Why are there discrepancies in my Chat Offer Effectiveness and ROI Report?
Oracle Engagement Engine Cloud - all releases
Oracle Service Cloud - all releases
The Chat Offer Effectiveness and ROI Report is a report showing performance and ROI broken down by specific offer i.e. by the widgets offering chat.
The report looks at the roi_conversion_stats, the roi_conversion_offer_stats and the visitor_offers tables, using filters on date, time and interval periods.
Due to how the data and the time boundaries are calculated, the end-user may sometimes perceive inconsistencies in what the report is displaying.
These calculations are done with respect to two major factors:
1. Time boundaries: each time an aggregation is done, it is aggregating and correlating data for a specific interval of time. This interval has definitive moments of time that mark the beginning and end of the interval.
2. Visitor sessions: certain events, such as conversions, are correlated back to previous events in the same session (i.e. a conversion is correlated to an accept event of a specific offer).
An example of data inconsistency is detailed below:
- Single visitor opens the customer site at 1:55 PM.
- Visitor receives an offer and accepts at 1:58 PM.
- Visitor completes chat at 2:02 PM.
- Visitor makes purchase at 2:05 PM.
- An aggregation for the 1:00 PM to 2:00 PM hour will show 1 session, 1 offer, 1 accept, 0 conversion
- An aggregation for the 2:00 PM to 3:00 PM hour will show 1 session, 0 offer, 0 accept, 1 conversion
- A calculation of conversion rate in the 2:00 PM to 3:00 PM hour would be 1 conversion/0 accept. This is obviously an invalid calculation with a divide-by-zero problem.
There are two important conclusions we can reach:
1. First is that these reports are most effective when dealing with larger quantities of data (i.e. greater volumes of traffic). The reports will be less effective on sites that only have a handful of visitors.
2. Second is that these reports are meant to show trending data where boundary conditions average out.