Discrepancies in spend data when using segmentation

[BuxApi]

Hi All,

I am using the twitter_ads library for Python to download spend data on campaign level, and noticed that there are some discrepancies in the data.
It would be great if you could help us out with this issue:

Data used for retrieval:

•   Account number: 18ce53wctv2
  

•   Campaign id: b2o9e
  

•   Date: 2018-07-01
  

I ran two asynchronous queries, one with segmentation by locations and one without and they resulted in different numbers (the one without segmentation resulted in numbers matching Twitter Ads portal).
With segmentation:

Campaign.queue_async_stats_job(tw_account,
                              [‘b2o9e’],
                              [METRIC_GROUP.ENGAGEMENT, METRIC_GROUP.BILLING],
                              **{'granularity':'DAY',
                                 'start_time': d,
                                 'end_time': (d + timedelta(days = 1)),
                                 'segmentation_type': 'LOCATIONS',
                                 'placement': 'ALL_ON_TWITTER'})

With d being the date.

This resulted in:

{'data': [{'id': 'b2o9e',
  'id_data': [{'metrics': {'app_clicks': [40],
     'billed_charge_local_micro': [7880196],
     'billed_engagements': [39],
     'card_engagements': [40],
     'carousel_swipes': None,
     'clicks': [32],
     'engagements': [118],
     'follows': [1],
     'impressions': [3352],
     'likes': [4],
     'poll_card_vote': None,
     'qualified_impressions': None,
     'replies': None,
     'retweets': None,
     'tweets_send': None,
     'unfollows': None,
     'url_clicks': None},
    'segment': {'segment_name': 'Austria',
     'segment_value': '30410557050f13a5'}},
   {'metrics': {'app_clicks': [372],
     'billed_charge_local_micro': [102828051],
     'billed_engagements': [358],
     'card_engagements': [372],
     'carousel_swipes': None,
     'clicks': [171],
     'engagements': [926],
     'follows': None,
     'impressions': [24547],
     'likes': [8],
     'poll_card_vote': None,
     'qualified_impressions': None,
     'replies': None,
     'retweets': None,
     'tweets_send': None,
     'unfollows': None,
     'url_clicks': None},
    'segment': {'segment_name': 'Germany',
     'segment_value': 'fdcd221ac44fa326'}}]}],
'data_type': 'stats',
'request': {'params': {'country': None,
  'end_time': '2018-07-01T22:00:00Z',
  'entity': 'CAMPAIGN',
  'entity_ids': ['b2o9e'],
  'granularity': 'DAY',
  'metric_groups': ['ENGAGEMENT', 'BILLING'],
  'placement': 'ALL_ON_TWITTER',
  'platform': None,
  'segmentation_type': 'LOCATIONS',
  'start_time': '2018-06-30T22:00:00Z'}},
'time_series_length': 1}

Without segmentation:

Campaign.queue_async_stats_job(tw_account,
                               [‘b2o9e’],
                               [METRIC_GROUP.ENGAGEMENT, METRIC_GROUP.BILLING],
                               **{'granularity':'DAY',
                                  'start_time': d,
                                  'end_time': (d + timedelta(days = 1)),
                                  'placement': 'ALL_ON_TWITTER'})

With the result

{'data': [{'id': 'b2o9e',
  'id_data': [{'metrics': {'app_clicks': [412],
     'billed_charge_local_micro': [110989197],
     'billed_engagements': [399],
     'card_engagements': [412],
     'carousel_swipes': None,
     'clicks': [203],
     'engagements': [1044],
     'follows': [1],
     'impressions': [27899],
     'likes': [12],
     'poll_card_vote': None,
     'qualified_impressions': None,
     'replies': None,
     'retweets': None,
     'tweets_send': None,
     'unfollows': None,
     'url_clicks': None},
    'segment': None}]}],
'data_type': 'stats',
'request': {'params': {'country': None,
  'end_time': '2018-07-01T22:00:00Z',
  'entity': 'CAMPAIGN',
  'entity_ids': ['b2o9e'],
  'granularity': 'DAY',
  'metric_groups': ['ENGAGEMENT', 'BILLING'],
  'placement': 'ALL_ON_TWITTER',
  'platform': None,
  'segmentation_type': None,
  'start_time': '2018-06-30T22:00:00Z'}},
'time_series_length': 1}

Any help would be appreciated!

[imit8me]

Hey @BuxApi

Thanks for reaching out. Could you point out the specific discrepancies you’ve noticed? Additionally, what is the expected vs actual outpout of the request?

Thanks!

[BuxApi]

Hey @imit8me

So using segmentation total billed_charge_local_micro for the campaign is 110708247, while without it is 110989197. The expected output is actually the value without segmentation, as this is in line with Twitter Ads portal. However, I am wondering why the API returns inconsistent values when I split the spend per country.

Cheers!

[imit8me]

Thanks for the details @BuxApi . Please note that

Segmented data is not expected to roll-up 100% to the non-segmented data, due to how this information is derived.

Source: Analytics Best Practices

Hope this helps!