Skip to main content
Industry

Your Secret Sales Management Tool: Machine Learning

As the Systems of Intelligence Era is upon us, we understand the transformative benefits that machine learning and advanced analytics can bring to businesses.

This powerful cloud-based service can help organizations mine and analyze their data for rich business and customer insights. It can be used for improving business processes such as predictive maintenance, allowing manufacturers to be notified about an equipment failure before it occurs. It can be used to recognize patterns, predict behavior, and adjust operations accordingly, such as the transformation Fujitsu is driving for farmers with its “connected cow” use case.

Are you leaving money on the table?
Let us now look at the application of machine learning from a whole new perspective: sales and marketing. Can machine learning’s predictive capabilities be used as a direct line to improve sales and sales management? Could it help to pinpoint your sales forecasts so accurately that you basically eliminate the need for sales teams to “sandbag” their sales opportunities?

The answer is “yes.” We are seeing this new and innovative implementation of machine learning proving quite lucrative for our customers. By applying machine learning to the sales process, we are seeing impressive results: organizations can boost sales by improving forecasting accuracy, and even improve productivity of global sales teams.

Machine learning can help you pinpoint exactly where you have missed revenue opportunities and where and how to apply just the right resources to maximize results. And, we are seeing up to an 90% accuracy rate in sales forecasting with this approach; needless to say, significantly better compared to human-made forecasts in the same sales environments. We are amazed, and our customers are amazed, with this level of improvement.

Machine learning for sales operations
Let’s look at how we helped one customer to increase sales and improve forecasting accuracy by applying machine learning algorithms. This customer was forecasting hundreds of thousands of opportunities in any given year and wanted to have a more accurate view of expected revenue as well as uncover hidden revenue potential. We examined their historical pipeline data and saw that they were 55% accurate in predicting wins and losses. So, essentially those odds were only slightly better than flipping a coin. They needed to do better.

Applying machine learning algorithms to two years of past sales data, we set up a training model to refine our testing and improve our learning. We analyzed the wins and losses, looking at details such as similar attributes, industries, services, products, account team structure and more. With our algorithms, we looked at 30 different data points and benchmarked against past successes and failures to get a prediction score for future wins and losses—down to each deal.

From this data, we could see sales trends at different times of the quarter, and even predict wins of deals that our sellers said were not going to happen. It was all done in a matter of seconds with machine learning. After a few iterations of this testing, our forecasting improved from 55% to an accuracy rate of 80%. This effectively reduced sandbagging and uncovered potentially lost revenue, which, in our customer’s case, was nearly $250M U.S

Machine learning for sales productivity
Machine learning isn’t just for improving sales forecasting. You can also take the same dataset and apply different algorithms to greatly improve productivity of your global sales teams, down to the individual seller level. You can help teams more deeply understand their pipeline, as well as expose them to new or different sales behaviors that you want them to embrace.

For example, when you look at deals that are potentially at risk, you can see what new or different resources you can apply to win the deal. Based on what has worked for similar deals, the data will tell you what will happen if you were to add a partner, a customer reference, or engineering support, for example. You can see, down to the percentage, how much you can improve your ability to win by applying each new option or by changing seller behaviors.

The results are truly mind-blowing and I encourage you to try this approach in your business. By using Microsoft Azure Machine Learning, the wealth of algorithms available from the Microsoft Azure Machine Learning Marketplace, and Microsoft’s Advanced Analytics services, you can get started today. Try a pilot on just one segment of your business, or a specific country, or even one product or service line. Reach out to your Microsoft account representative, who can guide you in a proof of concept or deeper conversations.

This is not about taking away the art of selling and replacing it with science. It’s about using the power of today’s technology to realize dramatic results for your business in a short amount of time. And, the results are truly transformative.