(FORTUNE Magazine) – Fidelity Magellan is the Coca-Cola of the mutual fund world--the closest thing the industry has to a signature brand. So it was big news in late October when the nation's largest mutual fund empire said that Robert Stansky was resigning as Magellan's manager and would be replaced by Harry Lange, a stock picker known for a more aggressive style. But the manager shuffle is just a small part of a much bigger story: Fidelity is in a funk, and it is finally making some dramatic moves to shake things up. The stage was set back in May, when Stephen Jonas replaced Abigail Johnson, the daughter of chief executive Edward "Ned" Johnson III and long considered his potential successor, as head of Fidelity's core money-management unit. (Johnson became president of Fidelity Employer Services Co., the subsidiary responsible for providing retirement and other benefits programs to employers.)

Fidelity insisted that the move wasn't a demotion for the boss's daughter, but a look at the numbers shows that the family-controlled fund group struggled during her reign, losing its spot as the nation's No. 1 manager of stock and bond funds. Vanguard took the title from Fidelity in February 2004, and Capital Research's American Funds bumped Fidelity from the No. 2 perch in June of this year. (Fidelity can still claim to be the largest fund company, but only when money market funds are included in the totals.)

This year has been particularly tough. Fidelity has taken in just $3.4 billion in new money since Jan. 1, according to Financial Research Corp. That's down 80% from the same period a year ago. American Funds has inhaled $61.2 billion this year, while Vanguard has collected $34.9 billion. And the stakes in this competition will get even higher in coming years as the giants duel over the trillions of dollars retiring baby-boomers will be rolling over from their 401(k) plans. Says Morningstar analyst Chris Traulsen: "Fidelity is really under the gun."

One reason for the slide is clear: performance. Many of Fidelity's signature large-company stock funds--including Magellan, Dividend Growth, and Growth & Income--rank in the lowest quartiles of their categories based on three-year performance. (Fidelity also announced last month that Tim Cohen would replace Steven Kaye as manager of Growth & Income.) Overall, Fidelity's diversified domestic stock funds have outperformed only 49% of their rivals in the past three years, according to Morningstar.

Jonas's mission is to improve those mediocre results. Soon after he took over, he and his team began scrutinizing the company's funds. "Domestic equity was a mixed bag," Jonas said during a recent interview. While half of Fidelity's $500 billion in domestic stock funds were "doing extremely well," he says, half were "not meeting our performance expectations." And that's not acceptable for the Boston fund behemoth. Unlike Vanguard, which is famous for its low fees and index funds, or the broker-sold American Funds, which excels at delivering reliable returns, Fidelity relies on hot funds to attract investors. "Fidelity is a performance shop," says Jonas, who wants all his funds to beat at least two-thirds of their competitors on a three-year basis, and for longer periods as well. "We will have good performance across all our asset classes."

To reach that goal, Jonas is revamping the firm's research operation. He will spend $100 million to double the number of U.S. equity analysts to 150 by next summer. And no longer will Fidelity hire all its analysts straight out of business school and groom them to become portfolio managers. Now the firm is bringing in experienced analysts and allowing them to follow companies for years. The hope is to develop a team of veterans who have a deep understanding of key industries.

The research staff will also be organized differently. To augment Fidelity's analyst pool, Jonas is creating a series of small teams of two to 20 to work closely with specific fund managers. Fidelity first started using this system in the mid-1990s in its bond division and generally liked the results. The teams will sort through research from the central pool of analysts and do intensive work on the most promising stocks. "The idea is to get a large organization to feel and play small," says Jonas. "We're trying to get the best of both worlds."

Vanguard and American Funds are thriving using very different models. Vanguard hires top firms such as Wellington Management to run many of its actively managed stock portfolios. The typical American fund has six to eight managers who operate autonomously, which in effect breaks each fund into several smaller portfolios. That approach allows the managers to focus on their best ideas. Jonas responds that there's no such thing as "the best" system. "You have a lot of different models from a lot of different organizations," he says. "I haven't observed one that wins all the time or one that loses all the time."

Magellan will be the big test for the new system. In many ways the fortunes of the fund mirror those of the company. Ned Johnson was the fund's first manager, serving from 1963 to 1971. But it was Peter Lynch's dazzling results--annualized gains of 29% for 13 years--that put the fund on the map and made it an enormously powerful marketing tool; long after Lynch left, the Magellan name helped persuade big companies to pick Fidelity to run their 401(k) plans. At its peak in 2000, the fund was the largest on earth, with $102 billion in assets, and delivered an estimated $500 million in annual fees to the parent company.

The fund continued to outperform the market--and pull in assets--under Lynch's successors Morris Smith and Jeff Vinik. But under Stansky, who took the reins in June 1996, Magellan has trailed the overall market for five of the past eight years and produced an annualized return of 6.9%, vs. 8.1% for the S&P 500. In response, investors have fled. Since 2000, the combination of market losses and shareholder defections has cut Magellan's assets to $52 billion. (Magellan has been closed to new investors since 1997, but existing account holders can add money, as can participants in 401(k) plans that offer it.)

Now it's up to Lange, 53, to turn the ship around. Lange first joined Fidelity as an industrial and machinery-stock analyst in 1987; he took over Fidelity Capital Appreciation in 1996. There, he displayed some of Lynch's gunslinging style. When the market plunged after the terrorist attacks on Sept. 11, 2001, for example, Lange doubled his technology-stock holdings to 30% of the fund's assets. As a result, the fund roared back as the market recovered. In his nearly ten years at Capital Appreciation, Lange posted an average annual return of 9.7%, topping the 8.4% of the S&P 500 and the 6% of the average large-company growth fund.

"This is good news for shareholders," says Eric Kobren, executive editor of Fidelity Insight, an independent newsletter based in Wellesley, Mass. "Lange will make Magellan a more aggressive, go-anywhere fund, as it was under Peter Lynch."

Size, however, remains a handicap. Even in its shrunken state, Magellan is seven times larger than the Capital Appreciation fund. That could hamper Lange's ability to maneuver quickly and take large positions in small stocks. "There's a practical limitation to running a fund of that size," says Morningstar's Traulsen.

Lange says he isn't worried. He expects to cope with the larger asset base by holding "a lot more" stocks than the roughly 200 securities he owned in Capital Appreciation. And he figures he'll rely on Fidelity's analysts much more than he has in the past. Lange also says he has discussed the possibility of "farming out" a small portion of the fund to an additional manager if he finds someone he's comfortable working with. For now, though, he's fired up about flying solo. "I'm going to make it a go-for-it kind of fund," Lange says. "I really want to shoot the lights out."

Feedback? investing@fortunemail.com