Oracle increases installed-base cloud sales using machine learning model

Using machine learning in Oracle Database for demand generation, Oracle sales reps pinpoint the most likely and profitable cloud customer prospects.


Reps felt they were given ML intelligence they didn’t have before. It encouraged all the right behaviors—reps were more enthusiastic about creating account plans and were willing to work harder to get a meeting. Meetings went better and we found new opportunities faster. That created a positive feedback loop.

Sanela HodzicVice President, Sales Operations, Oracle

Business challenges

Oracle Global Sales handles demand generation for Oracle Cloud Infrastructure (OCI) solutions. To motivate the representatives more fully and increase sales, the team needed to identify existing Oracle customers that were most likely to move to Oracle Cloud.

The aim was to help sales reps focus their attention on the most appropriate installed-base prospects while maintaining sales best practices. Additionally, these empowered reps could better connect Oracle customers to the cloud solutions that were the best fit for their use cases, improving overall satisfaction.

The team’s existing solutions were incapable of handling this level of sophisticated analysis in a timely way.

We have found consistently that 5% of accounts with the highest scores delivered 80% of the money won in any quarter.

Sanela HodzicVice President, Sales Operations, Oracle

Why Oracle chose machine learning

To encourage faster closes, and with better margins, Oracle Global Sales decided that it needed a sales-specific machine learning model. This would use historical data to rapidly create scores and pinpoint the best accounts. Managers would link the model to existing demand generation programs to streamline the process of feeding the Oracle cloud sales pipeline with the best leads.

The team tasked Oracle Labs to create a machine learning intelligence model to transform global demand generation and campaign design. Oracle Labs used Machine Learning in Oracle Database to create a robust and scalable model to manage 9 million product and customer data points and 2,400 scoring variables in Oracle Cloud. The scoring model required multiyear customer purchase and industry-specific data to predict future buying behaviors.

Oracle Global Sales linked the machine learning model to multiple demand generation programs for concrete, measurable results in the field. The organization’s sales representatives would train on this model, build trust in its usefulness, and rely on it for the long term, delivering immediate project return on investment.

Oracle’s machine learning-recommended leads result in deals three times larger than those in accounts in other territories, with a 160% higher win rate in the top accounts.


Oracle Machine Learning in Oracle Database enables Oracle Global Sales to quickly identify the installed-base clients most likely to migrate to Oracle Cloud. Each quarter, sales can identify the top 5% of accounts and what products they’re spending the most money on.

Oracle Global Sales can then accelerate the sales cycle and allow reps to focus on the most likely—and profitable—cloud prospects. Sales can also target the best sub-segments of a given account, which is especially helpful for larger international customers.

The scoring model and demand generation programs suggest the most likely cloud products for each account and boost customer satisfaction. Oracle Global Sales can also flag a cloud prospect that lacks a sales rep for its territory so that Oracle can hire appropriately to fill the need.

Over three years, leads recommended by Oracle Machine Learning resulted in deals three times larger than similar accounts in other territories, with a 160% higher win rate in the top accounts.

The machine learning model creates a large table that presents a score for each account. Closed sales are the model’s final success validation, and they factor into the model for future quarterly reports and scores. Because of the model’s recurring success, sales representatives are willing to trust it without excess technical explanations or training.

The score engine feeds leads into daily sales reports for representatives to act on immediately. In turn, sales reps work harder on the highest-quality accounts in the largest territories, enforcing best practices.

Published:June 24, 2022

About the customer

Oracle offers integrated suites of applications and secure, autonomous infrastructure in Oracle Cloud. The company operates in 175 countries, serves 430,000 customers, and drives US$40 billion yearly revenue.