Proactive Bid Adjustments based on Revenue & ACoS Forecasts
Based on the ACoS of the previous 3 days, as well as the predicted ACoS of the upcoming 4 days, Autopilot will constantly adjust your bids in order to reach the user defined target ACoS across 7 day time span.
How is the Upcoming Revenue Predicted?
Autopilot takes into account several factors, as detailed below:
- Keyword Clustering
Autopilot will predict the revenue per click of each keyword by analyzing which keywords behave in a similar way. Thereby, it will analyze historical data of all similar keywords within the whole account.
- Average Sales Attribution Latency
Sellics analyzes the latency for PPC orders by measuring past data with regards to how long it takes for orders to come in after their clicks for specific keywords and products, as well as how big the revenue difference is when looking at data 1 day after, compared to 2-14 days later.
The latent attributed revenue prediction also considers various other parameters such as day of week.
- Recent Performance Trends
Autopilot will look at each target's and product's recent historical data and analyze how the performance is evolving. Therefore, it gives more weight to more recent historical data.
The ACoS forecast will also consider recent performance data. Therefore, it will neither "overreact" to other factors nor predict unusually big changes. The forecast will instead expect the ACoS to be at least similar to recent performance data.
Additionally, Autopilot considers how the performance trends evolved in in the past for the same week days, as well as within the same months, if that history is available.
- Day of Week Seasonality
Autopilot looks at historical "day of week" data to anticipate revenues for specific days, eg. Mondays.
- Day of Year Seasonality
If historical data is available (ie. the Sellics account has been active for long enough), Autopilot can factor in the day of year seasonality.
- Special Days (eg. Prime Day)
Finally, Autopilot will also factor in historical data on special e-commerce relevant days and periods such as Black Friday, Easter or Christmas.
Nonetheless, we recommend preparing your bids / budgets manually according to your strategies for these holidays. See for additional webinars on how to navigate your business to best take advantage of these holidays.
- Max Bid Constraint based on Target ACoS + Predicted Revenue
In order to set the highest possible but still affordable bid, Autopilot calculates the average CPC you can afford for each target.
That CPC is based on your expected average revenue per click and Target ACoS.
Target ACoS = 30%
Expected Revenue per Click
= Expected Conversion Rate (10%) x Expected Revenue per Conversion (20$)
= 10% x $20 = $2
Autopilot can afford an average CPC (Cost per Click) of maximum:
= Expected Revenue per Click x Target ACoS
= $2 x 30%
CPC is generally a little lower than the bid -- Sellics assumes 0.8 times lower.
Therefore, Autopilot's bid is designed to stay below the affordable CPC = $0.6 ÷ 0.8 = $0.75
As an additional precaution, Autopilot will never increase the bid to a value higher than 15 times the CPC, as the bid would not bring more volume in that situation.