Assign algorithm

Assign the algorithm to a resource once it is ready.

Check if the algorithm is ready

An algorithm model must train on a minimum amount of data before it is ready.

Algorithms train a model for each advertiser that can use them. Models train on existing impressions data. It can take 1-3 days to train a model after the data requirements are met by the advertiser.

Retrieve the readiness of each model from the algorithm. The ReadinessState determines next steps:

ReadinessState
READINESS_STATE_NO_VALID_SCRIPT	There is no valid script. Upload a new script or rules file.
READINESS_STATE_EVALUATION_FAILURE	There is no script that could be evaluated in the allotted time. Upload a new script or rules file.
READINESS_STATE_INSUFFICIENT_DATA	There is not a sufficient amount of data available under the advertiser to train the model. Continue to run campaigns under the advertiser to meet the minimum data requirements.
READINESS_STATE_TRAINING	The model is training on existing data and not ready for serving. Wait 12-24 hours before checking to see if the model is ready for serving.
READINESS_STATE_ACTIVE	The model is fully trained and ready to be assigned to campaigns under the advertiser.

Tip: If you are retrieving an algorithm with an advertiser ID, only that advertiser's model will be included in the response.

Assign algorithm to line item

An algorithm is used to measure and adjust bidding performance. It can be used with bid strategies that optimize towards spending the whole budget or meeting a goal. In these cases, the strategy performance goal type would be BIDDING_STRATEGY_PERFORMANCE_GOAL_TYPE_CUSTOM_ALGO.

Warning: If no spend is associated with a model for 21 days, it is suspended. Suspended models don't train on new data. Suspension status can be found in the advertiser's model details. To use a suspended model, reactivate it by assigning the algorithm to an active line item. The line item will need to begin spending within 21 days.

Here's how to update a line item to use an algorithm with a bid strategy that optimizes for spending the whole budget:

Java

// Provide the ID of the advertiser that owns the parent algorithm.
long advertiserId = advertiser-id;

// Provide the ID of the parent algorithm.
long algorithmId = algorithm-id;

// Provide the ID of the line item to assign the algorithm to.
long lineItemId = line-item-id;

// Create the line item structure.
LineItem lineItem =
    new LineItem()
        .setBidStrategy(
            new BiddingStrategy()
                .setMaximizeSpendAutoBid(
                    new MaximizeSpendBidStrategy()
                        .setPerformanceGoalType(
                            "BIDDING_STRATEGY_PERFORMANCE_GOAL_TYPE_CUSTOM_ALGO")
                        .setCustomBiddingAlgorithmId(algorithmId)));

// Configure the patch request and set update mask to only update the bid
// strategy.
LineItems.Patch request =
    service
        .advertisers()
        .lineItems()
        .patch(advertiserId, lineItemId, lineItem)
        .setUpdateMask("bidStrategy");

// Update the line item.
LineItem response = request.execute();

// Display the new algorithm ID used by the line item.
System.out.printf(
    "Line item %s now uses algorithm ID %s in its bidding strategy.",
    response.getName(),
    response.getBidStrategy().getMaximizeSpendAutoBid().getCustomBiddingAlgorithmId());

Python

# Provide the parent advertiser ID of the line item and creative.
advertiser_id = advertiser-id

# Provide the ID of the creative to assign.
algorithm_id = algorithm-id

# Provide the ID of the line item to assign the creative to.
line_item_id = line-item-id

# Build bidding strategy object.
bidding_strategy_obj = {
    "maximizeSpendAutoBid": {
        "performanceGoalType": (
            "BIDDING_STRATEGY_PERFORMANCE_GOAL_TYPE_CUSTOM_ALGO"
        ),
        "customBiddingAlgorithmId": algorithm_id,
    }
}

# Build line item object.
line_item_obj = {"bidStrategy": bidding_strategy_obj}

# Build and execute request.
line_item_resp = (
    service.advertisers()
    .lineItems()
    .patch(
        advertiserId=advertiser_id,
        lineItemId=line_item_id,
        updateMask="bidStrategy",
        body=line_item_obj,
    )
    .execute()
)

# Print the algorithm ID now assigned to the line item.
print(
    f'Line Item {line_item_resp["name"]} now uses algorithm ID'
    f' {line_item_resp["bidStrategy"]["maximizeSpendAutoBid"]["customBiddingAlgorithmId"]}'
    " in its bidding strategy."
)

PHP

// Provide the ID of the advertiser that owns the parent algorithm.
$advertiserId = advertiser-id;

// Provide the ID of the parent algorithm.
$algorithmId = algorithm-id;

// Provide the ID of the line item to assign the algorithm to.
$lineItemId = line-item-id;

// Create the bidding strategy structure.
$maxSpendBidStrategy = new Google_ServiceDisplayVideo_MaximizeSpendBidStrategy();
$maxSpendBidStrategy->setPerformanceGoalType('BIDDING_STRATEGY_PERFORMANCE_GOAL_TYPE_CUSTOM_ALGO');
$maxSpendBidStrategy->setCustomBiddingAlgorithmId($algorithmId);
$biddingStrategy = new Google_ServiceDisplayVideo_BiddingStrategy();
$biddingStrategy->setMaximizeSpendAutoBid($maxSpendBidStrategy);

// Create the line item structure.
$lineItem = new Google_Service_DisplayVideo_LineItem();
$lineItem->setBidStrategy($biddingStrategy);
$optParams = array('updateMask' => 'bidStrategy');

// Call the API, updating the bid strategy for the identified line
// item.
try {
    $result = $this->service->advertisers_lineItems->patch(
        $advertiserId,
        $lineItemId,
        $lineItem,
        $optParams
    );
} catch (\Exception $e) {
    $this->renderError($e);
    return;
}

printf(
    '<p>Line Item %s now uses algorithm ID %s in its bid strategy.</p>',
    $result['name'],
    $result['bidStrategy']['maximizeSpendAutoBid']['customBiddingAlgorithmId']
);