Question 1

What can a ClickHouse model send to Facebook Ads?

Accepted Answer

Custom audiences and offline conversion events. Rows returned by the model become audience members, matched on hashed email or phone, and conversions let you report outcomes that happened away from the platform. Both come from ordinary models, so one behavioural table can drive targeting and measurement without maintaining two separate definitions.

Question 2

How do I connect if the cluster is not publicly reachable?

Accepted Answer

Through an SSH tunnel, which is supported natively. You supply a bastion host, port, user and private key, and the connection is made through it. For directly reachable clusters you connect on the native port with TLS, normally 9440, and certificate verification is configurable if you are using an internal certificate authority. ClickHouse Cloud has no separate authentication path, it is the same native endpoint with your ClickHouse credentials.

Question 3

Which credentials does the read actually use?

Accepted Answer

A ClickHouse database username and password, with SELECT on whatever the model touches. This trips people up because the connection form also carries AWS, Azure or Google Cloud credentials, and those are for destination staging rather than source authentication. If you are only reading from ClickHouse to push audiences out, the storage credentials are not part of the picture at all and the account needs nothing beyond read access.

Question 4

Does the audience shrink when people stop qualifying?

Accepted Answer

With mirror mode, yes. Add mode only adds and remove mode only removes, but mirror reconciles the audience against the model on each run. Meta supports it, which not every platform does. Bear in mind that the incremental read itself cannot see deletions in ClickHouse, so if your model relies on rows disappearing rather than being marked, mirror mode will not have the information it needs.

Question 5

Why model the audience rather than exporting one?

Accepted Answer

Because the segments worth targeting are computed from behaviour that changes constantly. Your most engaged users, or the cohort that churned after a particular pattern, is a query result rather than a fixed list. Reverse ETL reruns that query on a schedule and keeps Meta in step, whereas an export is accurate the day it is made and quietly wrong a fortnight later.

Sync ClickHouse to Facebook Ads

Sync ClickHouse to Facebook Ads with Polytomic

Audiences and conversions

TLS or SSH tunnel

Hashed before sending

Or sync Facebook Ads to ClickHouse

Trusted by businesses worldwide

Step-by-step integration guide

Connect ClickHouse in Polytomic

Connect Facebook Ads in Polytomic

Create a sync between ClickHouse and Facebook Ads

Map fields and enable the sync

Frequently asked questions

The sync and ETL platform that does it all

Powerful data sync

Advanced analytics

Enterprise security

Easy setup

Ready to connect ClickHouse with Facebook Ads?