Big Data in the Selling Process

At a fundamental level, business is all about value creation through innovation and value capture through the sales process. Yet, when we think about which aspect of business can benefit the most from a transformation, sales does not always come out tops. As business leaders, we relegate sales to ‘salesy’ people who somehow have managed to acquire persuasive skills. It does not appear that ‘Big Data’ and ‘Sales’ occur in the same train of thought. Big data is best understood as information that has volume (very large), velocity (rapidly changing) and variety (changing data fields). Big data has become an important issue in corporate meeting rooms because of the rise of mobile technology, affordable elastic cloud computing and the impact of network effects on how products scale. Companies like Netflix, Walmart and P&G are using big data across sales channels to drive topline growth. For almost 200 years P&G has been an innovative company. In partnership with Google, P&G uses cloud computing features to drive deep-selling across it’s products. For example the P&G OralB iO toothbrush uses user data to recommend parts replacement – thus growing customer lifetime value. Netflix uses contextual bandits to sell more shows on its platform. It modifies over 20 different features of each movie tile shown to drive engagement – some features it dynamically tweaks are tile pictures, movie descriptors, relevant rankings, location on the screen, etc.

More generally, the framework for sales process comprises 4 steps (see figure 1 below). Within each step, there are opportunities to utilize big data and analytics to increase value capture.

Lead (opportunity) Scoring

Predictive lead scoring applies big data and machine learning algorithms to evaluate the key behaviors of existing customers (using internal data) and prospects (using external deal data) and rank them using a metric (lead score) that can distinguish customers and prospects who are more likely to convert, retain, or buy from the company’s products and services. Lead scoring achieves two objectives. First, it allows sales teams move into sales engagement (actual selling) rapidly for hot leads. Second, it allows the marketing team to carry out more focused campaigns and investment to improve quality/receptiveness for cold leads. 

Symptom to look out for:

A key symptom to know where or not your sales process requires sales scoring is the ‘sales sameness’ test. If your sales team applies the same effort for every business opportunity, when price and value are held constant, then this is a strong indicator that you need lead scoring capabilities. 

Sales Engagement

Sales engagement optimizes the product assortment and sales approach problem. It answers the questions “what should I sell?” “how should I sell?” and “what price point is optimal?” Engagement scoring uses big data and machine learning to answer these questions. The ideal engagement scoring sales model comes with a recommendation engine that recommends to the sales team (in B2B scenarios), or sales channels e.g. websites (in B2C scenarios) which products or service to offer in the first sale – and at what price range. Most engagement scoring models use multi-armed bandits, or contextual bandits algorithms that continuously optimize sales engagement. These models are much better and stronger that regular A/B tests. They predict the optimal set of products, services and price points that maximizes a customer’s likelihood of making a purchase. This helps your sales team create a more personal experience and connect with your customer.

Symptom to look out for:

A key question to ask yourself is this “holding lead quality and client needs constant, am I offering the same product/service to my leads?” if your answer is yes, then this is indicative of an improvement opportunity in your sales engagement approach.

Deep-selling/Cross-selling

Deep-selling and cross-selling are sales approaches that answer the questions “what more should I sell?” “what other products can I sell?” “At what time after my first sale should I make the next proposition?” Deep selling increasing client wallet share by selling more of the same product to a client. Cross-selling increase client wallet share by selling a different related product to the client. Many B2B companies are implementing next-product-to-buy algorithms that draw on data about what similar customers have bought. The approach also helps retain customers. Engaging customers at risk of leaving for a competitor requires recognizing the signs of customer discontent well before they take action. These types of problems are perfectly suited to the pattern-recognition skills of machine-learning algorithms. Deep-selling/Cross-selling algorithms study related transactions from other clients to let you know what next to sell. Consider the case of a software company. Suppose that after purchasing 100 licenses on average (based on live data from other clients), there is often a need to upgrade to the enterprise version of a product, having an algorithm proactively identify when and what to pitch as the number of licenses purchased inches closer to 100, would dramatically increase value.

Symptom to look out for:

A key diagnostic question to ask yourself is this, “does the average sales staff on my team know when to sell more?” If your answer is “no” then there is an opportunity to innovate sales using big data.

So, What Do I Next?

There are two simple things to do to unlock the value of data and analytics and transform your sales function. First, focus only on one segment of your sales process (see figure 1) and implement – moving from left to right. There is typically no need to improve deep-selling when there is no lead scoring system in place. It helps to focus on one use case and optimize it deeply before moving right to the next step of the journey. This is especially true because big data becomes more useful as higher quality data becomes more available. Over time, data quality improves as positive early results justify greater investment in data infrastructure and quality. Second, ensure big-data solutions can be consumed by people on the sales force on the field. The acid test of a good sales function digitization efforts is if the insights can be used by the newest member of your sales team in the field. Sales efforts that release periodic reports to management teams are not sufficient to drive the depth of change needed.

Poatek is a leader in providing state-of-the-art services across the entire technology stack of firms. We have a unique combination of strong engineering expertise and business experience serving clients across Europe, US and Latin America who desire to implement sales function transformation initiatives drives by big data and analytics. 

Featured image by Freepik