You've pulled and mapped one of your CRM fields and you'd like to see its impact on conversions?

You've explored our list of computations and it gave you ideas of new data points to explore?

This article is made for you! It will guide you through the creation of a computation. 

Let's take 3 examples:

• You pulled your Salesforce field 'Number of developers' on the 'Lead' object, that your inbound leads have to fill in when they sign up to your product, and you mapped this field in the attribute mapping, calling it 'salesforce_lead_numbers_of_developers'
• You see the computation 'hg__tech' that detects various technologies, and want to use it to detect which competitor technology your leads use, such as PowerBI, Looker, or Tableau.
• You see the computation 'hg__tech' that detects various technologies, and want to use it to detect whether a lead has any integration with your product, such as GitHub, AWS, or MixPanel.

Prerequisites 

• You have access to the Data Studio
• You have the permissions of the Architect or Admin role
• You know what an Attribute is
• You know what a Computation is

Step 1: Create a computation

1. Go to the Data Studio either here: studio.madkudu.com, or through the App > Predictions > Data Studio 
2. Click on Computations then New computation
   
3. Configure the computation
   • Computations Name: use only lower case and underscore instead of space (we know it's not practical, this will be changed soon!). The computation name must be unique and different from attributes names (from the attribute mapping).  
     • Example Number of developers: Call your computation 'salesforce_lead_numbers_of_developers_comp' (the name of the attribute and the computation must be different)
     • Example Competitor: Call your computation 'competitor_tech_used'
     • Example Integration: Call your computation 'has_supported_integration'
   • Select the Type of computation
     
     
     • String: output is a text
     • Numeric: output is a number or a boolean (0 or 1, for true or false) 
       • Example Number of developers: You know the possible values for this field in your CRM are the buckets '1-10', '11-20'... So select String
       • Example Competitor: The output will be the name of the technology, so select String
       • Example Integration: The output will be a boolean (Does this lead use an integration? -> yes/no, true/false), so select Number/Boolean
   • Mode of Configuration: select among the 3 following choices
      
     • Conditional: for a computation based on the condition of other computations or attributes. In Configuration, select the rule of the computation e.g. "IF hg_tech contains Segment THEN 1" 
       
     • oneToOne: simple mapping of an attribute mapped in the attribute mapping from your custom field. In Configuration, select the field to use as a computation.
     • Advanced: for building advanced computation in SQL. This requires understanding of the MadML language. In Configuration, in the text box enter a SQL condition.
       e.g. "COALESCE (industry, industry_salesforce)" 
       "CASE WHEN (LOWER(zi_job_title) LIKE '%product%design%'  OR mk_job_title LIKE '%product%design%' ) AND LOWER(zoominfo_industry) LIKE  '%fintech%' THEN 1  ELSE 0 END
       • Example Number of developers: You don't need to transform this field and want to use is as is, so select oneToOne           
       • Example Competitor: You need to create a logic based on the field hg__tech, so select Conditional. List the technologies you want to detect from rarest to most common, because the output will be on the first condition in the list that is true.
       • Example Integration: You need to create a logic based on the field hg__tech, so select Conditional. List the technologies you want to detect. The output is 1 if any technology is detected, else it is 0.
4. In the section Options:
   
   • Check 'Mark as favorite' to see the computations starred in the Customer Fit models Insights 
     
   • If you would like to use this data in MadKudu API, check the box to make the computation available in the API. 
5. Click on Create computation.

Repeat the same for all the computations you would like to create.

Step 2: Deploy the computation(s)

Now you should be back on the computations list page.

You have just created a computation, which means you have added it to the library of computations.

If you would like to export it directly to your CRM or use it in a model, you'll need to deploy it. 

Click on Deploy and the computation will now being computed on the platform.

You will notice that the label of your computation has been automatically changed from its name to a more user friendly label, replacing underscores with spaces and capitalizing first letters.

Step 3: Visualize the impact of the computation on conversion 

To visualize the impact on conversion of this new computation, you can start from a new Customer Fit model or duplicate a Customer Fit model. 

Option 1: from a new model

1. Go to Models   
2. Click Create new model 
3. Configure the datasets (learn more on How to configure the Customer Fit datasets)
4. Once the dataset is enriched and loaded you should receive a confirmation email 
5. Head to Insights 
6. Search for the label of your computation and check out the result (learn more on How to read the Customer Fit insights)

Option 2: from an existing or duplicated model

1. In an existing or duplicate model, in the Overview section, click first on Reload computations to enrich the datasets with the new computation you created
   
2. Once the dataset is enriched and loaded you should receive a confirmation email 
3. Head to Insights 
4. Search for the label of your computation and check out the result (learn more on How to read the Customer Fit insights)

Congratulations! You've created a computation ;) 

F.A.Q

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

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

Why don't I see my field in the list to configure my computation?

You would first need to:

• make sure your CRM field is being pulled to the MadKudu platform.
• map your CRM field in the Attribute mapping  page in the app to be able to use it in the Data Studio.