A businessman is cruising in a hot air balloon. He’s lost.
He descends, looking for help.
He spots a man standing in a field, descends a bit further, and yells: “Hey! Can you tell me where I am?”
The man responds, “You’re about 30 feet above a field.”
The businessman says, “Oh, you must be an IT guy.”
“Why do you say that?” the man asks.
“Because what you’ve told me is technically correct, but it’s no use to me at all.”
“Well,” the man in the field retorts, “you must be in business.”
“Why?”
“Because you’re lost, you don’t know where you’re going, and now it’s my fault.”
This parable came courtesy of Christina Spade, senior vice president of affiliate finance and business operations for Showtime Networks, at this past week’s CFO Corporate Performance Management (CPM) Conference in Dallas. Along with its entertainment value, the tale is packed with the tensions that illustrate where business and IT find themselves today, struggling to make sense of and find value in the vast volume of data that organizations are collecting, storing, and processing. The pursuit of that value, the strategies both technological and operational that could unlock it to improve corporate performance, was the overarching theme of a conference that drew scores of finance executives to Texas, many in the same position as Spade’s balloonist: uncertain of where they are and in search of a map to guide them to where they want to go.
Why You’re Lost
The map to the treasure chest filled with gold nuggets of actionable business insight has long been understood to be business intelligence. BI, as Wayne Eckerson, president and founder of BI Leader Consulting defined it for the conference, is simply a way to use data to make smarter decisions. The problem, however, is that there’s nothing simple about BI.
The architecture upon which BI depends, Eckerson explained, is complex for both IT and the business. IT’s job is to extract data from various databases, transform and manipulate it in multiple ways so that it can be read, and finally load it into a data warehouse (a process called ETL) so the business can query it in order to design reports (financials, sales, human resources, whatever) that then can be analyzed and, perhaps, acted upon. This architecture, according to Eckerson, and the processes that arise from it make BI slow, hard to change, hard to use, and expensive. The historic leaders in the BI space — Hyperion (Oracle), Cognos (IBM), and Business Objects (SAP) — are gigantic enterprise applications priced out of the reach of most small-to-midsize businesses.
And even if BI were not so hard, Tom Davenport, professor of information technology and management at Babson College, and author of Competing on Analytics and most recently Analytics at Work, argued that the reporting that comes out of a business’s BI applications is based on the past. BI mines the data in the data warehouse to tell the person who receives the report where the business was at the time the data was first entered, either manually or automatically. In other words, it can tell you that you’re floating 30 feet above a field — which, as the businessman in the balloon would say, is of limited utility.
Where You Want to Go
Although it’s important to know where you are in order to get where you want to go, if you don’t know where you’re going, knowing where you are won’t help you.
“What makes an organization sustain profitability over time?” Kasthuri Henry, president of KasHenry Inc. and an expert in corporate finance, asked the conference. Her answer was the ability to look at the reports generated through BI and ask “Why?” Why are the sales numbers down? What’s the relationship between the kind of people you’re recruiting and the results you’re getting? “Managing solely with backward-looking reports is not going to help,” said Henry.
“You must identify the key performance indicators to guide investment,” noted Robin Washington, senior vice president and CFO of Gilead Sciences, a biopharmaceutical company specializing in developing treatments for HIV infection. To operate strategically — to figure out where Gilead should go, what markets it needs to address, where growth opportunities lie — Washington stressed the need for analytics that can give the organization the ability to develop predictive reports.
“It’s difficult to make predictions,” Mark Twain is reputed to have said, “especially about the future.” But in the first quarter of 2011, Apple sold 7.3 million iPads. In the last quarter of 2010, it sold 4.1 million. How did it know it needed to stock almost twice as many from one quarter to the next, especially as it (only) sold 3.2 million in Q3 2010? As Showtime’s Spade suggested, the answer could only come from predictive analytics, not from static Excel spreadsheets showing a curve of increasing iPad sales.
How to Build Your Map
When one gets right down to it, BI is basically reports — sales reports, P&Ls, audits — mostly born of Excel (the original BI tool). But Excel is primarily manual and therefore time-consuming and error-prone. Plus, all those spreadsheets issuing from all those silos within the organization are not easily rolled up to provide the single version of the truth necessary to know where you are with sufficient accuracy. Nor can they incorporate the new, often unstructured data flooding the organization that can point the way toward new opportunities and efficiencies. To do that, those reports must be informed by dynamic information feeds and massaged by business analysts who know technology — a relatively rare breed.
In order to do that cost-efficiently, Eckerson foresees the development of a new BI architecture based on leveraging cloud computing and new software-as-a-service analytics applications to lower costs and customize reports to two groups of users.
The first group is composed of executives and business leaders. They need to be able to monitor key performance indicators so they can see if the business is performing according to plan, and they need to be able to see this dynamically in order to make course corrections (adding leads or incentives to the sales pipeline, for example, or reducing head count). They may not need to see every report the business can produce. (Spade suggests withholding some reports. If no one complains, they’re probably not terribly useful.)
The second group is the business analysts who need to focus on what the business doesn’t know, what’s not in the data warehouse, and what’s changing. These users need access to Big Data, Hadoop clusters so that they can manipulate all that data, and a virtual warehouse so they can run virtual experiments without having to invest actual dollars.
And the business needs to be able to manage both groups of users, providing each with the technology and the reports each needs.
Davenport’s steps for reaching performance nirvana included accurate, timely reports (based on clean data); balanced scorecards; a strategy map; analytic models to test correlations between actions and financial performance; and, finally, the will to change.
For example, Davenport told a story about Victoria’s Secret. The company discovered (through sales reports) that although weekend sales generated a larger relative percentage of revenue than weekday sales, conversion rates were lower on the weekends, indicating lost opportunities. By comparing that with sales performance (using the balanced scorecard), the company realized that its best salespeople weren’t working weekends. It ran a model in which they did, and saw conversion rates increase. Consequently, the company implemented incentives to encourage its ace salespeople to work weekends and saw an increase in profitability. In other words, it matched its best salespeople to its best customers.
With BI that works, our balloonist is unlikely to get lost. Of course, it would help if he took that IT guy in the basket with him.