Lead Scoring set-up in Sales Cloud

[alinavol]

Hello,

prerequisites for a Lead Scoring setting in Sales Cloud are 1000 Leads and historical data. What workaround could it be, if i would like to set-up an empty DemoSystem? As i understood there is a possibility to import Leads, but not historical data. Can i train a model without it? And what Lead information is a must to work with?

[PU]

Hi Alina,

In my experience no, its a painful process. I used dataworkbench to load the leads and opportunities and then over time made alterations in the documents to build the models.

Regards,

PU

[alinavol]

mpuys which data for leads have you imported? And how long did it take after the import till the readiness report accepted the data?

[PU]

We set it up just when it was released and used the following fields :

External_Key, Name, Account_Party_ID, Qualification_Level, Status, Source, Priority, Start_Date, End_Date, Owner_Party_ID, Sales_Unit_Party_ID, Sales_Organization_ID, Distribution_Channel, Expected_Revenue_Value, Expected_Revenue_Amount_Currency

You might have another field(s) that you want use to affect the model.

Requirement for the last 12 months is 6k–10k, I loaded 5k leads
Requirement for total volume of data is 3k–5k, I loaded 5k leads.

Its great that they now only ask for 1000 leads.

[alinavol]

mpuys you mentioned, that you imported also opportunities. Without that, the model didnt work, right?

Could you let me know - which fields were necessary for Opps and how did u connect Opps with Leads?

Is it possible to connect with you in LinkedIn? 🙂

[PU]

Hi Alina, if you follow my profile in SAP, we can send private messages. Sorry I do not have LinkedIn.
Regarding the set-up, I set them up individually and there was no relation between my leads and opportunities.

[PrasanthArya]

Refer :https://help.sap.com/viewer/7947824c493f461d84e7bbc94c0dc3fd/CLOUD/en-US/dac6ab73cba14465a7ded68fde9bb372.html

Since manchine learning working on the concept of training and learning from exisitng data , without data , it wont be a feasible approach

So you can import some old data, basically a combication of all sorts of data (like lead of all status) that can represent an orignial business scnarion should do the job , but just random data and all data of a signle status wont enable the model as in fufture your model wil predict or show results based on what youhave given as input before to it in training satge