Realtime Tailored Audience API always causes audience size of '501'

ads
api
realtime

#1

I’m trying to push different lists of several million emails into different tailored audiences.

Current pattern:

  1. create CRM/EMAIL tailored audience with POST
  2. send updates of 100 at a time to
    https://ads-api.twitter.com/2/tailored_audience_memberships
    as specified here:
    https://dev.twitter.com/ads/audiences/ta-realtime-integration-guide

with an expiration date of 3 years out, and a hash function that looks like this (no salt, sha256 on stripped lowercase email):

def hash(email):
    raw = email.strip().lower().encode('utf8')
    return hashlib.sha256(raw).hexdigest()

After thousands of API calls sending millions of records, I get this from a list_audiences:

{'audience_size': 501,
  'audience_type': u'CRM',
  'created_at': u'2017-08-02T14:55:43Z',
  'deleted': False,
  'id': 'xxxxxxxxx',
  'is_owner': True,
  'list_type': u'EMAIL',
  'name': 'Current Customers',
  'partner_source': 'OTHER',
  'permission_level': 'READ_WRITE',
  'reasons_not_targetable': [],
  'targetable': True,
  'targetable_types': ['CRM', 'EXCLUDED_CRM'],
  'updated_at': '2017-09-07T15:00:20Z'},

All my different lists also show 501 too. What’s going on?

Thanks


#2

Now I’m seeing different numbers slightly larger than 501 (a few in 2000s) with ones I created many days ago a bit higher than the ones created recently. This would suggest that the update does not actually happen in realtime or near realtime as documented, but there’s perhaps a huge backlog in the queue to be processed once the updates are submitted? Is this the case?


#3

Hi @chet,

Someone else from the team will probably chime in, but the counts for the real-time audiences returns 501 once it is created so you can use the audience right away. The system takes a while to get the accurate count, and as the documentation states: “the true estimates for audience size may take significantly longer to finalize (24-60 hours).”

Carmen


#4

+1 Thanks, @carmenjyuen!