Cost Control
zinzu provides two independent controls to manage how much data the sequence engine processes. Use either one or both.
Sampling: Analyze fewer entities
Choose a percentage of customers, orders, requests or other entities to analyze. zinzu selects entities consistently across source tables and keeps their records within the selected time range, preserving their timelines.
ExampleSelect approximately 10% of customers to explore patterns before running a full analysis.
Data-skew control: Exclude unusually large timelines
A few entities can have far more records than the rest, creating disproportionately heavy processing. zinzu counts records per entity across the selected sources and time range, then excludes entities above your chosen limit.
ExampleExclude entities with more than 50,000 records, while continuing to analyze the remaining entities.
Understand the scope
Sampling reduces the number of timelines. Data-skew control limits unusually large timelines. Both change the population analyzed; actual cost savings depend on the data and query.