CONNECTIONS / Snowflake / Snowflake to TikTok Custom Audience
Snowflake's segment table, mirrored into TikTok Custom Audience.
Signals keeps watch on whichever table or view holds your segment, hashes every identifier as it uploads, and sends TikTok only the members who joined or left since the last run, not a full resend of the audience.
- ● snowflake → tiktok.custom_audience :: live
- > sync table VIP_SEGMENT rows=14,920
- hash sha256(email,phone)
- route tiktok.custom_audience upsert
- ✓ audience refreshed · 11,624 matched
WHAT THIS ENABLES
Snowflake to TikTok Custom Audience
Snowflake to TikTok Custom Audience keeps TikTok's Custom Audiences synced with a segment table your data team already maintains, adding and removing members automatically so suppression and lookalike campaigns reflect the warehouse's current view of your customers.
- Segment tables in Snowflake mirrored into TikTok Ads Manager with no export step in between.
- Lookalike seeds built from the rows your dbt models already rank as best.
- Renewal and expansion audiences that update the moment the underlying table refreshes.
WHAT FLOWS WHERE
Snowflake to Custom Audience, mapped.
A scheduled Snowflake read picks up changes to the nominated table, hashes every identifier, and pushes the resulting adds and removals to the TikTok Custom Audience.
A scheduled Snowflake read picks up changes to the nominated table, hashes every identifier, and pushes the resulting adds and removals to the TikTok Custom Audience.
Built for the teams that own the number.
Teams manually re-exporting warehouse segments to TikTok Ads Manager.
From kickoff to verified events.
-
Connect
A read-only Snowflake role scoped to the table or view that defines the segment.
-
Map
Each identifier column shaped to fit the upload format TikTok's Custom Audience endpoint expects.
-
Deliver
Incremental syncs to TikTok on the schedule the table refreshes, with matched-size reporting per audience.
What ships with this use case.
Segment logic that never leaves Snowflake, only its hashed output, serving directly inside TikTok targeting.
Each sync moving only the delta, so the full segment is never re-sent just to keep the audience current.
A single warehouse object driving every connected ad platform's audience, TikTok included.
What Custom Audience actually receives.
- ● signals :: audience upload
- > POST /dmp/custom_audience source=snowflake.VIP_SEGMENT
- id_schema ["EMAIL_SHA256","PHONE_SHA256"]
- action "add" · audience_id "snow_vip_9042"
- id_list_size 14,920 · matched 11,624
- ✓ accepted match=78%
Adjacent moves on the same stack.
from Snowflake
Snowflake to Meta Custom Audience
Snowflake to Meta Custom Audience keeps Meta's Custom Audiences built from whichever table…
Learn more
from Snowflake
Snowflake to Google Customer Match
Snowflake to Google Customer Match reads customer segments straight from your Snowflake…
Learn more
same use case
ActiveCampaign to TikTok Custom Audience
ActiveCampaign to TikTok Custom Audience: Signals reads your ActiveCampaign contacts and…
Learn moregeneric
TikTok Custom Audience
Datahash syncs hashed customer audiences from your CRM, warehouse, or files to TikTok for…
Learn moreAsked on almost every call.
Does this work with a Snowflake view, or does it need a physical table?
Either. Signals reads whatever object your team already uses to define the segment, a materialized table or a view built on a dbt model, so there is no requirement to rebuild your segmentation into a new format just to sync it to TikTok.
What refresh schedule keeps the TikTok audience honest?
Whatever cadence the underlying warehouse task already runs on, most commonly a daily rebuild. Because each cycle only ships the rows added or dropped since the previous one, tightening the schedule to hourly does not turn into a heavier upload, it just narrows the delta.
Does one segment table have to serve TikTok exclusively, or can it fan out further?
It can fan out. The same table defined once in Snowflake can drive audiences on TikTok, Meta, Google, and Snapchat at the same time, each with its own identifier mapping and refresh cadence, so the segmentation logic is written once rather than per ad platform.
NEXT STEP