A Little Bit PositiveBy David F. Carr | Posted 2003-08-01 Email Print
Unrealistic financial forecasts nearly tanked telecom giant Qwest. Can better analysis make a difference?
The first $2.75 billion phase of the sale, for the Dex East operations in Iowa, Minnesota, Nebraska, New Mexico, North Dakota and South Dakota, was completed in November 2002. The second partwith the possible exception of the phone-book business in Washington stateis set to take place this summer.
Tishkowski says his strategy team was asked to give advice on what the price had to be for the effect on Qwest's market value to be "at a minimum, neutral, and hopefully a little bit positive."
One of the basic analytical techniques Qwest uses to make this sort of judgment is discounted cash flow (DCF) analysis. Over the long run, stock prices are supposed to reflect the sum of a company's current and future free cash flowoperating cash flow, minus capital spending.
In DCF analysis, a dollar received today is assigned a higher value than a dollar expected in the future. So the projected benefit of any investment is "discounted" more steeply the longer it's expected to take to materialize. The "discount rate" in these calculations is an estimate of the average rate of return that could be expected from alternate investments.
QwestDex was valuable because it was a profitable business, requiring relatively little capital investment. While Tishkowski won't disclose the recommended price his group set, a Baseline analysis shows how a $2.75 billion price for the Dex East business could have been justified, based on a cash flow report published by the directory's new owners, Dex Media (see chart, next page). Tishkowski reviewed Baseline's calculations and found them valid, although not necessarily parallel with his own.
The quality of any forecast depends on the assumptions behind it. Qwest's strategy team takes some of the subjectivity out of its analysis by documenting the source of those assumptions. For example, growth projections start with telecommunications forecasts from research firms such as International Data Corp. The numbers from those general forecasts are then adjusted for Qwest's historical growth rates, market share and the competitive pressure it faces in each region.
This is where it helps to use business analytics software such as Cognos, rather than a series of spreadsheets, to construct the strategic plan. Tishkowski's team uses an Excel interface to the data model for some of its work. But the team's Cognos application, which they call the Enterprise Financial Model, makes it easier to organize the web of data, assumptions and calculations feeding into a forecast, and quickly show how projections would change in response to strategic moves like selling a business.
In analyzing the QwestDex sale, Tishkowski's team focused not only on what price would make sense to a buyer but also how much QwestDex contributed to the company's overall value. The trick was to weigh the immediate benefit of the cash Qwest would gain by selling the phone directory business against the future cash flow it would be giving up.
Financial analysts who follow Qwest doubt the company had the luxury of calmly applying a discounted cash flow analysis when figuring out the price at which it would sell the QwestDex business. "They were facing severe balance-sheet and liquidity issues, so they were at best a motivated seller and, at worst, a distressed seller," says Patricia Lee of the independent credit-analysis firm CreditSights. Qwest probably could have gotten $8 billion to $10 billion for the directory business if it could have afforded to wait for a better offer, she says.
Tom Friedberg, a Janco Partners telecom industry analyst and former U S West employee, thinks Qwest got a fair price for the directory unit, and sees the company practicing better financial management under difficult circumstances. "They've got to worry about return on value now, not some time in the future," he says.
Long-term, strategic planning is not a traditional use of the Cognos planning software. The program was designed to gather and analyze more short-term budgetary planning assumptions. Qwest originally bought the software for that purpose, but after learning of the finance department's implementation in early 2001, Tishkowski's group seized on the software as something that might also work for their purposes.
Since taking a weeklong training course in fall 2001, Tishkowski and two co-workers developed and maintained the computer model themselves, with no involvement from Qwest's information technology department beyond basic server maintenance. In its current incarnation, the model consists of 10 "cubes" (multi-dimensional data structures, as opposed to two-dimensional relational database "tables").
Also known as On-Line Analytical Processing (OLAP), multi-dimensional databases typically are used to analyze historical financial data in a flexible mannerfor example, allowing a report on "sales by region" to be expanded to show "sales by region by product line." The Cognos planning system uses a specialized OLAP database that tracks all the inputs into a forecast or budget and uses them to generate better projections.