Einstein Discovery does supervised machine learning to detect patterns. These patterns can help you to identify major business issues which otherwise could have gone unnoticed.
It also helps to Disseminate insights to rest of your organization.
It can also prescribe things that you can do improve your business.
Example usecase is how can we prevent customers from leaving.
Create a variable that tracks the thing that you want to solve.
Another example usecase is how can we improve value for our customers.
Another example usecase is how we can close deals faster.
Nobody's data is perfect. That does not mean that data is not useful.
Good to analyze different segments separately so that erroneous data in a segment does not interfere in analysis of other segments.
Sign up for dev trail org at : https://sfdc.co/ED-org
Create a story by going to dataset in Analytics Studio
Decide on what you want to maximize or minimize. Select variables to consider.
You can select a date range to consider sample data.
There is a feature where user can click on "why it happened", which gives details of why part.
Example shows coupons are doing well
It is recommended to first run insights and based on results tweak the parameters and once final run with predictions
You can compare models.
Once reviewed, deploy model to Salesforce.
You can deploy model to salesforce object. You can select actionable variables which you can change after prediction runs.
You can also deploy NOT into salesforce object.
You can use this in a dataflow. Prediction node need to be added in dataflow using the prediction model defined above.
