How We Picked The 50

The BusinessWeek 50

Any method for ranking companies will by necessity be imperfect. But for this year's BW 50 list, we have tweaked our methodology to better identify top performers, applying some of the lessons we've learned over the past 10 years.

First, we focused on two core financial measures, average return on capital and sales growth, both taken over the previous 36 months. For our measure of profits for nonfinancial companies we used earnings before interest and taxes. The earnings measure also excludes distortions from special and nonoperating items as defined by our data provider, Standard & Poor's Compustat. We then put these profit numbers into context by figuring them as a percentage of the value of invested capital, basically long-term debt and shareholder equity. For financial companies, we calculated return on shareholder equity using pretax profits as our earnings measure, the better to align our figures with the nature of the finance sector. For our measure of sales growth for all companies we included gains from operations and from mergers and acquisitions, as reported by Compustat.

Second, we compared companies with others in their sectors. (Ten sectors cover the companies in the Standard & Poor's 500-stock index.) This break from our past practices enables us to identify companies that are good performers relative to their peers, even if their sector of the economy is not booming. This avoids the situation, for example, where oil companies jump to the top of the list when oil prices are rising and then drop off the list when oil prices are falling.

We started with nearly all the companies in the S&P 500. We left certain ones out of our analysis because they had become public too recently, such as telecom Embarq (EQ ); because they had continuing accounting issues or insufficient data; because they were a government-sponsored enterprise, such as housing lender Fannie Mae (FNM ); or because they are likely going to be purchased, such as Biomet (BMET ). Education lender slm (SLM ) was omitted because it still had ties to the federal government in 2004, within our three-year window of analysis. The McGraw-Hill Companies (MHP ), BusinessWeek's parent, was omitted as well, to avoid any appearance of favoritism.

Within each sector we ranked companies separately by our two measures--return on capital and growth--and then combined the numerical rankings, giving substantially greater weight to return on capital. The top company in each sector, according to the combined ranking, was given a 1, and the bottom company received a 0, with all the intermediate companies receiving a score according to their ranking. For example, Coach (COH ) finished first in return on investment in the consumer discretionary sector and third in sales growth, which ended up making it the top performer among its peers.

A list was then constructed of all the companies we analyzed, according to their scores. We applied a tiebreaker where necessary. Finally, this list was reviewed by a panel of editors and correspondents, applying one thing we learned from 10 years of compiling the BW50: Financial measures applied mechanically sometimes miss the mark. Based on extensive discussion, the panel made a small number of adjustments in the ranking.

The new procedure does have some quirks. A company such as Sempra Energy (SRE ) may rank ahead of a company with higher return on capital and sales growth such as Lehman (LEH ). That's the result of focusing comparisons within sectors, which we think over the long run will produce a more reliable list of the best performers.

    Before it's here, it's on the Bloomberg Terminal.