How analytics turned sales from art to science at Hewlett-Packard

When you think of using analytics to improve business functions, you might think of supply chain, operations or finance – but not sales. Sales is about relationships, which are difficult to manage through data analysis, right?

If you thought this way, think again. At Hewlett-Packard, we applied analytics to improve sales coverage for HP PCs in the Americas region. The result: HP grew revenue and improved gross margin. We are making fact-based decisions in less time, with more accuracy and less emotion. It took an edict from the top to make it happen, and we had to learn that being directionally correct is good enough and to accept the “80/20 rule” for completeness of information.

Historically, sales management was more art than science. Decisions about sales coverage, or where to hire and place sales people to best address customers in a particular geography or territory, were based on the sales leader’s experience. Seasoned sales managers who knew the territory might not do badly. However, when they left the job or the company, all the institutional knowledge about customers and the business left with them. The information was not captured, and the team had to start from scratch to rebuild the knowledge, often resulting in inadequate coverage of key territories and accounts, which could take years to fix.

This was the situation at HP until 2005. Then, top management asked sales leaders to define the total addressable market (TAM) and share of wallet (SOW) for their areas. HP defines TAM as what all businesses in a particular area spend on information technology (IT), and SOW is the percentage they spent on HP. TAM indicates HP’s potential revenue in a region, territory or account – and share of wallet shows how well HP is capturing that potential.

In the beginning, when asked about TAM and SOW, most sales managers did not know how to respond. They quickly learned, since they had to include this data each time they presented to HP’s Executive Committee.

Building the Model

To estimate TAM, we started with information from sources like Dun & Bradstreet and Standard & Poor’s. Then we refined our model to align SOW with revenue using information about MSAs (metropolitan statistical areas), which are cities or counties in the United States. From this, we learned where our TAM was largest. While TAM is an obvious measure of sales potential, we incorporated share of wallet to reflect the opportunity for HP. For example, even if the TAM of New York City is larger, we might add more sales people in Kansas City because our share of wallet is smaller in Kansas City than in NYC. The opportunity is greater and time to revenue is faster.

Analytics transformed how HP sells PCs for specific customer segments in the Americas. We can grow the business more than we thought possible because we have the data and analytics to show us how. Today, we can see coverage gaps, or where are we not adequately addressing the market. We know where to hire first and where to cut back when budgets are constrained.

Almost seven years after the executive proclamation, we use analytics to manage sales coverage as part of our day-to-day processes – and we continue to develop and refine our capabilities. We can analyze TAM and SOW down to a specific account for many of the customer segments, from large enterprises down to mid-market and small businesses in the U.S., Canada and Latin America.

Since we want to maximize profit – not simply revenue – we are getting more sophisticated about measuring margins and return on investment. Analytics give us real-time “levers” to tune our sales activities to achieve our objectives for margin, market share or revenue – whatever our focus. Continuing with the example above, we can see the results of adding that sales person in Kansas City instead of NYC. We can see which territories or accounts generate higher margins, and we can hire the right people to help us meet our goals.

Our decision-making process is better because of analytics. We make decisions more quickly, using facts and less emotion. Decisions are more scientific, comprehensive and egalitarian. Analytics also help HP speak the same language across the company. Each product and service group uses similar metrics to make decisions about sales coverage, and communication between groups has improved.

HP has been managing sales coverage with analytics in the following customer segments of the PC market: corporate-enterprise, public sector and the high-end of the small-to-medium business (SMB) market. As a next step, we are applying our knowledge and analytics to more transactional-oriented businesses: the consumer segment and smaller SMBs, those businesses with less than 500 employees.

While we cannot manage to the account level as we do with other segments, we can still apply the same methodology to manage and grow these transactional businesses. To accomplish this, HP centralized analytics for the region under one team for consistency of measurement and comparison across the company.

Advice for Sales Leaders

When I step back and look at what HP has achieved, I can offer advice to any sales leader embarking on the analytics journey:

• Give it a chance. Remember that sales and analytics do not historically go together. Be patient. Make sure you have the right tools, processes and capabilities to get the right data, which is where you must start.
  
  
• Accept imperfection. Having the right data does not necessarily mean perfect data. Recognize that you will never be 100 percent accurate, and accept a solution that is 80 percent. In a company like HP that is accustomed to precise measurements, we had to accept that we would never have all the data. We learned that going in the right direction is good enough.
  
  
• Get executive-level backing. Support from the top made all the difference at HP in making analytics a part of our daily sales management process.

At HP, we are more effective selling PCs, we make better and quicker decisions, and we are continuing to refine and improve our models – thanks to data analytics.