Skip to main content
Docs
Feature Requests Onboarding Trust Center Status Page

Community

Connector Improvement: Optimizely Connector Refactor to Protect Outrageous MAR impact to customers

Completed
Andrew Scarani User

The optimizely connector schema is outrageous in the way that it syncs data through Fivetran and racks-up an MAR bill... That is because it syncs attributes one-at-a-time into a separate table, rather than syncing a collection of event attributes/properties with each event in Optimizely. Therefore, each individual attribute of an event counts towards MAR. At the rate that we currently use our Optimizely Fivetran connector... we may look at just turning off this connector all together because of how egregious this is. 

If the Fivetran connector could join these pieces together and log each event as a single entity to the downstream system, this would be much more cost-effective for end-users, and therefore more users would end up using this connector.

Please sign in to leave a comment.

Comments

2 comments

Date Votes
  • Official comment
    Drew Fivetranner

    Hi Andrew Scarani, Drew from the product team here!

    Thanks for this feedback, we are always looking for ways to improve the usability of our schemas.

    I'll reach out to you to learn more about your needs and how you would expect to receive event data from the Optimizely connector.

  • Drew Fivetranner

    Hi All, Drew from the product team here!

    Wanted to leave an update on this as we are rolling out some changes.

    Fivetran is rolling out support for a new setup option on the Optimizely connector which allows our users to select a "packed" or "unpacked" mode. In "unpacked" mode, the connector's behavior is unchanged.

    In "packed" mode, the connector will keep nested data from Optimizely as columns in the CONVERSION and DECISION tables, rather than splitting the data out into separate tables. That means that the CONVERSION_ATTRIBUTE, CONVERSION_TAG, CONVERSION_EXPERIMENT, and DECISION_ATTRIBUTE tables will no longer sync data (as that data will be nested in the parent tables).

    This change was made to address the MAR implications of the "unpacked" sync mode. The changes will first be rolled out to all new connectors, then turned on for existing connectors. If you are using an existing connector, it will continue to function in "unpacked" mode. You will be able to switch modes on the setup form and enable "packed" mode at any time.