Managing thousands of ads manually to improve ROI is very labor intensive. As the number and complexity of campaigns grows, it becomes beneficial and even necessary to automate bid and budget management with relevant technology.
Approaches to automated optimization of bid and budget allocation may roughly be put in one of two categories: rule-based and algorithmic.
A rule-based approach allows marketers to define changes in bid amounts and budget allocation when the observed performance of components of a campaign satisfies certain conditions. Rules are usually applied to individual constituents of a campaign (creative, keywords, etc.) in isolation.
For example, the incremental bid adjustments available in AdWords are a pretty good version of this, as are CPA and CPC caps. The popular automation site, If This Then That , is a literal version of rule-based decision-making - when a certain condition is met, complete an action.
The algorithmic approach adjusts bids and budget based on performance of campaign components together to increase the overall ROI - or value - of a campaign. This approach requires some combination of offline analysis of historical data and online learning of bid and budget allocation on campaign performance.
Maintaining an ongoing data set allows us to use probabilities
to determine whether a specific action is likely to occur for each
of the items observed. We then distribute budget based on that
probability. It's non-linear and far more active than a traditional
rules-based approach.
A closely-related concept is "portfolio optimization" that implies that a campaign is optimized as a whole rather than as a collection of unique components.
| Rule-based | Algorithmic | |
| Advantages |
|
|
| Disadvantages |
|
|