MadKudu | Turn more leads into paying customers | IBM Lead Scoring Case Study

IBM focuses sales efforts with predictive lead scoring

IBM Cloud (formerly known as IBM Bluemix and IBM SoftLayer) is a suite of cloud computing services from IBM that offers both platform as a service (PaaS) and infrastructure as a service (IaaS).

IBM Cloud IaaS allows organizations to deploy and access virtualized IT resources -- such as compute power, storage and networking -- over the internet. For compute, organizations can choose between bare-metal or virtual servers. While IBM Cloud PaaS allows developers to use IBM services to create, manage, run and deploy various types of applications for the public cloud, as well as for local or on-premises environments. IBM Cloud supports various programming languages, such as Java, Node.js, PHP and Python and extends to support other languages.

The business challenge

IBM Cloud platform supports access to other IBM tools and services -- including IBM Watson and IBM Cloud Functions for serverless computing -- as well as those from third-party vendors. The IBM Cloud Catalog lists about 200 services across categories. It's ease of use, flexibility in deployment have made it a great success on the demand generation side. The Platform attracts in the hundred thousand leads any given month.

The exact cost of IBM Cloud can vary based on resource usage, deployment models, support and other factors. IBM offers a free tier, a subscription tier and a "pay-as-you-go" pricing model.

The complexity of potential billing along with the high volume of inbound leads had led IBM to partner with MadKudu to help better identify the leads and accounts that were ready for a sales conversation.

How MadKudu helped

MadKudu Lead scoring performance
Leveraging MadKudu's advanced machine learning, IBM was now able to identify "A" leads which
converted 280 times betterthan the D leads!

IBM connected their Segment instance to MadKudu and replayed historical data to allow for the models to start training.

The MadKudu platform started pulling conversion information about IBM's historical leads. MadKudu automatically started enriching leads, deriving features (aka potential predictors) to predict conversion both with firmographic data as well as behavioral data. Using advanced machine learning, MadKudu identified which combinations of features were highly predictive of the potential of a lead and an account. The results were staggering and showed that MadKudu could identify the best 13% of leads that accounted for over 75% of the conversions.

When looking into the scores of leads that had been worked by the sales team historically, MadKudu found there was a high potential to improve the team efficiency by aligning their efforts against the predicted potential of leads.

Better alignment
MadKudu helped drive initial adoption by providing visibility into the scoring and opening the black box for the sellers and management.

The results

Because of the critical nature of the endeavor, IBM ran a pilot with MadKudu for 2 months to ensure the models would perform live better than their current in-house model. And it did, very much so!

The benefits of using MadKudu are now two-fold. It enables the sales team to leverage "Signals" for better context around a lead health, while working fewer leads to prioritize only the high potential ones. More importantly it has allowed IBM’s data science team to refocus their efforts on building intelligence and analytics capabilities for other products.

MadKudu Leadscoring in action
IBM is able to use scores and signals to enable sellers to hit their targets and more

Implementing MadKudu has allowed us to get sellers to have a more targeted outreach. I've been impressed by the high predictive rate of grades A and B towards lead conversions.