_summary
WHERE product IS NOT NULL
GROUP BY product
ORDER BY contract_count DESC
LIMIT 10;

SELECT 
    strftime('%Y-%m', created_at) as month,
    COUNT(*) as contracts_created,
    SUM(CASE WHEN contract_status = 'completed' THEN 1 ELSE 0 END) as contracts_completed,
    SUM(contract_total_amount) as total_value
FROM fct_contracts_summary
GROUP BY month
ORDER BY month DESC;

​
Architecture

​
Separation of Concerns

VoicePact uses a separate analytics database to:
  • Prevent query interference: Heavy analytical queries don’t impact production performance
  • Enable schema optimization: The analytics schema is denormalized for query performance
  • Support data retention: Historical snapshots can be maintained independently
  • Facilitate external integrations: BI tools can connect without production access

​
ETL Process

The ETL pipeline follows a classic three-stage approach:
  1. Extract: Queries the production database to retrieve raw contract, party, and payment data
  2. Transform: Cleans, enriches, and joins data to create analytics-ready summaries
  3. Load: Writes the transformed data to the analytics database
See the ETL Pipeline documentation for implementation details.

​
Use Cases

​
Business Intelligence

  • Track contract completion rates by type
  • Identify bottlenecks in the contract lifecycle
  • Monitor payment release patterns
  • Analyze seasonal trends in contract creation

​
Operational Monitoring

  • Detect anomalies in contract processing times
  • Monitor escrow fund flows
  • Track user engagement by phone number
  • Identify high-value contracts requiring attention

​
Product Development

  • Understand which contract types are most popular
  • Analyze feature usage patterns
  • Identify opportunities for automation
  • Measure the impact of product changes

​
Next Steps

Build docs developers (and LLMs) love

Get started for freeTalk to us