Using Big Data to Optimize Pricing

April 12, 2013- 

By Doug Fuehne

Big data is undeniably a big deal for sales teams and executives. To give you a sense of what that means, research from Chango shows that people around the world send 145 billion emails, blast out 340 million tweets, and send 2 million search queries to Google every day.

Buried in that massive mountain of data — from credit card transactions, frequent-buyer points, and cellphone calls to web clicks and emails — lies important, game-changing value. The volume and velocity of this data is growing at an unprecedented rate. But for companies that successfully respond to the challenge and opportunity of big data, the result can be a series of insights that can have a rapid and significant impact on win rates, deal sizes, and sales cycle length — and, according to Gartner, can increase gross margins more than 2 percent on average (registration required).

Collective judgement of a free market
How does that happen? In many ways, it’s like the principles that James Surowiecki identified in his landmark book, The Wisdom of Crowds, which describes how large groups of people can collectively come up with precise answers to different questions. Big data gives us an unprecedented opportunity to aggregate information from diverse, disparate, and independent sources and replicate crowd wisdom in a marketing context, particularly pricing. This question is ideally suited for the wisdom of crowds. Fundamentally, what is a price if not the collective judgment of a free market?

At many companies today, data warehousing is the foundation of reporting and analysis initiatives, and that strategy does deliver some value. Sales representatives can answer some basic questions: What is this customer’s industry? How large is its business? In what region is it located? But that leaves unanswered the most important question in any sale: What is this particular prospect truly willing to pay? And answering that question means finding important correlations among a much wider variety of factors tucked in piles of data from inside and outside the organization.

Pricing analytics today seeks to define thousands of customer segments based on commonalities among dozens, or even hundreds, of independent variables. By finding the hidden correlations based on rich, deep data histories, we can make smarter decisions and pursue targeted sales, marketing, and pricing strategies much more likely to succeed. That means more revenue and faster sales cycles.

Let the crowd help you answer the key question
Using a range of sophisticated statistical analyses — as well as breakthrough IT infrastructures equipped to plow through and process terabytes and even petabytes of data — data scientists are unlocking the hidden value of big data for more effective sales and pricing strategies. They’re delivering crucial guidance right at the moment of transaction. They’re tapping ERP systems, CRM histories, competitive analyses, third-party econometric data, subscription sources, and much, much more. They’re pursuing the best answer to the key question: What is the maximum amount this prospect is willing to pay?

That analysis starts with an effective segmentation strategy: scanning and rescanning mountains of historical data to create peer groups of customers and prospects who demonstrate similar purchasing behaviors that are meaningfully predictive. But we’re not talking about simply lumping prospects into rudimentary A, B, and C buckets. We want to create hundreds, thousands, or even tens of thousands of narrowly, sharply defined segments defined by dozens of variables (or data dimensions), such as company size, industry/NAICS code, state, ZIP code, order frequency, average order size, preferred provider status, shipping preference, market competitiveness, sales channel (phone vs. web vs. in-store), preferred payment method, D&B credit score, time spent on website in past 30 days, and number of telephone calls received in past six months.

Suppose you sell office supplies. One peer group segment could be consumers based in eastern Massachusetts who have a purchase history of at least six months and make purchases every 30-40 days using the website. Their preferred credit card is MasterCard. They’ve accepted one upsell recommendation in the past year. They spend an average of 10 minutes per month on the company website, and they place an average of eight calls to the customer care center every six months.

With the data volumes usually available, this kind of segmentation can give rise to sparsity problems. But with big data, even a massive number of cells can be populated with enough data to derive statistically significant results. We can verify the value of these dimensions with simple analyses and assemble them into a predictive model. That lets sales and marketing tap that crowd wisdom more effectively and formulate optimal pricing levels and strategies.

Where big data has the greatest impact
One might be tempted to assume that big data-driven predictive models would be highly valuable to direct sales teams pushing complex, big-ticket sales. In reality, we’re finding that big data’s impact can be highest at sales and marketing organizations with two key characteristics: a large number of SKUs (in the tens of thousands) and a network of horizontal customers spanning different industries and geographies. In direct sales of high-priced items, sales teams have fewer transactions to manage, and they can devote resources to individual pricing analyses. However, in distribution organizations, margins are typically quite thin, and the power of pricing is often magnified.

Another strategic advantage of data-driven pricing analyses comes from speed. As the saying goes, time kills all deals. We can increase win rates significantly without sacrificing margins by presenting the optimal price as early in the sales cycle as possible. By avoiding protracted, outdated negotiation tactics near closing time, the sales team can avoid disruptions and delays.

For sales and marketing organizations seeking a surer footing as they optimize pricing, the best path may be to tap into the hidden power of big data and gain new insights into the factors contributing to price elasticity and sensitivity.