Computations are some form of permutation on top of attributes coming from your systems as well as MadKudu's enrichment database. 

Computations could be:

• Directly taken from your CRM attributes 
• Combination of your CRM attributes and MadKudu enrichment attributes
• Combination of MadKudu enrichment attributes

All MadKudu standard enrichment attributes are by default a computation. 

However, you can also create your own computation. There are several modes to this and I will explain each use case according to the different modes. 

1) One-to-one

Let's go through an example of why building a "one-to-one" computation would be necessary. If you want to pull a custom attribute from your CRM which is basically a form input for "industry" because this is the most accurate information as compared to enrichment, you would typically create a one-to-one computation.

Here's a step by step guide to create a computation in the Data Science Studio.

1. First, you need to go on the app > click on the Integrations tab > navigate to the CRM you're using (say Salesforce or HubSpot) and click on Pull. Here, you can select which object to pull fields from.
2. Once you've pulled the field (this takes about a couple of hours), you can then proceed to the next step. 
3. Second, navigate to the Mapping tab > click on Attribute Mapping.
4. Create the attribute by selecting the: (1) source system (Salesforce or HubSpot), corresponding object and field name. 
5. Once the processing has been completed (takes up to 12 hours), you can now proceed to the next and third step.
6. Navigate to the Data Science Studio and select any model (computations are model-agnostic). Once you land on there, click on Customer Profile and Computations. Here is where you are able to create a new computation by clicking on "Create new computation".
7. Select the type of computation
   1. Numeric: output is a number of boolean (true or false)
   2. String: output is text
8. Configure the computation
   1. Select the mode of computation: one to one
   2. Select the attribute recently created in Step #4.
9. If this is computation is solely based on MadKudu attributes, select "Configured". Otherwise, select "Custom (batch)" under tags for contact-level attributes or "Custom (realtime)" for account-level attributes.
10. Always select "Customer Fit Insights" under tags.
11. Save the computation.
12. Click on Release and Recompute when you are back on the computations homepage.
13. Once you received a success email from engineering@madkudu.com (this typically takes up to 2 hours or more depending on volume data), the computation has been created successfully.

2) Conditional 

Let's go through an example of why building a "conditional" computation would be necessary. Say you want to create a multi-conditional computation such as whether leads or prospects are part of a specific segmentation that meets your targeting criteria. For example, a lead belonging to a buyer job title in an Ideal Customer Profile (ICP) account that is hiring more than 5 roles at this time. 

Here's a step by step guide to create a computation in the Data Science Studio.

1. If the computation requires a custom attribute from your systems, please follow steps #1 to #5 in the "one to one" computation setup guide. If not, please proceed to the next.
2. Navigate to the Data Science Studio and select any model (computations are model-agnostic). Once you land on there, click on Customer Profile and Computations. Here is where you are able to create a new computation by clicking on "Create new computation".
3. Select the type of computation
   1. Numeric: output is a number of boolean (true or false)
   2. String: output is text
4. Configure the computation
   1. Select the mode of computation: conditional
   2. Select the attributes from the picklist and define the output of the computation.
5. If this is computation is solely based on MadKudu attributes, select "Configured". Otherwise, select "Custom (batch)" under tags for contact-level attributes or "Custom (realtime)" for account-level attributes.
6. Always select "Customer Fit Insights" under tags.
7. Save the computation.
8. Click on Release and Recompute when you are back on the computations homepage.
9. Once you received a success email from engineering@madkudu.com (this typically takes up to 2 hours or more depending on volume data), the computation has been created successfully.

3) Advanced

Let's go through an example of why building an "advanced" computation would be necessary. If you want to create a sophisticated computation that requires a complicated SQL query, this is the mode for you. However, this requires understanding of the MadML language. Therefore, please reach out to your Customer Success Manager to enable this capability.

FAQ

1) Do I need special user permissions to be able to create a computation

You will need to have the architect user permissions. This can be provided by any admin on the platform.