Question 1

Which Customer.io connector do I use for this direction?

Accepted Answer

Customer.io Warehouse Exports, not the main Customer.io connector. This trips people up, so it is worth being explicit: the main connector is a destination, used for pushing audiences and attributes into Customer.io. Reading Customer.io data into a data warehouse goes through Warehouse Exports, which is a separate connector with its own setup. Choosing the wrong one is the most common reason this pairing appears not to work.

Question 2

What data is available?

Accepted Answer

What Customer.io includes in its warehouse export, which covers people and their attributes, segments and memberships, campaigns and newsletters, deliveries, and the message level events such as sent, delivered, opened, clicked, bounced and unsubscribed. The event tables are usually the point of the exercise, because per message outcomes are what let you measure messaging against behaviour rather than against other messaging.

Question 3

How does incremental loading work?

Accepted Answer

It follows file arrival rather than polling an application API. The first run consumes the export and backfill files available. Later runs list files whose modification time is at or after the saved checkpoint, coalesce fragments belonging together, and discard duplicate records. The cadence you actually experience is therefore set by how often Customer.io produces exports, so plan the ETL schedule around the export frequency rather than the other way round.

Question 4

Do we need our own storage bucket, and does Redshift need another one?

Accepted Answer

Yes to both, and they serve different purposes. Customer.io writes export files to a bucket you own and control, and Polytomic reads from there, which means retention of the raw exports is your decision. Separately, every Redshift write stages through an S3 bucket in the same region as your cluster before running COPY. Keeping the export bucket and the staging bucket distinct avoids any chance of a sync reading its own output.

Question 5

Are deletions represented?

Accepted Answer

Only to the extent that the export itself represents them. This is file based rather than change based ETL, so there is no tombstone stream to rely on and a record removed in Customer.io will not necessarily be marked as removed in Redshift. If reconciliation matters, plan on periodic full reads so that rows no longer present are cleared, rather than assuming incremental file loads will handle it.

Sync Customer.io to Redshift

Sync Customer.io to Redshift with Polytomic

Warehouse Exports is the source

File arrival incremental

Messaging entities as tables

Or sync Redshift to Customer.io

Trusted by businesses worldwide

Step-by-step integration guide

Connect Customer.io in Polytomic

Connect Redshift in Polytomic

Create a sync between Customer.io and Redshift

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 Customer.io with Redshift?