Advanced Metrics

We have released a new set of features for Custom Metrics. You can now:

• Use custom formulas for a metric value, incorporating aggregations, operators, and any schema field,
• Define multiple aggregations per metric;
• Exclude or include the current event from an aggregation calculation.

This unlocks a wide range of new use cases, including:

• Ratio metrics:

  • Calculate how the transaction amount compares to usual transactions for a user current transaction / avg(transaction) [per user]
  • Detect suspicious login activity by calculating the ratio of failed to succeeded loginsfailed logins / succeeded logins per user or the ratio of logins outside of usual countries logins outside of US / total logins

  

  
  

• Time difference between events to detect quick actions:

  • Detect if a transaction was attempted immediately after logintransaction [attempted] timestamp - last(successful login timestamp)
  • Detect if a profile update was attempted immediately after loginprofile update [attempted] timestamp - last(successful login timestamp)

  

  
  

• Threshold calculations for a policy acting as a limiter:

  • Set a limit on the transaction amount allowed per day transaction [attempted] amount + sum(transaction [succeeded] amount)

  

  
  

You can find more details on the updated metrics documentation along with examples in the advanced metrics guide.