Custom bidding predicts the value of each impression based on conversion values and custom variables, such as transaction values from Floodlights. Identify key conversion points to direct the system to deliver towards metrics more aligned with your business goals.
Typically, not all conversions are created equal. Weighted conversions provide a good opportunity to differentiate conversions based on their value within the funnel (e.g. page visit, sign up, and transaction). The time-decay model is the optional model for weighing conversions and calculating appropriate weights. It’s based on the total value of the final goal and the total number of events recorded. Remember to factor in how many times every event has been tracked by the Floodlight when building your model, and adjust values accordingly.
To calculate conversion weights:
- Decide on a time-decay rate (1-100%) to calculate proxy value of conversion events.
- The earlier the event occurs in the purchase journey, the less value it has.
- Calculate the conversion value of events based on the value of a real conversion.
- Weigh every conversion by dividing the calculated event value by the number of events.
Once you have defined what is most important for your business goals, you can rely on Google’s Custom Bidding to analyze over 40 signals (including auction dynamics and probability of conversion) to deliver the best outcomes for your investment.
Custom bidding combines your unique insights with Google’s AI to create a bespoke algorithm in order to achieve results specific to your business needs. To inform your custom bidding algorithm, you may use data from:
The what:
Use u-variables to define the importance of certain actions, such as loyalty tier, purchase value, and items purchased. This helps the system identify and target impressions most likely to lead to the desired outcome.
The why:
Without this steer, all actions, such as purchases, will be valued the same, which can lead to wastage and means you fall short of maximizing your revenue.
The what:
Analytics goals can provide a deeper insight into consumer actions on the website. This can help you understand what a valuable consumer looks like and guide the algorithm to bid accordingly, optimizing towards engaging the most valuable customers.
The why:
You can combine goals to describe specific behaviors that are considered to be valuable consumer actions; for example, the amount of time a user spends on site, or if they subscribe to a newsletter.
The what:
Use durable, quality-based signals, such as location, audibility, exchange, and time on screen, to improve impression quality, which is not captured by tracking technologies like Floodlights and plays a key role in delivering value.
The why:
You can rely on your reports to identify the specific impression signal levels (for example, TOS10) that add value to your business and include them in custom bidding scripts (custom script use case).
To begin your custom bidding journey, start by defining which actions are significant in the customer journey (e.g. newsletter subscription, favoriting a product). These vary in importance and in their ability to bring value to your business.
Once you know the impact these actions have, you can then weigh them accordingly and provide the platform with this information. Doing so allows you to steer the system to prioritize one customer over the other and bid accordingly. This enables the algorithm to set more precise bids that align with the real business value of a conversion.
Within the DV360 UI, you can make use of the goal builder to create custom rules for the system to bid towards, without having to write a script.
Once the required parameters have been set up, you need to choose a scoring method of conversions and an attribution model to complete the elements needed to create your bidding model.
DV360 provides the option to create a custom bidding script to accurately represent the KPIs you want to improve when using your custom bidding model. This allows you to articulate your bidding intentions with more granularity.
You may also choose to create custom bidding algorithms by using the DV360 API to upload and verify scripts and assign algorithms to campaigns.
