This chart shows you the history of ROAS predictions Kohort provides that you can use to get confidence in the predictions and also see what impact product changes are having on cohorts over time
Important items to note
When looking at the global view of all segments, volatility in predictions can be influenced by newly launched networks and segments that are still stabilising. Filter into your main segments to see how the predictions are tracking
ROAS predictions are saved as a snapshot when ever a scenario is run which will depend on what schedule is set for scenarios to run
You will only get ROAS prediction snapshots for the granularity chosen when we have all cohorts in that time period have been acquired and at least 7 days old. So with monthly resolution selected you will get your first result for the prior month when a scenario has been run on the 8th of the month of later
Its important to be aware of what product changes are being made that affect LTV and view the relevant scenario and milestones
Baseline scenario with no milestones - Usually the best choice. Increasing ROAS should tie up with increasing LTV improvements
Baseline scenario with milestones reflecting your budget - View if ROAS is tracking as planned with product improvements. ROAS results should be higher than target if UA team is buy against what LTV they are currently seeing rather than what LTV they expect to get if they deliver product improvements
Results are purged for cohorts older than 6 months and start collecting when you start running the scenario in question. To keep history its better to change settings on an existing scenario than duplicate and create a new scenario with new settings