Hey everyone, I'm currently working with fact tables to build metrics, and things are going well so far. However, I’ve run into something I don’t quite understand, and I’m hoping you can help me make sense of it. I have a fact table that includes all purchase events. Based on this, I created two metrics: 1. A simple count of orders 2. The average order value – calculated as a ratio (revenue sum / order count), using a ratio metric For the ratio metric, I applied percentile capping at 0.95. Now here’s what’s confusing: The number of transactions (orders) used in the ratio metric is lower than the number in the basic order count metric. When I remove the capping, both numbers match – as expected. According to the documentation, “the purchase will still be counted, but the aggregated value for that user will be capped.” Based on that, I’d expect the count to remain unchanged, even with capping. So I’m not sure why the order count differs when capping is applied. For context: • The "revenue: all transactions" metric is built using a fact table • The "revenue" metric is based on custom SQL queries Any ideas what might be causing this discrepancy? Thanks a lot in advance!