Transactional analysis

Transactional analysis is the Glass Data layer that evaluates individual events — typically payments and transactions — in real time and returns a risk recommendation to support operational decisions.

Unlike metrics and monitors, which observe aggregated trends over time, transactional analysis processes each event in isolation, at the moment it happens.

How it works

For every event received, the platform runs the sequence below:

  1. Risk score: the platform calculates a score from 0 to 100 based on predictive models and the event data.
  2. Default recommendation: the score is translated into a recommendation — allow, challenge, or deny — according to the account’s analysis configuration.
  3. Rules: if the API call references a rule, its conditions are evaluated and may override the default recommendation.
  4. Response: the final recommendation is returned in the API response, along with the details behind the decision.

Components

  • Analysis: score ranges that define the account’s default recommendation.
  • Rules: custom logic that overrides the default recommendation based on business criteria.
  • Lists: data sets used as parameters in rules.

When to use

Use transactional analysis when you need an automated per-event decision — for example, to authorize or decline a payment, require additional verification, or route a transaction to manual review. To track operational behavior over time, combine it with metrics and monitors.

Table of contents