Given the nature of data typically available from sales activities, the revenue side of many organizations is a great place to start when you’re assessing the feasibility of machine learning use cases. Sales managers and sales operations leaders who want to use data intelligently to drive top line growth and reduce their costs of sales have endless opportunities to use machine learning for impactful results.

All Hail Sales Data

The unofficial and often messy source of truth for sales activities is an organization’s customer relationship management (CRM) software platform. Often there are multiple CRM platforms in use, due to technical debt from historically uncoordinated business units, acquisitions that were never fully integrated, and other reasons. This reality can drive sales managers and sales ops leaders to pull their hair out when trying to make data-driven decisions.

No matter how many CRM platforms an organization has, the key ingredient for machine learning is historical sales activity data and the more granular, the better, wherever it’s currently recorded.

Actionable Sales Cycle Intelligence

Most sales funnel and operational downstream activities are somewhat linear, moving from one stage or document process to the next. Think of the Marketing Qualified Lead (MQL) that matures into an outbound email and then into a proposed initial product configuration. If it continues to mature, it might grow up to be a quote someday, and if it’s the best of the best, into a signed contract. (Virtual high fives all around!)

Throughout the sales cycle, there are often dozens or hundreds of distinct stages and activities where individual contributors and managers make important transactional decisions. These decisions are typically made by relying on individuals’ abilities to manually process large volumes of data, either in their heads or in spreadsheets. Think of the sales rep crunching records and variables in Excel or maybe, if she’s lucky, on a Tableau dashboard, hunting in a sea of data and scenarios for who to reach out to, or when, or which configuration or price is most likely to hit the sweet spot.

Thankfully, DataRobot has reduced the barrier to entry for anyone interested in machine learning. You no longer need to rely completely on your sellers’ and partners’ gut instincts in these common scenarios. It’s now possible to feed those troves of historical data into a machine learning model that can suggest or, in some cases, completely automate the most favorable, revenue-impactful next step in the sales cycle. This is sales working smarter in your organization, not harder. 

Which sales leads should you focus on? Which of these new prospects are most likely to result in the greatest lifetime value (LTV) for your organization? Which partner is most likely to win this opportunity if sourced to them? What’s the optimal outreach channel, product configuration, or discount that’s most likely to result in a sale?  

Identifying the Opportunity

Machine learning use cases in the sales function are just like any other area of the business, limited only by teams’ imaginations to develop them (and perhaps IT’s willingness to help). Sales organizations that want to use AI and machine learning to grow revenue today should focus on the following:

• The problems or opportunities generally believed to be worth investing teams’ time and resources, for value-focused use cases to avoid pure science experiments.
• The critical data required to train a machine learning model, with sufficient historical granular observations in readily accessible systems.
• The downstream processes impacted by the model’s predictions (for example, by automating insights or suggesting an optimal decision to sellers and managers).

It’s critical to know which of these activities is most likely to bring the greatest financial impacts to your organization, based on perceived growth opportunities or known pain points in your sales processes. As one colleague likes to remind me, if you focus on the two comma ROI use cases, everything else just seems to fall into place.

There are countless opportunities to start using the historical data organizations typically collect to drive better decisions. If you’re working in the sales function and want to get started or expand your organization’s ability to use AI, here are 20 machine learning use cases to explore. As they say, when the sellers are happy, everyone’s happy. Happy selling!

20 Machine Learning Use Cases to Drive Revenue by Predicting: 

1. Conversion likelihood for prospects, for lead qualification and prioritization.
2. Customer Lifetime Value (LTV) for improved lead scoring or account management and expansion efforts.
3. Prospect year one and future years’ profitability for lead scoring and prioritization.
4. The next best offer for products or services, or next best sales action to improve cross-sell and upsell activities.
5. Payment defaults from pre-customers to improve financing and credit risk scoring for assigning account payment terms or payment tier.
6. Seller performance, including quota attainment for improved sales management. 
7. Pipeline conversion at individual or segment level for improved sales management or revenue forecasting.
8. Which marketing-sales outreach channel will be most effective per prospect for increasing campaign optimization and effectiveness. 
9. Optimal timing of outreach based on individual buyers.  
10. Difference between initial quoted price and final sale price for improved discount optimization and strategy.
11. Optimal total price for deal size maximization.
12. Likely products or services configuration from initial inquiry or lead for improved seller efficiency and effectiveness.
13. Likelihood of a deal to close once quoted, or at any other stage in deal lifecycle.
14. Deal size to better prioritize sellers’ time and attention.
15. The number of days until deal close based on known information at time of quote or customer legal review, to identify opportunities for interventions.
16. Quoting and configuration anomalies or errors based on products and services included, to proactively detect pricing variations and corrections required.
17. Deals most likely to be stalled or otherwise stuck in the cycle, identifying those that require manual intervention or special terms to advance.
18. Fraud or errors in inbound sales information requests and RFPs.
19. Likelihood of churn for resource allocation and targeted customer interventions.
20. Customer complaints for insights and/or proactive customer care and interventions.

DataRobot offers an end-to-end platform that appeals to a broad mix of users, helping organizations around the globe transform their data into value. In an ever-changing market, our strategy for delivering the industry’s first end-to-end platform for enterprise AI is game-changing for customers, and is key to enabling the success that many companies achieve with AI. If you’re interested in accelerating how your organization leverages AI to drive sales, contact us for a demo.

About the author

Darren Stevenson

AI Success Director at DataRobot

Darren Stevenson is a Director in DataRobot’s AI Success organization where he helps customers and partners leverage the power of machine learning. He lives in suburban Chicago with his wife and kids. Prior to joining DataRobot, he was a management consultant at PwC US and has held a variety of advisory roles in strategy, data, and technology over the past 15 years. Darren received his Ph.D. in Communications from the University of Michigan and is a regular contributor to academic, government, and industry consortia in the US and abroad.

Meet Darren Stevenson