Predictive Insights - Salesforce Integration Guide

Overview

The enrichment process pulls objects from Salesforce which were recently modified and if relevant fields were modified their scores and insights will be refreshed. Additionally, the integration also pulls objects whose scores are older than a certain refresh period even if they were not modified within Salesforce. That way, scores and insights always remain fresh.

Enrichment, Scoring and Integration Overview and Options

  • We support integration for Salesforce Lead, Contact and Account objects.

  • Leads, Contacts and Accounts can be scored and enriched using Account Based Markets (Models).

  • Leads and Contacts can also be scored and enriched using Lead Based Markets (Models).

  • You can run scoring and enrichment using more than one market (model) for the same type of object. In this case, you should create different fields in Salesforce in order to maintain scores and ranks for each market separately.

  • Alternatively, you can enrich same object using different markets (models) based on your own criteria.

Details of different integration types will be described below.

To facilitate the integration, certain fields used by internal Predictive Insights processing will need to be created. In addition to that, you will need to create fields for Predictive Score, Predictive Rank, and Insights.

Steps to Integrate

  1. Connect Predictive Insights and Salesforce Platforms.
  2. Create fields in your Salesforce instance as described below.
  3. For every market and object type configure integration as described below.

You will be able to repeat steps 2 and 3 as needed and add more fields in Salesforce and then to the integration. You will also be able to add more markets to the integration.

Connecting Predictive Platform to Salesforce

To get started, a connection needs to be established between Predictive Insights and Salesforce – you’ll need to be both a Predictive Insights and SFDC Administrator to do so. Go to Settings External Platforms and toggle ON Salesforce. The application will prompt you to login to Salesforce instance and you will need to allow Predictive Insights to perform operations inside your Salesforce instance.

Once you have this turned on and authenticated, you will be able to adjust certain settings for integration within the Predictive Insights platform.

Note: If you would like to test the SFDC integration in your SFDC sandbox, it is possible, however the integration cannot be authenticated the SFDC sandbox and production at the same time. After you have tested the integration in sandbox, you will have to go into the Predictive Insights platform and authenticate the product instance and walk through the process below again.

Screen Shot 2020-07-30 at 9.58.37 PM.pngScreen Shot 2020-07-30 at 9.59.48 PM.png

  • Daily API Calls Limit Limit number of API calls that Predictive Insights can use for this integration.

  • Score and Insights Refresh Period max age (in days) for scores and insights.

  • Daily Pull Time Time of day for daily pulls of data for enrichment. As described above, objects that are modified are constantly pulled, on an additional daily basis we also pull objects that were not modified, but should be refreshed, nevertheless. Daily pull times will adjust to your local time zone based on your IP address.

Pre-requisites and Configurations in Salesforce

To enable the integration, we need the following fields to be created on every type of object that needs to be integrated. We recommend that you create those fields on Lead, Contact and Account objects. Please make sure that case of the API names is as follows below.

    • Mintigo Hashcode TextArea(255) API Name: Mintigo_hash__c

    • Mintigo Update Datetime – Date/Time API Name: Mintigo_update_datetime__c 

Note: These fields are case sensitive; The API names must be exactly as written above. Please also ensure the correct field type for each. The Salesforce user used for the integration should be able to modify (write) to those fields. These fields do NOT need to be mapped during the integration in the platform. These fields allow to track changes in the object (record) data.

To consume Predictive Insights information you will need to create additional fields. The following would be the typical fields for you to create:

  • Predictive Insights Score number (3) you should create it for every market you integrate. Possible values would be 0 100.

  • Predictive Insights Rank text (1) you should create it for every market you integrate. Possible values would be A, B, C, D, F (F is optional and depends on configuration in the Predictive Insights platform).

  • Predictive Insights are typically of one of the two types, boolean (Checkbox) or string (Text). For every insight you will need to create a relevant field.

Note: For every object that the integration will be working with, the Salesforce user used for the integration should have a Create permission for this object.

Data Flow

Every few minutes:

  1. Fetch all integrated objects that have been modified since the last pull.
  2. If any of important (*) fields were modified, run scoring and enrichment for a given object.
  3. Push updated objects back to Salesforce.

Note: Important fields are the fields that are being used as input for the integration.

Once a day, perform a daily pull of integrated objects that were not modified in the “Score and Insights Refresh Period” number of days. All those objects will be scored, enriched, and pushed back to Salesforce.

Note: Those settings are configured in Settings External Platforms.

Enrichment Integration

We will review simple integration steps for a single market, we will enrich Account objects using the same account-based market. Other object's configuration will be very similar.

Note: Please make sure that you have created relevant fields in Salesforce as described above. The Salesforce user used for the integration should be able to modify (write) to those fields. Also, make sure that you have enabled the insights and Predictive Insights fields that you are interested at under the Enrichment tab of the relevant market.

 

To enable integration, go to the market of your choice and then choose ‘Market Integrations’ from the Other drop-down menu. Then choose ‘SFDC RT’ tab.

Screen Shot 2020-07-30 at 10.10.54 PM.png

It is required to define integration settings for every Salesforce object type separately. Currently you are in the Accounts configuration tab.

Screen Shot 2020-07-30 at 10.12.11 PM.png

Step 1:

Map fields that are being pulled from Salesforce in the left-hand pane, in this case, we will map Account Name and Website. Fields that were used by the market building are marked by an asterisk.

Screen Shot 2020-07-30 at 10.16.43 PM.png

Note: Those input fields are monitored for changes by Predictive Insights. If the Salesforce object had one of those fields modified, it would trigger refresh of scores and insights in Predictive Insights


Step 2:

Map fields that you want to export to Salesforce in the right-hand pane. In this case, we mapped Predictive Insights Score and Rank, as well as, two sample insights. Additional insights can be added or removed from the integration as needed once the integration is live.

Screen Shot 2020-07-30 at 10.21.36 PM.png

The platform will not allow you to send data to incompatible fields, for example you will not be able to send textual insights to Boolean (Checkbox) fields in Salesforce. Field type is the first thing to make sure is correct if you are not seeing data for an insight sync to your SFDC.

Step 3:

Once, you have chosen the fields, you should:

  • Turn on the integration with the ‘Market Integration’ toggle at the bottom right of the screen.

  • Enable persistence of data in Predictive Insights platform this step is optional due to compliance requirements, however, we recommend that you persist data in Predictive Insights.

Screen Shot 2020-07-30 at 10.22.50 PM.png

Step 4:

Hit ‘Save’

You are now integrated for Account object. If you want to integrate the same market with Leads and Contacts you will need to repeat the steps above within their respective tabs. As mentioned above, configuration for each object type is independent of others and you will need to separately map fields that are being pulled from Salesforce as well as mapping of insights into fields within Salesforce.

Note: The integration will allow saving even if you forget to map input or export fields, however both mappings are required for correct operation. The platform will remind you that mapping is required.

 

 ⚠️ Please map both input and export fields in order to ensure the integration is processed correctly.  ⚠️


Special Notes on Integration of Contact Object

Contact objects in Salesforce are reliant on some fields from the parent account object. For example, Account Name and Website are on account object while contact name or email are on the contact object.

To help you easily distinguish account originated fields from contact originated fields, we have added ‘Account:’ prefix to account originated fields. For example:

Screen Shot 2020-07-31 at 12.36.38 AM.png

Advanced

You can have specific control over enrichment of each object within Predictive Insights. We will discuss two possible use-cases when this might be required.

Use-case 1 - You only want to send certain Leads and Contacts for enrichment in Predictive Insights. To control which Leads and Contacts are enriched, we suggest that you create a special Boolean (Checkbox) field on Lead and Contact objects within SFDC, call it ‘Predictive Insights Enrichment Candidate’ for example and then choose it in the drop down. (You can name the field anything you would like, but we encourage you to put Predictive Insights in the field name).

 

Screen Shot 2020-07-31 at 1.20.28 AM.png

 

After you ‘Save’ the integration, only Leads and Contacts that have ‘Predictive Insights Enrichment Candidate’ checked in Salesforce will be enriched in Predictive Insights.

Use-case 2A: You might have markets for SMB, Mid-Market and Enterprise.

  • You want to use those markets to enrich and score different accounts based

    on their properties.

  • To achieve this, create Boolean (Checkbox) fields in SDFC on Account objects

    for every market type and use those fields to route the accounts to the correct market for enrichment.

OR

Use-case 2B: You might have markets for SMB, Mid-Market and Enterprise.

  • You will score all desired records against all three models. Therefore, there

    will be a Predictive Insights Rank and Predictive Insights Score SFDC field for each model. (e.g. Predictive Insights ABM SMB Rank; Predictive Insights ABM SMB Score; Predictive Insights ABI Enterprise Rank; Predictive Insights ABM Enterprise Score, etc.).

  • Once all scores and enrichment has taken place, create a Boolean (Checkbox) field filter in SFDC to apply/show the correct Predictive Insights Rank and Score based on the model criteria (e.g. Annual Revenue, Number of Employees, etc.)

  • This option will avoid any delays between enriching a record for firmographic details via Predictive Insights and THEN scoring against one of the three desired models.

 Score Records After Integration

  • How do I score all records against the model?
  1. Option 1: Manually
    1. Manually score all records against the correct model in the Predictive Insights platform.

    2. Once scored, upload the records to your SFDC instance.

  2. Option 2: Through the API

    1. Once the integration is turned on, Predictive Insights will start scoring records in your SFDC instance. At this time, there is a limited of 10,000 records that can be scored each day through the integration. If you need all records scored faster than that limitation, please see ‘Option 1: Manually’ above.

  • How do I rescore all my accounts against the new model(s)?
  1. If you had a different model integrated and now are integrating a new model, please take the following steps rescore your records.
    1. Step 1: Remove all the “Predictive Insights Created on” data from your records. This can be done in SFDC across your records. This will trigger Predictive Insights to start rescoring account records.
    2. Step 2: The SFDC integration in place will see the missing fields and ‘rescore’ the desired records. This should happen in the daily push to update any records that have been changed or altered in SFDC. 
  • How long will it take to score records through this new integration?
  1. Mentioned above, currently there is a limitation on 10,000 records a day. You can calculate time to score based on the total records you are looking to score.
  2. Note, if this calculated time takes too long, you can manually score all desired records against the models in the Predictive Insights platform and upload to SFDC manually. (See FAQ #1).
  3. If you score items manually, all new or adjust records with be scored by the Predictive Insights integration.
 

Monitoring Integration

 Ensuring records are being scored and monitoring for errors can be done in the clients SFDC instance. To monitor, visit the Monitor Bulk Data Load Jobs page in SFDC. This can be found by going to Setup →Monitor→Jobs→ Bulk Data Load Jobs.

Screen Shot 2020-07-31 at 12.50.36 AM.png

On this page you will see the list of jobs in progress. The example highlighted show you a job that is attempting to update three records. You see the job identified that two records in that update failed.

Note, if a job is updating a large number of records and a small number fail, it doesn’t not stop the update from happening to the other records, as seen in this example below.

Screen Shot 2020-07-31 at 12.51.22 AM.png

To find out more about those two records that failed, click on the Job ID under the in progress list.

Here you can see the details of the job. If you click on ‘view results’ on this screen, you will get the SFDC error message for the failed records.

 

Screen Shot 2020-07-31 at 12.52.36 AM.png

Screen Shot 2020-07-31 at 12.53.18 AM.png

Q & A

Q: How often does Predictive Insights pull data?
A: Predictive Insights pulls data for updates every 60 minutes. Daily data pulls happen according to user definition.

Q: What if I add insights to the integration at a later date? Will the data populate on the account records?
A: Yes, the additional fit or intent insights will be populated through the Predictive Insights integration.

Limitations

  1. Only Lead, Contact, and Account objects are supported. Custom Objects or Opportunities are not supported by this integration.
  2. Location-Based Intent is currently not supported by this integration.
  3. Integrations: Either a sandbox or production integration is permitted on each model at once. Once you test the integration in the sandbox, the process has to be done in production with an SFDC production username and password.

Troubleshooting and Technical Notes

  1. Changes in Account fields such as Account Name or Website do not trigger change on Contact objects that are linked to the same account.
0 Kudos
Contributors
Latest Articles
Modeling Data Requirements
Predictive Insights
4 weeks ago
09-08-2020
08-04-2020
Labels (1)