Algorithmic optimization maximizes a campaign's performance by adjusting the CPM bid price and status based on the campaign goals that you set.
The optimization algorithm changes the CPM bid and/or status of inventory sources that are generating clicks, conversions, or video views.
If a placement or domain meets the goal that you set, Basis DSP will continue to bid higher in paced increments.
If the placement or domain doesn't meet the goal, Basis DSP lowers the bid and/or turns off the placement. You have the option to set additional rules for how and when the bid and status adjustments take place.
How It Works
If click-through rate is the chosen goal, optimization creates rules that maximize the probability of clicks, using the theory that a domain or placement that provided a click in the past has a higher probability of providing a click again in the future. The algorithm increases bids for domains or placements that have produced more clicks, similar to what a buyer would do manually.
Note: Because its predictions are based on clicks from the past, a campaign must see clicks before the optimization algorithm can act. Click probabilities are quite low, and when the number of clicks is low – often in the early phases of a campaign – expect larger fluctuations in CTR, and then smaller transitions once the campaign has a larger sample size of clicks available.
Enabling Algorithmic Optimization for a Campaign
Create a new campaign.
In the Budget section of the General tab, be sure that the campaign has a start date and an end date defined. If the group specifies flight dates, campaign flight dates are optional.
In the Optimization section of the General tab, select Algorithmic optimization.
Choose a Goal type and target value.
Note: When your campaign is running, don't change the goal type mid-flight. If you need to run the campaign with a different goal type, make a copy of the campaign and edit the goal type in the new copy.
The target value can be a dollar amount or percentage, depending on the goal type you choose:
- Click-Through Rate (CTR): Clicks to Impressions ratio, as a percentage: (Clicks / Impressions) * 100
- The default value is 0.05% CTR.
- Cost Per Click (eCPC): Cost to acquire a click: (Media Cost / Clicks)
- The default value is $5.00 eCPC.
- Cost Per Acquisition (eCPA): Cost to acquire an acquisition or conversion: (Media Cost / Conversions)
- The default value is $25 eCPA.
- Cost Per Completed Video (eCPCV): Cost to run a video through completion: (Media Cost / Completed Video Views)
- The default value is $0.10.
- Video Completion Rate (VCR): Video Completions to Impressions ratio, as a percentage: (Completed Video Views / Impressions)*100
- The default value is 50% VCR.
Choose a Maximum Bid, the highest CPM bid that can be a given to placement. The maximum bid should be higher than the campaign's Default CPM Bid. For more on setting a Maximum Bid, see Tips for Setting a Maximum Bid.
Choose to optimize on Placements or Domains. Optimizing by domains (the default) considers all placements in the domain together.
Optional: Configure the Advanced optimization settings.
On the Inventory tab, choose the exchanges you want the campaign to run on. Optimized campaigns can be enabled to Run on Selected Exchanges or Run on All Exchanges.
Tips for Setting a Maximum Bid
Setting a maximum bid doesn't mean a placement is going to reach that bid. Assuming a placement is meeting the desired goals, the bid step, learn budget, and campaign duration will impact the bid price.
Remember that smaller maximum bids give the optimizer fewer options when making changes to the default bid to win inventory.
Consider what inventory you want a campaign to win: in the inventory screen, locate some domains that you would be interested in winning impressions at and check their price. Set your max bid at $1 - $2 higher than this.
eCPC, eCPA, or eCPCV
For these goal types, choose a maximum bid based on the campaign goal as well as a reasonable estimation of "best case" CTR, conversion rate or completion rate.
Other Optimization Types
You can't use machine learning optimization and algorithmic optimization at the same time. If you switch a campaign from algorithmic optimization to machine learning optimization, the campaign returns inventory bid and status settings to their defaults. If you've manually turned off any inventory, it remains off.
You can use ad optimization while using algorithmic optimization.