1. Send your data to S3 following the supported format.
2. Give MadKudu access to your S3 bucket.
3. Let us know when you have added a file to your bucket so we'll validate if we can pull it.
   All the columns you plan to send to MadKudu on a recurring basis need to be present in this test file.
4. Set up a recurring dump (at least daily) of your fresh data to S3 and give us a green light to activate the recurring pull of your data. 

Once this is setup and MadKudu has access to your data, you'll be able to: 

1. Map the data pulled in MadKudu platform in the event mapping.
2. Start building a Likelihood to Buy model or aggregations based on your behavioral data.

How to enable to send data from MadKudu to S3

This section describes how to set up Amazon S3 as a destination.

How to score contacts coming from your data warehouse (or a CSV) 

1. Please follow the instructions above to enable Amazon S3 as a source first and sending to S3 the list of your contacts following the supported formats. 
2. Make sure MadKudu has write permissions (PutObject and DeleteObject) for the bucket. 
3. Contact your Account Manager at MadKudu or submit a support ticket to our support team with the following details: 
   1. Bucket path where MadKudu can find the file(s) containing the list of contacts to pull and score.
   2. Bucket path where MadKudu can add the files of scored contacts.
   3. Desired format of the file (supported format: CSV, JSON, PARQUET).
   4. The models, computations, or aggregations for scoring the contacts.

We'll take care of the rest of the set up and contact you for more information if necessary or to confirm when the setup is finalized. It should only take a few days but it all depends on volume, format respected, availability, etc. 

Once the setup is completed, you'll start receiving data in your s3 bucket at every sync process (every 4h to 12h, which means at the same time MadKudu pushes updates in your CRM or other integrations that are set up) in a different file timestamped containing the full list of contacts with the result of the model(s). Each file is independent of the other so that you have the full history of scores, not just the updates.

For example, the file could contain the following properties:

• email
• mk_customer_fit_score
• mk_customer_fit_segment
• mk_customer_fit_signals
• mk_likelihood_to_buy_score
• mk_likelihood_to_buy_segment
• mk_likelihood_to_buy_signals
• mk_lead_grade_score
• mk_lead_grade_segment
• persona (a computation named persona flagging people as buyer, influencer, etc.) 
• email_number_activities_last_90_days (an aggregation counting the activities of the person) 

How to export all the records pulled by MadKudu from your CRM, MAP, etc and scored back into S3

You'll only need to set up an S3 bucket and give MadKudu access to this S3 bucket.

1. Make sure MadKudu has write permissions (PutObject and DeleteObject) for the bucket 
2. Contact your Account Manager at MadKudu or submit a support ticket to our support team with the following details: 
   1. Bucket path where MadKudu can add the files of scored contacts
   2. Desired format of the file (supported format: CSV, JSON, PARQUET) 
   3. The models, computations, or aggregations for scoring the contacts 

F.A.Q.

What CSV parameters are sent to Amazon S3 when it is enabled as a Destination for MadKudu?

• delimiter: quote
• row delimiter: \n (line break)
• quotes: around multiline elements only (signals)