Back from the brink

As the Internet and the technology stock mania led to a surge in the kind of electronic trading that Instinet had pioneered three decades earlier, the company nearly came unraveled. Finally, Reuters in April 2002 installed chief financial officer Mark Nienstedt as acting CEO.

Sam Shapiro had a problem. It was a spring afternoon in 1999, Nasdaq stocks were surging, and his Instinet terminal was on the fritz. He called customer service. “They said: ‘Your terminal’s down? We’ll send you a new one right away,’” recalls Shapiro. “I asked them what I should do with the old one, and they said: ‘Keep it. Throw it away. We really don’t care.’”

Shapiro, an Atlanta money manager, was thrilled at how quickly he got his new machine. In those heady, bull market days, he relied heavily on Instinet to buy and sell stocks. Its electronic system enabled him to match trades directly with other institutional investors, avoiding the high commissions of mainstream broker, who might have divulged his strategy to other clients to boot. But trashing terminals? Shapiro, whose Shapiro Capital Management Co. has $1 billion under management, didn’t think much of the way the company ran its business. “They used to throw away money,” he says. “I don’t think they had any cost controls at all.”

Shapiro had that right. Instinet’s lack of discipline, in fact, cost it dearly. Even as the Internet and the technology stock mania led to a surge in the kind of electronic trading that Instinet had pioneered three decades earlier, the company nearly came unraveled. Costs spiraled and upstart rivals stole large swaths of its market share, even as Instinet Group CEO Douglas Atkin set out on an expensive and ill-conceived expansion into ancillary markets such as fixed-income trading and retail brokerage.

Like the lavish headquarters it fashioned in the Times Square skyscraper occupied by Reuters Group, its corporate parent since 1987, some of Instinet’s spending was pure waste. But the company also had to fight off a host of more nimble start-ups, armed with newer technology, that exploited government-mandated reforms in the Nasdaq Stock Market. And it was losing the fight. By February 2000 Instinet had been surpassed as the largest electronic communications network by Island ECN, which handled 10 percent of Nasdaq trades. Island’s share had doubled by the start of 2002, while Instinet fell behind Archipelago into third place in a market that many experts believed would support just one or two dominant players.

The company’s financials weren’t much prettier. Between 1996 and 2001 Instinet’s revenues nearly tripled on surging stock trading volumes. But net income grew by a meager 21 percent during this stretch -- with profits actually declining in 1999 and 2001 -- as expenses soared. Aftertax profit margins plunged by more than half, to 9.7 percent. Instinet’s shares, floated in May 2001 at $14.50 each, have recently been trading at about $3.

Finally, Reuters took matters into its own hands, pushing aside Atkin this April and installing chief financial officer Mark Nienstedt as acting CEO. Nienstedt, who has since been replaced as CFO by Island’s John Fay and now holds the title of president, delegated day-to-day management to chief operating officer Jean-Marc Bouhelier, an ambitious 37-year-old Frenchman. Nienstedt and Reuters officials swiftly negotiated the acquisition of Island, for $364 million in Instinet stock. The deal closed in September. Meantime, Bouhelier hit the road to reassure and seek input from customers. He slashed prices and expenses while reorganizing the company around three client segments. He initiated a long-overdue technology upgrade and rolled out two successful new products -- front-end systems for active and passive money managers, dubbed Portal and Newport, respectively -- ahead of schedule.

His cold-blooded moves were wrenching but necessary. In all, Instinet wrote off $552 million in impaired goodwill related to the retail, fixed-income and other misadventures of the late 1990s. That resulted in a third-quarter loss of $528 million on $254 million in revenues. But Instinet decisively regained its market position: The firm now controls about one third of Nasdaq trading, double the 16 percent of No. 2 Archipelago.

Shapiro is so sanguine about the company’s prospects that he is now Instinet’s largest shareholder other than Reuters, which retains 62 percent. He began buying last spring after Atkin’s removal; by last month he had amassed roughly 10 million shares, a 4 percent stake.

“If you can buy a company for $3 a share that has close to $2 a share in cash and accounts for 30 percent of the Nasdaq trading market, that’s a no-brainer opportunity,” says Shapiro, who’s not alone. Other value investors, like Martin Whitman’s Third Avenue Funds, are also jumping into the stock, which has recovered from its 2002 low of $2.40 on October 16 but remains more than $7 below its January 7 peak of $10.68.

Are they in too soon? Instinet’s top brass may have halted the company’s harrowing decline, but many obstacles and questions remain. For starters, market share gains have yet to flow through to the bottom line, largely because they were obtained through price-cutting and the introduction of new products, which cost money to develop. Excluding the massive goodwill-impairment charge, Instinet made a mere $3 million profit, or 1 cent a share, in the third quarter. Despite cutting more than 700 jobs -- about a third of its workforce -- this year, the company remains bloated with unnecessary personnel and infrastructure, say top management and outside observers alike. More layoffs are imminent as Instinet CEO Edward Nicoll tries to make the company far more profitable at current revenue levels.

Combining Instinet and Island won’t be easy. Island has a brash, entrepreneurial style that is likely to clash with the more risk-averse and bureaucratic culture of Instinet and Reuters. Even top management concedes that some infighting is inevitable. Though Instinet was the acquirer, Nicoll, who was Island’s chairman, won the top job at the combined company. The team that led Instinet’s resurgence -- particularly Bouhelier -- may chafe at playing second fiddle. Already there has been fallout in the top ranks, though the first major casualty of the merger came from Island: CEO Matthew Andresen, who jumped ship in October to become head of global trading at Sanford C. Bernstein & Co.

Most critically, Instinet will have to contend with these cultural and integration issues while facing a fresh challenge from Nasdaq. In October the over-the-counter market launched SuperMontage, a modernized trading platform that poses a direct competitive threat to all ECNs. So far SuperMontage has not made a dent in the market share of Instinet and other ECNs, but Nasdaq has essentially bet its entire future on the system and can thus be expected to pull out all the stops before conceding defeat (Institutional Investor, September 2002). That might make the business of matching buy and sell orders for OTC stocks, no matter how much of it Instinet can garner, a fairly unattractive one to be in.

“They’ve warded off the market share erosion they were experiencing, addressed the issue of pricing and set up the company to compete in the market environment we have today,” says analyst Stephen Laws at W.R. Hambrecht + Co. “But there are always question marks in this space. Instinet definitely has its share of question marks.”

CEO Nicoll knows he has a tough assignment. “This is an organization that on an operating basis is breaking even, and with the market share that it has, it ought to be making money,” he acknowledges. “We know we’ve got a lot of work to do.”

INSTINET WAS CREATED TO REPLACE WALL Street. Instead, Wall Street became its biggest customer. Over time that became a big problem.

Founded in 1969 as Institutional Networks Corp., the New Yorkbased company aimed to link institutional investors directly with one another through an electronic trading network that bypassed the New York Stock Exchange and its member firms. But big pension funds and mutual funds were slow to abandon the NYSE for a new system that lacked the liquidity, or trading interest, to ensure that they would regularly find counterparties at the best prices. “They would say, ‘Yes, we’ll subscribe,’” recalls Jerome Pustilnik, the company’s principal founder and CEO until 1983. “And they did. And then they made very, very little use of us.”

Instinet didn’t become much of a factor in the trading world until the early 1980s, when Pustilnik shifted its focus from NYSE-listed stocks to the over-the-counter market, which lacked an exchange or any other mechanism for centralizing liquidity. The brokerages that made markets in OTC stocks wound up flocking to Instinet because it let them anonymously trade with each other at better prices than those they had to honor for public customers in the Nasdaq quote “montage.” Pustilnik’s successor, former Pacific Stock Exchange specialist William Lupien, continued to focus on this end of the business. And Instinet essentially became a private, more efficient market where dealers could carry out their principal transactions while executing customer orders using the public Nasdaq quotes.

“Instinet became the layoff market of choice for market makers in the 1980s,” says Bernstein’s Andresen.

With so much liquidity from market makers, Instinet finally attracted institutional order flow. And with no competition to speak of, it became so profitable that in 1984 several brokerage houses, including Merrill Lynch & Co., Shearson Lehman Brothers and E.F. Hutton, paid a total of $11.2 million for minority equity stakes totaling 14 percent. In 1987 Reuters bought them all out for $110 million. By the mid-1990s Instinet had gone from being a profitable side investment to a key subsidiary that accounted for one third of Reuters’ earnings and nearly one tenth of its head count.

But what made Instinet so attractive to Wall Street later contributed to its difficulties. In 1996 the Justice Department and the Securities and Exchange Commission accused market makers of colluding to keep artificially wide the spreads between the best bid and offer prices quoted on Nasdaq. Under terms of the settlement of the charges, the SEC in 1997 imposed new order-handling rules requiring dealers to publicize customer orders whose prices were better than their own quotes. The rules essentially nationalized the tighter spreads that had been present only on Instinet, shrinking the profits of dealers -- Instinet’s biggest customers -- in the process. When stock price quotes shifted from fractions to decimals in 2001, spreads shrank to a penny, completely erasing dealing profits.

The biggest consequence of the order-handling reforms was unintended: Most dealers weren’t technologically equipped to catalogue and integrate their customers’ limit orders into the public quote, so they began to rely on a new breed of trading platform, the ECNs. Rivals to Instinet offered newer and faster technology -- and lower prices. Some catered to the growing ranks of active individual traders who under the new rules could drive the public quote, just like market makers always had, by entering limit orders into ECNs. By 1999 Instinet was battling with 11 ECNs.

Plumped from years of plenty, Instinet failed to respond nimbly. Its system of subscriber terminals hardwired into a proprietary network became obsolete in the face of Internet technology and other inexpensive, open-standard communications protocols. Tech-savvy traders mocked the Instinet device as the “green screen” for its retro 1980s-era appearance. But rather than update its infrastructure and reduce prices to stave off competition, Instinet kept its fee schedule at the old monopoly levels and went on a spending binge in an ambitious bid to further automate the securities industry’s traditionally manual processes. In December 1998 it began developing an online retail brokerage called, on which it spent tens of millions before pulling the plug two years later. In October 1999 Instinet acquired German fixed-income brokerage Montag Popper & Partner, around which it built an online fixed-income trading operation. When shut down in May, it was losing $16 million a quarter.

Instinet is still paying the price for losing touch with its customers. “For so long we were just making too much money, and we got arrogant,” says Bouhelier, a former technology executive in London for Banque Paribas Capital Markets and Merrill Lynch, who joined Instinet in 2000 as COO of Instinet Global Equities.

Bouhelier describes Instinet’s malaise as “a little bit of the IBM disease: You had an organization that had been a leader in its sector for quite some time -- arrogant, poorly priced and trying to lock customers into the value chain. In IBM’s case Microsoft and other computer companies appeared. We had exactly the same problem. We had 15 percent of something that was getting bigger every day, with no pricing pressure, and we didn’t realize we were losing ground.”

Ex-CEO Atkin won’t comment on his three-and-a-half-year tenure as CEO, citing the terms of his separation agreement. But people familiar with Instinet’s inner workings say that he clashed often and bitterly with Reuters over his proposals for shoring up the ECN business, which centered primarily on making acquisitions to improve technology and eliminate competitors.

In 1998, for example, Atkin wanted to buy Austin, Texas based CyBerCorp, a pioneer in so-called smart order-routing systems, to narrow the gap with Island and other ECNs that already had similar technology for searching out the best price of a security across multiple markets. The Reuters board balked, and Charles Schwab Corp. bought CyBerCorp, which it renamed CyberTrader, in March 2000 for $500 million. And in 1999, mindful not only of competition from ECNs but also of Nasdaq’s initial work on what would become SuperMontage, Atkin went to the Reuters board with a plan to buy Island and the electronic brokerage company that started it, Datek Online Holdings Corp.

“The Reuters board just went nuclear,” says one person familiar with the matter. “They didn’t want to pay a big premium, and they didn’t want to get into what they regarded as a commodity business.”

Things got so ugly during this period that Atkin and Peter Job, then Reuters’ CEO, didn’t speak for several months, according to knowledgeable sources.

Though they rejected Atkin’s ECN proposals, Reuters officials approved of his efforts to enter new businesses. Indeed, say people familiar with the matter, as Instinet became a bigger contributor to the struggling Reuters’ results, Job and other top executives got increasingly involved after years of hands-off ownership. They encouraged Atkin to push into the retail and fixed-income sectors. Reuters also pressed Instinet to shed its entrepreneurial trappings and behave more like a big, mature company. During the mid- to late 1990s, for example, Instinet grew its legal department from just three lawyers to more than 20, and added similar heft in human resources and risk management. In 1999 the company more than doubled its spending on marketing and business development, to $23 million, even though revenues had grown by only 15 percent. In slick print and television ads, the company started calling itself “the world’s largest agency broker.”

Atkin couldn’t convince his bosses at Reuters to go after Datek and Island, but that didn’t stop Nicoll, chairman of both of those companies and CEO of Datek, from negotiating directly with Reuters management to do the very same deal -- except this time with the proviso that Nicoll would come in as co-CEO or sole CEO, accompanied by other Island personnel. In April 2000 the two sides signed a term sheet to merge Island and its parent into Instinet. Specific management roles had not been settled, but before they could be ironed out, antitrust issues scuttled the deal. Corporate lawyers vetting 4(c) documents, which include all correspondence related to the deal and items presented during negotiations, said that the papers contained too many red flags and that the government would likely hold up the acquisition as a result.

“Our lawyers said these were the worst 4(c)s they’d seen in their careers,” says one person who was involved in the negotiations. “There was all kinds of stuff in there like, ‘We want to use our market share to raise prices.’”

Reuters, for its part, takes a different view of the recent past. Thomas Glocer, CEO of the market data company and an Instinet director, says that Reuters wanted Instinet to acquire another ECN all along, and denies that the parent company pressured its subsidiary to add bureaucracy or invest in things like retail and fixed income against its will. “Instinet wanted to expand out of its core institutional business itself,” he says. “Reuters was very supportive of both the abandonment of the retail strategy and ultimately of the fixed income.”

As for why a deal with Island didn’t get done earlier, Glocer says: “The issue was always one of valuation and who should the management be, not one of whether we wanted to do a deal. And you can see where this time we came down in favor of Ed Nicoll.”

Instinet made yet another pass at Island in the fall of 2001. By this time, Datek had spun off the ECN; each company was majority-owned by a private equity consortium consisting of Bain Capital, Silver Lake Partners and TA Associates. With retail market conditions deteriorating, Instinet was no longer interested in Datek. (Ameritrade Holding Corp. acquired Datek this September for $1.3 billion.)

Instinet and Island could not agree on either a price or who would run the company, and talks broke down before year-end. Island came out fighting, cutting its already low prices in half.

“We said, Okay, it’s bare-knuckle time,” recalls former Island chief Andresen. “We felt that we were the only ones that could make money at lower price levels.” By January Island had boosted its share of Nasdaq trade volume to 22.8 percent, versus Instinet’s 9.9 percent. In mid-2001, Instinet had been only two or three percentage points behind.

With Instinet reeling, Reuters ended its increasingly acrimonious relationship with Atkin, 39, who had spent his entire adult life with the company, starting as an institutional salesman at age 21. Keener than ever on hooking up with Island, Reuters installed Nienstedt, a 17-year Instinet veteran, as interim CEO, and promoted Bouhelier to COO.

They inherited a mess. When Bouhelier reached out to clients, he found that some had not been called on for years. He decided that the company needed to be reorganized along client-segment lines. Previously, it did not differentiate, for example, between active and passive money managers, who can have vastly different needs. Now there are distinct product development, sales and support professionals serving each segment, in addition to the traditional ECN for broker-dealers.

Instinet’s technology is no longer a laughingstock. The company built an open front-end trading architecture based on the financial information exchange protocol, enabling customers to reach the network from their desktops rather than from a dedicated terminal. Instinet also unveiled a new trading engine that cut the average turnaround time on a transaction from 1.3 seconds -- the slowest of the ECNs -- to a highly competitive 0.2 seconds. And this summer it rolled out the Portal and Newport buy-side trading platforms. Some 300 institutions have deployed Portal since its introduction in April, and 22 major banks and passive money managers are using Newport, which provides program trading alongside execution cost analysis services.

“They’re certainly talking to their customers quite a bit,” says Richard Leibovitch, head of trading at Putnam Investments, of Instinet. “And they’re finally catching up on the technology side with some interesting new products.”

To assist with price-cutting, Bouhelier chipped away at the overhead that bogged Instinet down in the 1990s. Spending on employee compensation and benefits during the first six months of 2002 fell 34 percent from the comparable period in 2001. Outlays for communications infrastructure and professional fees were down 33 percent for the same period. Instinet’s total fixed costs are $175 million lower on an annualized basis than they were at this time last year.

Meanwhile, the company has cut the commissions and fees it charges customers by more than half. This step, combined with technological improvements, closer attention to customers and new products for the buy side, has restored Instinet to ECN leadership. The Island acquisition only solidifies that No. 1 position, while bringing in a customer base heavy with professional traders and active retail investors who complement Instinet’s emphasis on market makers and institutions.

“I just don’t see anyone in the sector who has a better opportunity right now,” says Bernstein’s Andresen. “Nobody’s ever had 31 percent of the market before. I think it can be an incredibly successful business.”

PLENTY OF CLOUDS LINGER ON THE HORIZON, however, and former Datek chief Nicoll faces a tough balancing act. The systems and cultures of Instinet and Island have to be integrated at a time when competition and cost pressures are not letting up and when it’s an open question whether the ECN business will even be worth staying in for the long haul.

Nicoll, 49, has taken a roundabout path to heading a public company. After graduating from Passaic High School in New Jersey in 1971, he was accepted by the Richard Stockton College in Pomona, New Jersey, but instead moved to western Pennsylvania and spent a few years raising sheep. He developed an interest in commodity markets, and later securities, and moved to Pittsburgh in 1975 to become a stockbroker. A year later he and Lawrence Waterhouse, a colleague at the brokerage firm Reynolds Securities (now part of Morgan Stanley), left to start discount brokerage Waterhouse Investor Services, now Toronto-Dominion Bank’s TD Waterhouse subsidiary. Nicoll departed as Waterhouse president in 1994 to earn a law degree from Yale University (which made the rare move of accepting him without an undergraduate degree) and ran the successful 1998 campaign of Yale classmate Cory Booker for a seat on the Newark, New Jersey, city council. In May Democrat Booker nearly defeated longtime Newark Mayor Sharpe James in a long-shot candidacy that Nicoll backed financially but did not manage.

Nicoll, who became president of Datek in 1999, is accustomed to lean operations. Before the merger Island employed a mere 175 people, yet had managed to nearly bring its new owner to its knees. Instinet, even after two rounds of layoffs, still employed 1,570. Nicoll knows what he has to do.

“We’re going to cut across the board,” he asserts. “This company can and will be profitable at September market volumes.”

Analysts and others who follow Instinet expect that Nicoll will hack off at least an additional 25 percent of jobs and slash real estate costs. One option he is considering is moving a large number of employees from Times Square to cheaper digs in Jersey City, where both Instinet and Island already have some staff. Reuters already has begun to occupy some of Instinet’s excess space at Three Times Square, notes CEO Glocer.

Potentially more vexing are the internal politics, both within Instinet and with Reuters, still majority owner and grappling with revenue and margin problems of its own. Nicoll must get along better than Atkin did with the people at Reuters who ultimately call the shots, beginning with Glocer, the first American CEO of the British media and information conglomerate. So far, Glocer says he’s quite happy with the current management team. Nicoll says he’s aware that some of his predecessors clashed with Reuters and notes that in his 14 years running Waterhouse he learned that more mature businesses like Reuters often need to behave differently: “I think I represent sort of a middle ground between the two extremes.”

The biggest concern internally is the Nicoll-Bouhelier relationship. After several months of autonomy working under CFO Nienstedt, a numbers guy who didn’t want the top job on a permanent basis, the hard-driving Bouhelier may recoil if Nicoll pulls the reins too hard.

Bouhelier won’t be shy if he feels affronted. In February he fumed to Glocer and Instinet chairman Andre Villeneuve that his 2001 bonus -- roughly $850,000, down almost 50 percent from 2000 as business slowed -- did not reflect his true contribution. To keep from losing him, the Instinet board authorized an increase to $1 million. That’s in addition to his $350,000 base salary and $92,600 in allowances for housing and school tuition, according to the firm’s proxy statement filed in April.

Bouhelier won’t comment on his compensation but isn’t afraid to toot his horn: “My view is that an organization should reward its most talented people in good years and in bad years. And I strongly believed -- and continue to believe -- that I brought more to Instinet in the past two years than anyone else. I never felt that the previous regime necessarily recognized how much value I had brought to the organization.”

That kind of talk doesn’t surprise people who know Bouhelier. “He’s incredibly talented,” says one Instinet veteran who has worked closely with him. “But he’s not a team player.”

Even Nicoll acknowledges his second-in-command’s extraordinary zeal. “Talented individuals are often seen by other people as being difficult to work with, because they have egos and because they care a lot and because they’re passionate about what they do,” he says. “I would put Jean-Marc in that category. But I like working with talented individuals, and I think that he and I can work well together.”

As Island’s chairman, Nicoll left much of the day-to-day operation to Andresen while learning the ECN business and managing relationships with his board of directors. But in his short time with Instinet, Nicoll has been more hands-on. Last month, for instance, he began a six-month global tour of Instinet customers, which allows him only about one day every two weeks in his New York office. And he replaced Nienstedt as CFO with Island’s Fay. But with Andresen, widely regarded as the architect of Island’s success, now departed, Nicoll would be loath to lose the remaining high-ranking insider, who could hold the key to the company’s long-term recovery.

For his part, Bouhelier says: “I think the fair assessment is that I’m a pretty determined guy, that I’m committed to doing what I’m doing. I think the people who would say I’m difficult to work with -- and I can think of quite a few -- are people who are less demanding and less passionate. But I certainly don’t feel that my authority has been in any way, shape or form reduced in the organization. In a way I feel more empowered than I’ve ever been, and that’s because Ed and I believe in some of the same things. It’s easier to be a change agent when you work for a change agent. I say to him, Now you can be the bad guy and I can be the good guy for a change.”

Investors also seem to welcome the change at the top. “Ed Nicoll is an excellent operator,” says shareholder Shapiro. “That’s what the company lacked. That alone makes me very optimistic about the future.”

So far, so good. But Nicoll must also wrestle with Nasdaq. It’s too early to tell whether Nasdaq’s SuperMontage will fulfill its stated purpose of reclaiming market share from the ECNs. Some observers say that Instinet can’t win, regardless of whether SuperMontage works as planned. If it succeeds, Instinet will be forced to shave costs and prices even more to stay in the game. But if SuperMontage falters after all that Nasdaq has invested in it, then Nasdaq may play rougher than ever. In one scenario, Nasdaq would exploit its monopoly in the business of gathering and selling market data, using these revenues to subsidize massive price cuts in the transaction services market where it competes with Instinet, putting the ECN at a tremendous competitive disadvantage.

And there are other bumps in the road for Nicoll. In October, for example, Instinet received two subpoenas from the SEC for information related to the trading of exchange-traded funds and to Island’s past rebating of market-data revenues from regional exchanges to its customers. Neither the company nor the SEC will comment on the inquiries. Nicoll also may have to improve Instinet’s governance. Its board is dominated by insiders, with nine of the 13 seats held by current or former executives of Instinet, Reuters and the investment firms that owned stakes in Island.

Instinet may not be throwing computers in the garbage anymore, but the hard work of growing earnings is just getting under way. It will need to do that soon to keep shareholders like Sam Shapiro happy.