Connector

Collect Facebook Ads data in Snowflake

Bring Facebook and Instagram ad performance into your Snowflake account with OWOX Data Marts: ads, creatives and insights as separate tables, reloaded on a schedule. Then join spend to your web analytics and CRM tables in SQL to see which ads brought customers.

Facebook AdsConnectedAdAd AccountAd Account UserAd Creatives
Data MartPublishedone definition
SnowflakeDateCostClicksDaily 09:00
Refreshed on a schedule. No exports, no copy-paste.

Get Started Free

Hand this to your AI

Facebook Ads reports what Facebook says it delivered, and Snowflake holds no spend to check it against revenue.

Paste this into your AI assistant. It creates a Facebook Ads Data Mart on your Snowflake storage, reloads ads, creatives and insights on a schedule, and joins spend to your web analytics and CRM.

The prompt

load Facebook Ads ad insights into Snowflake on a daily schedule and build a Data Mart with spend, clicks and CRM revenue by campaign and week, so I can see cost per customer

Endpoints

Facebook Ads Endpoints

Documentation

15 Facebook Ads endpoints you can import. Each lands in its own table in your data warehouse.

S00 · Why

Why Facebook Ads belongs next to your CRM in Snowflake

Ads Manager counts results with Meta's own attribution, so it reports what Facebook says it delivered. To check that against revenue, spend needs to sit beside website touch points and CRM deals as tables in one data warehouse, where SQL can join them.

  • Line Facebook up with Google, TikTok and Microsoft spend in one query
  • Follow a sign-up on your site to the order or deal it became
  • Query spend, reach and ROAS from your own account instead of the ad account
Facebook AdsAd spend, reach
Web analyticsVisits, sign-ups
CRMOrders, deals
Your data warehouseSnowflakeAdVisitDealJoined by idAd seen or clickedDeal months laterA SQL join links the ad to the deal.

S01 · Step 1

Step 1: Connect Facebook Ads and choose the fields

Create a Data Mart on your Snowflake data storage and pick Facebook Ads as the data source. Sign in with Facebook or paste an access token, list the ad account ids and tick the fields you want.

  • 15 endpoints and 1172 fields are available
  • Ad Insights has 100 fields, from spend and reach to conversions and ROAS
  • Give each breakdown its own Data Mart so it lands in its own table
New Data MartChoose a sourceGoogle AdsFacebook AdsTikTok AdsAuthorizeConnectedEndpointAd InsightsAdAd AccountAd Account User
Ad Insights100 fieldsdate_startspendimpressionsclicksreachpurchase_roas

S02 · Step 2

Step 2: Load Facebook Ads into a Snowflake table

Choose the Snowflake database, then the schema and table; all three are pre-filled and editable. The first run loads the previous month through today into your account.

  • Each endpoint lands in its own table, for example the ad account insights
  • Each run is logged in Run History
DestinationSnowflakeDatabaseSchemafacebook_adsTablead_account_insightsRun
Run HistoryPendingRunningSuccessIncremental load
Tablefacebook_ads.ad_account_insig…

S03 · Step 3

Step 3: Read the output schema in plain words

Once the first run finishes, OWOX shows every column with an alias and a description a colleague can follow. Your data quality checks can use it as well.

  • actions, action_values and purchase_roas arrive as JSON text with documented meaning; parse them once in a view
  • A stakeholder and an analyst see the same column names
facebook_ads.ad_account_insightsOutput schemaFieldDescriptiondate_startThe start date for your data…spendThe estimated total amount of…impressionsThe number of times your ads…clicksThe number of clicks on your ads.reachThe number of people who saw…purchase_roasThe total return on ad spend…

S04 · Step 4

Step 4: Join the Facebook Ads tables in SQL

Ad accounts, ads, creatives and insights stay separate tables that share ids. A few joins in Snowflake turn them into one Facebook Ads data model.

  • Ad Insights holds ad_id; Ads holds creative_id for the creative
  • Campaign and ad set ids come on every insights row, so you can sum spend at either level
Ad Accountnameaccount_id
Ad Creativesidname
Adsidaccount_idcreative_id
Ad Insightsad_idspend

S05 · Step 5

Step 5: Schedule the load and backfill history

Pick a cadence once and a scheduled refresh keeps the Snowflake tables current. A manual run fetches any earlier period.

  • Daily, weekly, monthly or interval runs
  • A long backfill is cut into pieces of up to 31 days for you
ScheduleRun on a scheduleDaily 09:00
BackfillRun backfill
Run HistoryDaily scheduleManual backfillDaily schedule

S06 · Step 6

Step 6: Use the same Snowflake tables in every report

The same tables serve your data model, spreadsheets, dashboards, AI assistants such as Claude and ChatGPT, and AI insights delivered to Slack or email.

  • Blend Facebook with Google and TikTok spend in one place
  • Reports stay consistent because they read the same tables
Facebook AdsData MartPublished
Blended ad spend
Google Sheets
Data Studio dashboards
Claude, ChatGPT
Slack, email

Recent updates

Why OWOX

Why OWOX for Facebook Ads in Snowflake

A Marketing API script or a manual export gets the rows out. OWOX leaves them in Snowflake tables that you own and that are ready to model.

  • Inside your Snowflake account. Your roles and grants decide who can query the tables, not an ad account login.
  • A table per breakdown. Country, region, age and gender, device, placement and product breakdowns each load separately, so you keep only the cuts you use.
  • Ready to model. Tables share ad, ad set and campaign ids, so they join to web analytics and CRM.
  • Looked after for you. When Meta changes the Marketing API, there is no script of yours to repair.
  • Free when self-hosted. Run the open-source OWOX Data Marts on your own infrastructure at no licence cost.

Book a callGet Started Free

Trust

The Data Partner You Can Trust

users

175K Source: Internal OWOX data as of 2026-11-11

Executives and data analysts across 20+ industries

G2 rating

4.9★2

“A trusted data analytics partner” — verified review

npm installs

171K3

In Open Source We Trust

Customers