Create a decision dataset in Analytics

A dataset is a tabular view of your data used for analytics. It represents business entities or activities such as users, sessions, tickets, accounts, or events. Datasets in Analytics are logical views mapped to tables or views in your data warehouse. 

What is a decision dataset?

Decision events (also known as impressions) are special events that are triggered when Optimizely Experimentation "decides" that a visitor is bucketed into a certain experiment and variation pair. See Data specification for information.

Create a decision dataset to integrate your Experimentation data. This helps you centralize and analyze your experiment results along with additional information in your data warehouse. 

1. On the navigation panel, go to Data > Datasets, click + New Dataset, and select Source Dataset.

   

2. On the Pick a source page, select your data from the connection you previously configured in the Create a connection section and click Confirm. You can use the search box to search for your tables in the warehouse. 
   

   

3. Give your dataset a name and click Save.

4. On the Semantics tab, select Decision Stream from the drop-down list.

   

5. Complete the Decision configuration by configuring the following fields:
   • Actor dataset
   • Experiment ID
   • Experiment Name
   • Variation Id
   • Is Holdback
   • Timestamp
   • Custom Partition Time Column.

     

See Create New Datasets for information.