Gerar métricas de estimativa

As métricas de previsão fornecem métricas para campanhas propostas ou existentes, incluindo:

• Impressões
• Taxa de cliques
• Custo médio por clique
• Cliques
• Custo

Teste diferentes configurações de campanha para otimizar cada uma dessas métricas. Dados como a pontuação de qualidade histórica e a taxa de cliques podem ser usados na previsão para simular a performance esperada, como a taxa de conversão.

As métricas de previsão geralmente são usadas depois das métricas históricas. As métricas históricas reduzem uma lista de palavras-chave grande a um tamanho mais gerenciável. As métricas de previsão oferecem uma estimativa mais exata da performance futura da campanha. Comece com as métricas históricas das palavras-chave. Quando a lista estiver gerenciável, use as métricas de previsão para otimizar a performance da sua campanha.

Gerar métricas

Para gerar métricas de previsão, siga estas etapas:

1. Crie uma campanha que você quer estimar CampaignToForecast.
2. Chame KeywordPlanIdeaService.GenerateKeywordForecastMetrics com esse CampaignToForecast e qualquer outro parâmetro, como forecast_period.

Criar uma campanha para fazer estimativas

Confira algumas dicas para criar uma campanha para fazer a previsão:

• Crie os grupos de anúncios com base em temas, como relevância do criativo, categoria de produto ou custo por clique.
• Inclua palavras-chave negativas comuns. Por exemplo, se você não está em recrutamento, exclua a palavra-chave "empregos".
• Use uma conta relacionada à campanha. Não faça previsões em uma conta não relacionada. O Google considera elementos como índice de qualidade e conteúdo criativo na previsão.
• Se uma campanha não estiver relacionada ao restante da sua conta, use uma nova conta limpa para a previsão.
• Repita as previsões com horizontes diferentes para ter uma visão holística.

Gerar métricas de previsão

O próximo exemplo de código faz uma solicitação para KeywordPlanIdeaService.GenerateKeywordForecastMetrics. Em seguida, ela itera em cada um dos resultados e exibe cada uma das métricas da previsão.

A resposta também pode retornar IDs temporários no recurso, por exemplo, customers/CUSTOMER_ID/keywordPlanCampaigns/-1.

private void runExample(GoogleAdsClient googleAdsClient, Long customerId) {
  CampaignToForecast campaignToForecast = createCampaignToForecast(googleAdsClient);
  GenerateKeywordForecastMetricsRequest request =
      GenerateKeywordForecastMetricsRequest.newBuilder()
          .setCustomerId(String.valueOf(customerId))
          .setCampaign(campaignToForecast)
          .setForecastPeriod(
              DateRange.newBuilder()
                  // Sets the forecast start date to tomorrow.
                  .setStartDate(new DateTime().plusDays(1).toString("yyyy-MM-dd"))
                  // Sets the forecast end date to 30 days from today.
                  .setEndDate(new DateTime().plusDays(30).toString("yyyy-MM-dd")))
          .build();
  try (KeywordPlanIdeaServiceClient keywordPlanIdeaServiceClient =
      googleAdsClient.getLatestVersion().createKeywordPlanIdeaServiceClient()) {
    GenerateKeywordForecastMetricsResponse response =
        keywordPlanIdeaServiceClient.generateKeywordForecastMetrics(request);
    KeywordForecastMetrics metrics = response.getCampaignForecastMetrics();
    System.out.printf(
        "Estimated daily clicks: %s%n", metrics.hasClicks() ? metrics.getClicks() : null);
    System.out.printf(
        "Estimated daily impressions: %s%n",
        metrics.hasImpressions() ? metrics.getImpressions() : null);
    System.out.printf(
        "Estimated average CPC (micros): %s%n",
        metrics.hasAverageCpcMicros() ? metrics.getAverageCpcMicros() : null);
  }
}

/**
 * Creates the campaign to forecast. A campaign to forecast lets you try out various
 * configurations and keywords to find the best optimization for your future campaigns. Once
 * you've found the best campaign configuration, create a serving campaign in your Google Ads
 * account with similar values and keywords. For more details, see:
 *
 * <p>https://support.google.com/google-ads/answer/3022575
 *
 * @param googleAdsClient
 * @return
 */
private CampaignToForecast createCampaignToForecast(GoogleAdsClient googleAdsClient) {
  CampaignToForecast.Builder campaignToForecastBuilder =
      CampaignToForecast.newBuilder()
          .setKeywordPlanNetwork(KeywordPlanNetwork.GOOGLE_SEARCH)
          .setBiddingStrategy(
              CampaignBiddingStrategy.newBuilder()
                  .setManualCpcBiddingStrategy(
                      ManualCpcBiddingStrategy.newBuilder().setMaxCpcBidMicros(1_000_000L)));

  // See https://developers.google.com/google-ads/api/reference/data/geotargets for the list of
  // geo target IDs.
  campaignToForecastBuilder.addGeoModifiers(
      CriterionBidModifier.newBuilder()
          // Geo target constant 2840 is for USA.
          .setGeoTargetConstant(ResourceNames.geoTargetConstant(2840)));

  // See https://developers.google.com/google-ads/api/reference/data/codes-formats#languages for
  // the list of language criteria IDs. Language constant 1000 is for English.
  campaignToForecastBuilder.addLanguageConstants(ResourceNames.languageConstant(1000));

  // Create forecast ad group based on themes such as creative relevance, product category, or
  // cost per click.
  ForecastAdGroup.Builder forecastAdGroupBuilder = ForecastAdGroup.newBuilder();
  forecastAdGroupBuilder.addBiddableKeywords(
      BiddableKeyword.newBuilder()
          .setMaxCpcBidMicros(2_500_000)
          .setKeyword(
              KeywordInfo.newBuilder()
                  .setText("mars cruise")
                  .setMatchType(KeywordMatchType.BROAD)));

  forecastAdGroupBuilder.addBiddableKeywords(
      BiddableKeyword.newBuilder()
          .setMaxCpcBidMicros(1_500_000)
          .setKeyword(
              KeywordInfo.newBuilder()
                  .setText("cheap cruise")
                  .setMatchType(KeywordMatchType.PHRASE)));

  forecastAdGroupBuilder.addBiddableKeywords(
      BiddableKeyword.newBuilder()
          .setMaxCpcBidMicros(1_990_000)
          .setKeyword(
              KeywordInfo.newBuilder()
                  .setText("jupiter cruise")
                  .setMatchType(KeywordMatchType.BROAD)));

  forecastAdGroupBuilder.addNegativeKeywords(
      KeywordInfo.newBuilder().setText("moon walk").setMatchType(KeywordMatchType.BROAD));

  campaignToForecastBuilder.addAdGroups(forecastAdGroupBuilder.build());
  return campaignToForecastBuilder.build();
}
GenerateForecastMetrics.java
      

public void Run(GoogleAdsClient client, long customerId)
{
    CampaignToForecast campaignToForecast = CreateCampaignToForecast();

    KeywordPlanIdeaServiceClient keywordPlanIdeaService =
            client.GetService(Services.V18.KeywordPlanIdeaService);

    GenerateKeywordForecastMetricsRequest request = new GenerateKeywordForecastMetricsRequest()
    {
        CustomerId = customerId.ToString(),
        Campaign = campaignToForecast,
        ForecastPeriod = new DateRange()
        {
            // Set the forecast start date to tomorrow.
            StartDate = DateTime.Now.AddDays(1).ToString("yyyy-MM-dd"),
            // Set the forecast end date to 30 days from today.
            EndDate = DateTime.Now.AddDays(30).ToString("yyyy-MM-dd"),
        }
    };

    try
    {
        GenerateKeywordForecastMetricsResponse response =
            keywordPlanIdeaService.GenerateKeywordForecastMetrics(request);

        KeywordForecastMetrics metrics = response.CampaignForecastMetrics;

        Console.WriteLine($"Estimated daily clicks: {metrics.Clicks}.");
        Console.WriteLine($"Estimated daily impressions: {metrics.Impressions}.");
        Console.WriteLine($"Estimated average cpc (micros): {metrics.AverageCpcMicros}.");
    }
    catch (GoogleAdsException e)
    {
        Console.WriteLine("Failure:");
        Console.WriteLine($"Message: {e.Message}");
        Console.WriteLine($"Failure: {e.Failure}");
        Console.WriteLine($"Request ID: {e.RequestId}");
        throw;
    }
}

/// <summary>
/// Creates the campaign to forecast. A campaign to forecast lets you try out
/// various configuration and keywords to find the best optimization for your
/// future campaigns. Once you've found the best campaign configuration,
/// create a serving campaign in your Google Ads account with similar values
/// and keywords. For more details, see:
/// https://support.google.com/google-ads/answer/3022575
/// </summary>
private CampaignToForecast CreateCampaignToForecast()
{
    CampaignToForecast campaignToForecast = new CampaignToForecast()
    {
        KeywordPlanNetwork = KeywordPlanNetwork.GoogleSearch,
        BiddingStrategy = new CampaignToForecast.Types.CampaignBiddingStrategy()
        {
            ManualCpcBiddingStrategy = new ManualCpcBiddingStrategy()
            {
                MaxCpcBidMicros = 1_000_000
            }
        }
    };

    // See https://developers.google.com/google-ads/api/reference/data/geotargets
    // for the list of geo target IDs.
    campaignToForecast.GeoModifiers.Add(new CriterionBidModifier()
    {
        // Geo target constant 2840 is for USA.
        GeoTargetConstant = ResourceNames.GeoTargetConstant(2840)
    });

    // See https://developers.google.com/google-ads/api/reference/data/codes-formats#languages
    // for the list of language criteria IDs.
    // Language constant 1000 is for English.
    campaignToForecast.LanguageConstants.Add(ResourceNames.LanguageConstant(1000));

    // Create forecast ad group based on themes such as creative relevance, product category,
    // or cost per click.
    ForecastAdGroup forecastAdGroup = new ForecastAdGroup();

    forecastAdGroup.BiddableKeywords.Add(new BiddableKeyword()
    {
        MaxCpcBidMicros = 2_500_000,
        Keyword = new KeywordInfo()
        {
            Text = "mars cruise",
            MatchType = KeywordMatchType.Broad
        }
    });

    forecastAdGroup.BiddableKeywords.Add(new BiddableKeyword()
    {
        MaxCpcBidMicros = 1_500_000,
        Keyword = new KeywordInfo()
        {
            Text = "cheap cruise",
            MatchType = KeywordMatchType.Phrase
        }
    });

    forecastAdGroup.BiddableKeywords.Add(new BiddableKeyword()
    {
        MaxCpcBidMicros = 1_990_000,
        Keyword = new KeywordInfo()
        {
            Text = "jupiter cruise",
            MatchType = KeywordMatchType.Broad
        }
    });

    forecastAdGroup.NegativeKeywords.Add(new KeywordInfo()
    {
        Text = "moon walk",
        MatchType = KeywordMatchType.Broad
    });

    campaignToForecast.AdGroups.Add(forecastAdGroup);

    return campaignToForecast;
}
GenerateForecastMetrics.cs
      

public static function runExample(
    GoogleAdsClient $googleAdsClient,
    int $customerId
): void {
    $campaignToForecast = self::createCampaignToForecast();
    $keywordPlanIdeaServiceClient = $googleAdsClient->getKeywordPlanIdeaServiceClient();
    // Generates keyword forecast metrics based on the specified parameters.
    $response = $keywordPlanIdeaServiceClient->generateKeywordForecastMetrics(
        new GenerateKeywordForecastMetricsRequest([
            'customer_id' => $customerId,
            'campaign' => $campaignToForecast,
            'forecast_period' => new DateRange([
                // Sets the forecast start date to tomorrow.
                'start_date' => date('Ymd', strtotime('+1 day')),
                // Sets the forecast end date to 30 days from today.
                'end_date' => date('Ymd', strtotime('+30 days'))
            ])
        ])
    );

    $metrics = $response->getCampaignForecastMetrics();
    printf(
        "Estimated daily clicks: %s%s",
        $metrics->hasClicks() ? sprintf("%.2f", $metrics->getClicks()) : "'none'",
        PHP_EOL
    );
    printf(
        "Estimated daily impressions: %s%s",
        $metrics->hasImpressions() ? sprintf("%.2f", $metrics->getImpressions()) : "'none'",
        PHP_EOL
    );
    printf(
        "Estimated average CPC (micros): %s%s",
        $metrics->hasAverageCpcMicros()
            ? sprintf("%d", $metrics->getAverageCpcMicros()) : "'none'",
        PHP_EOL
    );
}

/**
 * Creates the campaign to forecast. A campaign to forecast lets you try out various
 * configurations and keywords to find the best optimization for your future campaigns. Once
 * you've found the best campaign configuration, create a serving campaign in your Google Ads
 * account with similar values and keywords. For more details, see:
 *
 * https://support.google.com/google-ads/answer/3022575
 *
 * @return CampaignToForecast the created campaign to forecast
 */
private static function createCampaignToForecast(): CampaignToForecast
{
    // Creates a campaign to forecast.
    $campaignToForecast = new CampaignToForecast([
        'keyword_plan_network' => KeywordPlanNetwork::GOOGLE_SEARCH,
        'bidding_strategy' => new CampaignBiddingStrategy([
            'manual_cpc_bidding_strategy' => new ManualCpcBiddingStrategy([
                'max_cpc_bid_micros' => 1_000_000
            ])
        ]),
        // See https://developers.google.com/google-ads/api/reference/data/geotargets for the
        // list of geo target IDs.
        'geo_modifiers' => [
            new CriterionBidModifier([
                // Geo target constant 2840 is for USA.
                'geo_target_constant' => ResourceNames::forGeoTargetConstant(2840)
            ])
        ],
        // See
        // https://developers.google.com/google-ads/api/reference/data/codes-formats#languages
        // for the list of language criteria IDs. Language constant 1000 is for English.
        'language_constants' => [ResourceNames::forLanguageConstant(1000)],
    ]);

    // Creates forecast ad group based on themes such as creative relevance, product category,
    // or cost per click.
    $forecastAdGroup = new ForecastAdGroup([
        'biddable_keywords' => [
            new BiddableKeyword([
                'max_cpc_bid_micros' => 2_500_000,
                'keyword' => new KeywordInfo([
                    'text' => 'mars cruise',
                    'match_type' => KeywordMatchType::BROAD
                ])
            ]),
            new BiddableKeyword([
                'max_cpc_bid_micros' => 1_500_000,
                'keyword' => new KeywordInfo([
                    'text' => 'cheap cruise',
                    'match_type' => KeywordMatchType::PHRASE
                ])
            ]),
            new BiddableKeyword([
                'max_cpc_bid_micros' => 1_990_000,
                'keyword' => new KeywordInfo([
                    'text' => 'jupiter cruise',
                    'match_type' => KeywordMatchType::BROAD
                ])
            ])
        ],
        'negative_keywords' => [
            new KeywordInfo([
                'text' => 'moon walk',
                'match_type' => KeywordMatchType::BROAD
            ])
        ]
    ]);
    $campaignToForecast->setAdGroups([$forecastAdGroup]);

    return $campaignToForecast;
}
GenerateForecastMetrics.php
      

def main(client, customer_id):
    """The main method that creates all necessary entities for the example.

    Args:
        client: an initialized GoogleAdsClient instance.
        customer_id: a client customer ID.
    """
    campaign_to_forecast = create_campaign_to_forecast(client)
    generate_forecast_metrics(client, customer_id, campaign_to_forecast)


def create_campaign_to_forecast(client):
    """Creates the campaign to forecast.

    A campaign to forecast lets you try out various configurations and keywords
    to find the best optimization for your future campaigns. Once you've found
    the best campaign configuration, create a serving campaign in your Google
    Ads account with similar values and keywords. For more details, see:
    https://support.google.com/google-ads/answer/3022575

    Args:
        client: an initialized GoogleAdsClient instance.

    Returns:
        An CampaignToForecast instance.
    """
    googleads_service = client.get_service("GoogleAdsService")
    # Create a campaign to forecast.
    campaign_to_forecast = client.get_type("CampaignToForecast")
    campaign_to_forecast.keyword_plan_network = (
        client.enums.KeywordPlanNetworkEnum.GOOGLE_SEARCH
    )

    # Set the bidding strategy.
    campaign_to_forecast.bidding_strategy.manual_cpc_bidding_strategy.max_cpc_bid_micros = (
        1000000
    )

    # For the list of geo target IDs, see:
    # https://developers.google.com/google-ads/api/reference/data/geotargets
    criterion_bid_modifier = client.get_type("CriterionBidModifier")
    # Geo target constant 2840 is for USA.
    criterion_bid_modifier.geo_target_constant = (
        googleads_service.geo_target_constant_path("2840")
    )
    campaign_to_forecast.geo_modifiers.append(criterion_bid_modifier)

    # For the list of language criteria IDs, see:
    # https://developers.google.com/google-ads/api/reference/data/codes-formats#languages
    # Language criteria 1000 is for English.
    campaign_to_forecast.language_constants.append(
        googleads_service.language_constant_path("1000")
    )

    # Create forecast ad groups based on themes such as creative relevance,
    # product category, or cost per click.
    forecast_ad_group = client.get_type("ForecastAdGroup")

    # Create and configure three BiddableKeyword instances.
    biddable_keyword_1 = client.get_type("BiddableKeyword")
    biddable_keyword_1.max_cpc_bid_micros = 2500000
    biddable_keyword_1.keyword.text = "mars cruise"
    biddable_keyword_1.keyword.match_type = (
        client.enums.KeywordMatchTypeEnum.BROAD
    )

    biddable_keyword_2 = client.get_type("BiddableKeyword")
    biddable_keyword_2.max_cpc_bid_micros = 1500000
    biddable_keyword_2.keyword.text = "cheap cruise"
    biddable_keyword_2.keyword.match_type = (
        client.enums.KeywordMatchTypeEnum.PHRASE
    )

    biddable_keyword_3 = client.get_type("BiddableKeyword")
    biddable_keyword_3.max_cpc_bid_micros = 1990000
    biddable_keyword_3.keyword.text = "cheap cruise"
    biddable_keyword_3.keyword.match_type = (
        client.enums.KeywordMatchTypeEnum.EXACT
    )

    # Add the biddable keywords to the forecast ad group.
    forecast_ad_group.biddable_keywords.extend(
        [biddable_keyword_1, biddable_keyword_2, biddable_keyword_3]
    )

    # Create and configure a negative keyword, then add it to the forecast ad
    # group.
    negative_keyword = client.get_type("KeywordInfo")
    negative_keyword.text = "moon walk"
    negative_keyword.match_type = client.enums.KeywordMatchTypeEnum.BROAD
    forecast_ad_group.negative_keywords.append(negative_keyword)

    campaign_to_forecast.ad_groups.append(forecast_ad_group)

    return campaign_to_forecast


def generate_forecast_metrics(client, customer_id, campaign_to_forecast):
    """Generates forecast metrics and prints the results.

    Args:
        client: an initialized GoogleAdsClient instance.
        customer_id: a client customer ID.
        campaign_to_forecast: a CampaignToForecast to generate metrics for.
    """
    keyword_plan_idea_service = client.get_service("KeywordPlanIdeaService")
    request = client.get_type("GenerateKeywordForecastMetricsRequest")
    request.customer_id = customer_id
    request.campaign = campaign_to_forecast
    # Set the forecast range. Repeat forecasts with different horizons to get a
    # holistic picture.
    # Set the forecast start date to tomorrow.
    tomorrow = datetime.now() + timedelta(days=1)
    request.forecast_period.start_date = tomorrow.strftime("%Y-%m-%d")
    # Set the forecast end date to 30 days from today.
    thirty_days_from_now = datetime.now() + timedelta(days=30)
    request.forecast_period.end_date = thirty_days_from_now.strftime("%Y-%m-%d")

    response = keyword_plan_idea_service.generate_keyword_forecast_metrics(
        request=request
    )

    metrics = response.campaign_forecast_metrics
    print(f"Estimated daily clicks: {metrics.clicks}")
    print(f"Estimated daily impressions: {metrics.impressions}")
    print(f"Estimated daily average CPC: {metrics.average_cpc_micros}")
generate_forecast_metrics.py
      

This example is not yet available in Ruby; you can take a look at the other languages.
    

sub generate_forecast_metrics {
  my ($api_client, $customer_id) = @_;

  my $campaign_to_forecast = create_campaign_to_forecast();

  my $keyword_forecast_metrics_response =
    $api_client->KeywordPlanIdeaService()->generate_keyword_forecast_metrics({
      customerId => $customer_id,
      campaign   => $campaign_to_forecast,
      # Set the forecast range. Repeat forecasts with different horizons
      # to get a holistic picture.
      forecastPeriod => Google::Ads::GoogleAds::V18::Common::DateRange->new({
          # Set the forecast start date to tomorrow.
          startDate => strftime("%Y-%m-%d", localtime(time + 60 * 60 * 24)),
          # Set the forecast end date to 30 days from today.
          endDate => strftime("%Y-%m-%d", localtime(time + 60 * 60 * 24 * 30))})
    });

  my $metrics = $keyword_forecast_metrics_response->{campaignForecastMetrics};

  printf "Estimated daily clicks: %s.\n",
    defined $metrics->{clicks} ? $metrics->{clicks} : "undef";
  printf "Estimated daily impressions: %s.\n",
    defined $metrics->{impressions} ? $metrics->{impressions} : "undef";
  printf "Estimated average cpc (micros): %s.\n\n",
    defined $metrics->{averageCpcMicros}
    ? $metrics->{averageCpcMicros}
    : "undef";

  return 1;
}

# Creates the campaign to forecast. A campaign to forecast lets you try out
# various configuration and keywords to find the best optimization for your
# future campaigns. Once you've found the best campaign configuration,
# create a serving campaign in your Google Ads account with similar values
# and keywords. For more details, see:
# https://support.google.com/google-ads/answer/3022575
sub create_campaign_to_forecast {
  my ($api_client) = @_;

  # Create a campaign to forecast.
  my $campaign_to_forecast =
    Google::Ads::GoogleAds::V18::Services::KeywordPlanIdeaService::CampaignToForecast
    ->new({keywordPlanNetwork => 'GOOGLE_SEARCH'});

  # Set the bidding strategy.
  $campaign_to_forecast->{biddingStrategy}->{manualCpcBiddingStrategy} =
    Google::Ads::GoogleAds::V18::Services::KeywordPlanIdeaService::ManualCpcBiddingStrategy
    ->new({maxCpcBidMicros => 1000000});

  # See https://developers.google.com/google-ads/api/reference/data/geotargets
  # for the list of geo target IDs.
  $campaign_to_forecast->{geoModifiers} = [
    Google::Ads::GoogleAds::V18::Services::KeywordPlanIdeaService::CriterionBidModifier
      ->new({
        # Geo target constant 2840 is for USA.
        geoTargetConstant =>
          Google::Ads::GoogleAds::V18::Utils::ResourceNames::geo_target_constant(
          2840)})];

  # See https://developers.google.com/google-ads/api/reference/data/codes-formats#languages
  # for the list of language criteria IDs.
  $campaign_to_forecast->{languageConstants} = [
    # Language criteria 1000 is for English.
    Google::Ads::GoogleAds::V18::Utils::ResourceNames::language_constant(1000)];

  # Create forecast ad groups based on themes such as creative relevance,
  # product category, or cost per click.
  $campaign_to_forecast->{adGroups} = [
    Google::Ads::GoogleAds::V18::Services::KeywordPlanIdeaService::ForecastAdGroup
      ->new({
        biddableKeywords => [
          Google::Ads::GoogleAds::V18::Services::KeywordPlanIdeaService::BiddableKeyword
            ->new({
              maxCpcBidMicros => 2500000,
              keyword => Google::Ads::GoogleAds::V18::Common::KeywordInfo->new({
                  text      => "mars cruise",
                  matchType => 'BROAD'
                })}
            ),
          Google::Ads::GoogleAds::V18::Services::KeywordPlanIdeaService::BiddableKeyword
            ->new({
              maxCpcBidMicros => 1500000,
              keyword => Google::Ads::GoogleAds::V18::Common::KeywordInfo->new({
                  text      => "cheap cruise",
                  matchType => 'PHRASE'
                })}
            ),
          Google::Ads::GoogleAds::V18::Services::KeywordPlanIdeaService::BiddableKeyword
            ->new({
              maxCpcBidMicros => 1990000,
              keyword => Google::Ads::GoogleAds::V18::Common::KeywordInfo->new({
                  text      => "jupiter cruise",
                  matchType => 'EXACT'
                })})
        ],
        negativeKeywords => [
          Google::Ads::GoogleAds::V18::Common::KeywordInfo->new({
              text      => "moon walk",
              matchType => 'BROAD'
            })]})];

  return $campaign_to_forecast;
}
generate_forecast_metrics.pl
      

Mapear para a interface

KeywordPlanIdeaService.GenerateKeywordForecastMetrics tem uma funcionalidade semelhante no Planejador de palavras-chave da interface do Google Ads.