Performance Planner determines the optimal bids and average daily budget allocations across all the campaigns to help increase the number of conversions it can be achieved for any future spend scenario. Performance Planner uses a combination of account history and machine learning to power forecasts.
Process follows as below:
Forecasting >> Simulation >> Machine Learning >> Validation
1. Forecasting: Google Forcast Engine is made up of Google search ad auction i.e billions of searches per week.
2. Simulation: Forecast engine simulates relevant ad auctions with query level variables, including seasonality, click through rate, competitors, landing page, and time of day.
3. Machine Learning: It uses machine learning to fine tune forecasts and achieve a higher level of accuracy.
4. Validation: Planner perform forward and backward measurements of accuracy for thousands of campaign samples across one, seven, 30, and 90 day periods to ensure the valid recommendations.