The Nativo Optimizer (a) improves the performance of ads and (b) ensures all your budgets deliver in full, while taking available inventory, prices, floors, serving priorities, and reservations into account.
The Nativo Optimizer (a) improves the performance of ads and (b) ensures all your budgets deliver in full, while taking available inventory, prices, floors, serving priorities, and reservations into account.
The Optimizer’s job is to provide input for the Ad Server to:
How to allocate impressions optimally (globally, for all budgets)
What specific inventory a given budget will run in
What creative should run
How frequently an ad should run (indirectly)
Provide pacing cues to ad server
Provide information used to enforce publisher floors for performance-based rate types like CPC
It is important to understand that the Optimizer does not create inventory or performance, it tries to predict the best way to maximize things given conflicting goals and limited inventory. The Optimizer balances tradeoffs between prices, performance, and delivery globally.
How does the optimization work?
Placement & Creative Optimization
Optimization operates at the placement and creative levels. The Budget (including all targeting and other settings) is used as input, and while Campaigns help group Budgets, they don’t inform the optimization itself. There are 2 important decisions we need to make for any given budget: (1) where it will run and (2) what ad creative and creative variation to show.
Placement Optimization
Our Ad Server depends on the Optimizer’s input to help it decide what Placements will provide the best performance, given a Budget’s “Optimize By” goal and targeting constraints. In fact, the Optimizer looks at performance signals for Placements sliced by device and category, since these dimensions often have unique characteristics, to provide even more accurate performance estimates. The Optimizer provides its input to the Ad Server in the form of weights for each Budget and Placement slice.
Creative Optimization
Creative optimization can be thought of as a second step after placement optimization. Once a placement slice is identified for the budget, the Ad Server must select a specific creative, headline, preview image, and content variation to run for a given impression.
This involves two components: a) selecting the Ad creative and b) selecting the creative variation. For Budgets that include multiple Ads, step (a) involves choosing the best performing Ad creative, given the optimization goal. Next, step (b) chooses from among the headlines, preview images, or content variations available in the creative itself.
Creative optimization also uses historical performance of the creative across other campaigns and budgets to help inform decisions. This is particularly useful early in the flight, when there is less data available from the current budget. As new data comes in throughout the flight, the optimizer will adjust based on that new data.
NOTE: When “No Optimization” is selected, ad creatives and ad variations are chosen at random, evenly cycling through all the options regardless of performance.
Optimization Frequency
The Optimizer evaluates the entire network and all your campaigns every hour. This frequency allows it to be responsive to changes while allowing time to accumulate enough real-world performance data before changing the optimization / inventory allocation.
Since newly started budgets don’t have historical performance data, they start running in “sampling mode” in the first 1-2 hours. Sampling mode distributes budgets more broadly than normal (within the specified targeting parameters) in order to identify the best performing placements. As a result, performance typically improves after the sampling period.
Note that the Ad Server uses information provided by the Optimizer between runs to react to real time changes in inventory during the hour, allowing for continued improvement in performance between runs.
FAQ
Can I turn off the optimizer and run my campaigns without it?
Yes, but it isn’t recommended. By setting your “Optimize By” goal to “No Optimization”, you can disable any goal-based optimization (e.g. Clicks) from being performed. The Optimizer still creates an allocation for your campaign, but does not take into account any performance considerations.
Why don’t the highest CTR headlines and preview images get the most impressions?
There are a number of reasons why the headline and preview image with the highest average CTR won’t always get the most volume.
Recall that the optimizer looks at not only what creative combination to run but also where to run it. Both of these inputs influence performance. As a result, the same headline and preview image may be a star in one placement and a dud in another. Couple that with the fact that inventory is not distributed evenly and the optimizer may just not find enough opportunities to run in the perfect spot.
Downstream optimization goals don’t always translate into the highest CTR upstream. For example, optimizing for time on content may mean taking a hit in CTR (since low value-clicks are avoided).
How do Budget and Deal priorities affect Optimizer decisions?
The Optimizer takes budget priorities into account whenever multiple budgets are competing for the same inventory. This doesn’t mean that lower-priority budgets won’t deliver, but it may mean that they will be “squeezed out” of their ideal placements or in more extreme cases might even miss their delivery goal. On inventory-heavy days, you might not notice an impact of priorities at all, but on tight days you might find that lower-priority budgets see reduced performance.
Note that raising priority for a given Budget will only have an impact if most other Budgets are set to lower priorities. If all budgets are at the same priority (whether Standard or Critical), the Optimizer can’t use priority to decide which budget is more important than another.
How do Demand Channel priorities effect Optimizer decisions?
The Monetization UI allows you to set priority levels for each Demand Channel. This gives publishers granular control over prioritization of demand.
Demand (Budgets and programmatic bids) within Demand Channels that have the same priority will compete with each other head-to-head. If one demand channel has a higher priority than another then its budgets or bidders will win over a lower-priority channel regardless of price.
You can use Demand Channel priorities to tier your supply, creating negotiating leverage with your buyers.