How to Import Data from the Linkedin Ads Source
Try it freeBefore proceeding, please make sure that:
- You have created a refresh token (as described in CREDENTIALS) and securely saved your Client ID and Primary Client Secret.
- You have set up OWOX Data Marts and created at least one storage in the Storages section.

Create the Data Mart
Section titled “Create the Data Mart”- Click New Data Mart.
- Enter a title and select the Storage.
- Click Create Data Mart.

Set Up the Connector
Section titled “Set Up the Connector”- Select Connector as the input source type.
- Click Set up connector and choose LinkedIn Ads.
- Fill in the required fields:
- Credential – select the Credential with your Client ID, Primary Client Secret and Refresh Token (see the CREDENTIALS tutorial), or click + Create new Credential.
- Account URNs – you can find this value on your LinkedIn Campaign Manager.
- Leave the other fields as default and proceed to the next step.



Configure Data Import
Section titled “Configure Data Import”- Choose one of the available endpoints.
- Select the required fields.
- Specify the dataset where the data will be stored (or leave the default).
- Click Save, then Publish & Run Data Mart. Publishing starts the first import automatically.

Run the Data Mart
Section titled “Run the Data Mart”Publish & Run Data Mart already started the first import. To import data again, you have two options:
Option 1: Import Current Day’s Data
Choose Manual run → Incremental load to load data for the current day.


ℹ️ If you click Incremental load again after a successful initial load,
the connector will import: Current day’s data, plus Additional days, based on the value in the Reimport Lookback Window field.

Option 2: Manual Backfill for Specific Date Range
Choose Backfill (custom period) to load historical data.
- Select the Start Date and End Date.
- Click the Run button.

The process is complete when the Run history tab shows the message:
“Success”

How do I see spend by country, region, or county?
Section titled “How do I see spend by country, region, or county?”Choose one of these endpoints:
- Ad Analytics by Country – table
linkedin_ads_ad_analytics_by_country - Ad Analytics by Region – table
linkedin_ads_ad_analytics_by_region - Ad Analytics by County – table
linkedin_ads_ad_analytics_by_county
Each row holds one campaign’s metrics for one location on one day. The location comes as a LinkedIn URN (for example urn:li:geo:103644278) and as its English name (for example United States). The campaign column holds the campaign URN; its number is the id in the Campaigns endpoint.
LinkedIn reports one location level at a time, so country, region, and county are separate endpoints and separate tables.
What to expect from these endpoints:
- More API requests than Ad Analytics. LinkedIn breaks location down only for one campaign at a time. For every day, the connector first reads which campaigns ran, then asks for each of those campaigns separately. LinkedIn returns up to 18 metrics per request, so every further 18 selected metrics add one more request per campaign.
- Clicks and spend by location are incomplete. LinkedIn treats location as a member demographic and describes such metrics as approximate. Impressions by location can add up to nearly all of the campaign’s impressions in Ad Analytics, while clicks and spend by location add up to only part of the campaign’s. Spend by location can be well below the campaign’s spend, even for a campaign that runs in a single country, and a campaign spread over many locations can show no clicks or spend at all. To protect member privacy, LinkedIn also drops a location with fewer than 3 events, returns at most the top 100 locations for each ad on each day, and returns location data only for ads that reached enough members. Rows LinkedIn returns without a location or a date are not imported, and the run says how many.
- Data arrives 12–24 hours late. The default Reimport Lookback Window of 2 days re-imports those days on the next run.
- Two years of history. LinkedIn keeps location breakdowns for two years, so a backfill further back returns no rows.
- Some metrics are missing. LinkedIn does not break down
approximateMemberReach,audiencePenetration,conversionValueInLocalCurrency, or the carousel card metrics by location, so these endpoints do not offer them.
To estimate spend by location, split each campaign’s daily spend from Ad Analytics by each location’s share of that campaign’s impressions. The estimate is only as good as the impressions by location, so first check that they add up to nearly all of the campaign’s impressions in Ad Analytics.
In Ad Analytics, pivotValues holds the creative, campaign, campaign group, and account URNs in that order, so the campaign URN is its second value. It is stored as text: a JSON array, or a comma-separated list in AWS Redshift. For example, in BigQuery:
WITH by_country AS ( SELECT dateRangeStart, campaign, countryName, impressions FROM `your_dataset.linkedin_ads_ad_analytics_by_country`),campaign_spend AS ( SELECT dateRangeStart, JSON_VALUE(pivotValues, '$[1]') AS campaign, SUM(costInUsd) AS costInUsd FROM `your_dataset.linkedin_ads_ad_analytics` GROUP BY dateRangeStart, campaign)SELECT c.dateRangeStart, c.campaign, c.countryName, c.impressions, s.costInUsd * SAFE_DIVIDE(c.impressions, SUM(c.impressions) OVER (PARTITION BY c.dateRangeStart, c.campaign)) AS estimatedCostInUsdFROM by_country AS cJOIN campaign_spend AS s USING (dateRangeStart, campaign)The query leaves out a campaign’s day that only one of the two tables has, so import both endpoints for the same days. On a day when the campaign’s locations have no impressions, the estimate stays empty, because there is nothing to split the spend by.
Location names:
- If LinkedIn returns no name for a location, its name column stays empty, and the run finishes with a Warning status that lists those locations.
- If LinkedIn does not allow the name lookup for your app, the run imports the metrics and the location URNs, leaves every name empty, and finishes with a Warning status.
- If LinkedIn rejects the lookup for some locations, only those names stay empty, and the Warning lists them.
- If the name lookup fails for a while (for example LinkedIn is overloaded), the affected days are not imported and the run fails, so the next run imports them again with their names.
Why does my Ad Analytics run finish with a Warning status?
Section titled “Why does my Ad Analytics run finish with a Warning status?”LinkedIn’s adAnalytics endpoint does not support pagination and caps each response at 15,000 rows. To avoid silently losing data on long date ranges, the connector fetches analytics one day at a time. A single day can still exceed the limit, for example an account with more than 15,000 active creatives. The run then finishes with a Warning status and lists the affected days in the log. Data for those days may be incomplete.
Access Your Data
Section titled “Access Your Data”Once the run is complete, the data will be written to the dataset you specified earlier.

If you encounter any issues:
- Check the Run history for specific error messages
- Please visit Q&A first
- If you want to report a bug, please open an issue
- Join the discussion forum to ask questions or propose improvements