(FORTUNE Magazine) – The best research analysts are often a bit like investigative journalists. They might spend months sifting through arcane information while competitors, and sometimes even bosses, wonder whether they've gone off the deep end. Eventually, however, they emerge with a story no one else has.

One of the very best snoops in the securities business, and possibly the top researcher on our list, is three-time FORTUNE All-Star Andrew McQuilling of UBS, who covers household products companies such as Gillette and Colgate-Palmolive. He's earned a reputation for his prescient stock picking and willingness to be unpopular. Since 2000 he's averaged a 10% annualized return as his typical peer has lost money. Among his most contrarian recommendations: He told investors to bet on Procter & Gamble in 2000 when its business was in disarray and shares were plunging. The stock soon rebounded from $53 to as high as $93. In 2003 he picked Estee Lauder as shares remained stuck at their lowest level in years. It gained 77% within a year (including 35% after McQuilling told FORTUNE readers to get in last May).

In an era when everyone presumably has access to the same data--and when CEOs are prohibited from passing along information to their favorite analysts--why is one analyst able to do the job so much better than others? A lot of it comes down to what an analyst thinks his job is supposed to be. At some firms analysts are graded on their relationship to large institutional investors. Arranging meetings with companies as a sort of scheduling secretary becomes the main part of their work. Polls are also a huge driver in the way the job is done. The annual Institutional Investor report that names the favorite analysts of major investors, weighs heavily in determining compensation for many of them. Other firms--even after the Spitzer investigation--still give consideration to investment-banking activity.

To McQuilling, however, the job is about making money for investors, period. He in turn profits because the more home runs he recommends, the more likely investors are to trade through UBS. Though commissions have fallen to as low as 3 cents a share, the large volume of trading in household products stocks means that the money can still add up.

Beyond compensation, however, McQuilling thrives on the game. He loves the challenge of finding winners. Unlike many analysts in his sector, McQuilling couldn't care less about household products beyond what they mean to valuations. He likes to say that he covers "stocks, not companies." To do that, he compiles reams of original data on consumer habits, currency fluctuations, and historical valuation trends. He spends weeks searching for, say, obscure nutrition professors who can decipher numbers he has collected on eating habits. He writes his research with the help of two associates and crunches the numbers himself. Most days McQuilling, 38, can be found sitting between two massive computer monitors in a seven-by-ten office. The tiny room is jam-packed with spiral-bound notebooks, industry studies, and half-empty soda cans.

McQuilling's own research reports are not your typical sleep-inducing industry overviews. Rather they're edgy pieces of specific advice, such as "buy 50% now and the rest after the earnings call" or "wait three months and then get in." They're never boastful, however, because when he makes those recommendations he almost always looks as though he is making a mistake. That is, in fact, his goal: He wants to be early on a call because it takes time for large investors to buy enough shares.

"I'm used to being wrong for months at a time," he explains. "I don't know everything. I don't have a crystal ball." And so during those "months at a time" when he's usually the butt of jokes if not the target of investor wrath, he smiles and keeps digging. That was the case with P&G and Estee Lauder, and it's currently the situation with Weight Watchers. McQuilling began recommending Weight Watchers in November when the stock was at $35.88 a share. (See the main story for more on why he rates it a buy.) With shares near a 52-week low at $34.88, McQuilling likes it even more.

When FORTUNE met with McQuilling one recent afternoon, Weight Watchers shares had just fallen 11% on disappointing quarterly results. McQuilling did not appear embarrassed that he had told investors to start buying before the report, although he admitted he didn't expect such a negative reaction. Gripping a tiny soccer ball in his hand, he explained to an investor on the phone that Weight Watchers has compounded its sales at a 13% annualized rate since 1977, and that the company is about to unveil several new initiatives for boosting enrollment. "Don't abandon ship," he said. "Millions of Americans are sitting at home in front of their TVs, eating low-carb chocolate bars and wolfing down steaks, and they're wondering why they haven't lost weight. It's only a matter of time before they figure out why." --D.R.