Data validation (DV) is a monitoring tool that provides an in-depth overview of the collected data, it will allow you to identify and troubleshoot data issues in the early stages.
DV status for each integration will appear on the ad network integration screen, it compares the volume of the ad network API data (clicks) vs the data Soomla SDK data.
For that purpose, the user can early drill down and investigate data discrepancies.
DV data is presented with the following break down along scores for each:
- integration(Ad Network)
- location (US or Rest of the world)
- Ad type
- Revenue proportion - indicate the portion of the total revenue for the past 7 days(excluding last day) generated by this ad network
- Data health score - an average of the revenue coverage weighted by API Revenue for the past 7 days (excluding last day)
- Health signs:
- When revenue proportion is larger than 10 % and the health score is more than 85%
- When revenue proportion is less than 10 % and the health score is more than 85%
- When revenue proportion is larger than 10 % and the health score is less than 85%
Integration screen including an expandable section that indicates SDK data healthiness with scores.
Data healthiness is measured by comparing API data (Clicks) vs SDK data
Our scoring mechanism is based on the daily revenue coverage and daily impression coverage, scores representing the average of the revenue coverage weighted by API Revenue.
** last day is excluded from this logic as it may represent partial data
- API revenue is the ad revenue that was reported by the ad network and collected by Soomla through ad network API
- Revenue coverage is a ratio between the revenue that was calculated using the SDK divided with the revenue that was reported by the ad network
- Impression coverage is the ratio between the Impression that was calculated using the SDK divided with the Impression that was reported by the ad network
- Daily indication
- Is indicate that both revenue & impression coverage is above 85% and less than 115%
- Is indicate that revenue or impression coverage less than 85% or over 115%
- Last day data is in process, for this reason, it may be partial. Therefore, the last day is excluded from scoring logic
- Ad type - Data health sign (refers to the previous 6 days, excluding last day)
- all previous 6 days there were no discrepancies for this ad type
- all previous 6 days there were discrepancies for this ad type
- on the previous 6 days, there were some discrepancies for this ad type
- Location breakdown (US & Rest of the world) with a weighted average
- Platform weighted average score
- The number of daily discrepancies that were detected during 6 days (excluding last day)