Making it fit

The success of Big Board CEO John Thain’s deal making depends largely on whether the NYSE’s technology chief can streamline its infrastructure.

Once, New York Stock Exchange members painted the bourse’s windows black so that outsiders couldn’t spy on the frenzied gesticulations of its traders and somehow divine the market’s direction. Now, of course, computers instantly disseminate information from the exchange floor to the farthest burgs. And trades are increasingly executed electronically, a fact the NYSE acknowledged in April when it acquired automated exchange operator Archipelago Holdings. That deal had barely closed when Big Board CEO John Thain struck another: a proposed combination, also valued at $10 billion, with the fully electronic pan-European stock and derivatives market Euronext.

The aim of all this deal making is that NYSE Group, as the newly demutualized and publicly traded company is now known, will be able to supercharge its profits by executing more trades in more kinds of securities and in more locales around the globe while consolidating the infrastructure through which those transactions are processed. The latter part of that job falls to Steven Rubinow, NYSE Group’s chief technology officer. His task isn’t easy: If the Euronext deal wins approval from shareholders and regulators, the NYSE will have no fewer than three different trading systems and ten data centers. Minimizing overlap between them will be a key to the Big Board’s future earning power. One particularly thorny problem is how to marry systems so that the NYSE’s planned hybrid market, which seeks to meld the best attributes of both floor-based and automated trading, can share technology with the electronic matching platforms run by Archipelago and Euronext.

Another big challenge for the company is business continuity planning. The NYSE remained shut for four days following the 9/11 terrorist attacks - even as competing electronic exchanges stood ready to do business - in part because it lacked sufficient backup for its trading floor. Five years later the exchange seems resigned to the likelihood that if the floor is rendered inoperable, customers will, as Rubinow puts it, “switch their business to another surviving trading venue.”

Rubinow was CTO at Archipelago for five years before the merger with NYSE. In the early 1990s he worked as a vice president of corporate management information systems at Fidelity Investments - where he was responsible for integrating systems among more than 40 divisions - before leaving Boston for Silicon Valley. There he put in stints as CTO of online consumer credit upstart NextCard and chief information officer of online advertising company AdKnowledge, which was acquired by Internet investment fund CMGI in 2001.

The CTO says his start-up experience is the most valuable asset he brings to the NYSE. “I have an acute sense of time-to-market issues,” he says. “You do things at a small organization that a large company would deem impossible. When I came to NYSE, I said, ‘Don’t let the nonessential things distract you.’”

Rubinow recently discussed the NYSE’s technological future with Institutional Investor Contributor Julie Segal.

Institutional Investor: What is the NYSE’s technology strategy?

Rubinow: It’s the three zeros: zero latency, zero cost and zero downtime. That’s what customers expect. While those are noble goals and you can’t really achieve them, you can strive for them. What I don’t want is to sacrifice one attribute for another. I’m also a big fan of simplicity. Simple systems are cheaper and easier to maintain. When things inevitably go wrong, it’s not hard to figure out where the error is. For instance, I admire Google, but it’s a highly distributed company with hundreds of thousands of servers in a complex model. We need simpler systems with fewer numbers of moving parts. I also want to get systems out the door as quickly as possible. We don’t want to be laggards and sacrifice an opportunity because we took too long.

What’s the biggest technology challenge you face?

The biggest challenge is not technology at all. It’s not hardware or software. It’s getting the people of NYSE, Arca and SIAC [Securities Industry Automation Corp., a technology systems entity owned jointly by the NYSE and the American Stock Exchange] on the same page, and that takes time. A lot of processes and bureaucracy develop over time. In a start-up you don’t have that legacy, and there’s the appetite to get stuff done. We need to get the right amount of process to keep things together without slowing things down. Some people think that if you just follow the process, then the operation was a success even if the patient died. Some people here think I’m being too aggressive and that I don’t give enough credit to process. But it’s not up to me - it’s the competitive market we face.

You’ve cut technology staff in recent months. Will that continue?

There’s never a good time for layoffs, and I can’t predict the future and say we’re done. But we have to do more with less. The answer is to boost productivity, which is not the same as working 80-hour workweeks. We have to work harder, with a greater sense of urgency.

What are your disaster backup plans?

We’re prepared for typical disaster scenarios - power failures, software bugs, hardware failures - with data centers from both Arca and the NYSE that are geographically spread out as well as available locally. We have diversity of power sources, network vendors and other systems to keep business flowing. In more catastrophic scenarios, such as 9/11, if a market center went off-line, people would switch their business to another surviving trading venue. From a broad perspective, financial markets will be maintained because the markets are interconnected and customers also have backup plans. It’s not by preagreement - we’re not asking anybody else to back us up, or vice versa - but remaining operational sites will take over as people seek an exchange location with which to do business.

What’s your strategy for integrating NYSE’s systems with Arca’s?

A couple things are guiding us. One, Arca has low costs and low latency, but reliability is a tiny bit behind the NYSE systems. On the other hand, NYSE has emphasized reliability, but cost and latency have not been as important. If you could shuffle those two decks of cards, it would be a perfect combination. We also need to consolidate common elements between the two systems. Each market has its own customer gateways for orders and its own way to distribute market data, for instance. We only need one of each. NYSE has superb floor technology. The trades submitted from the floor have to go to a trading engine, and both NYSE and Arca have those. So we’re putting those together too.

What about merging Euronext’s platforms with your own?

We have many concrete ideas about what we want to do. Overall, we have too much of everything, and we need to boil it down. For example, between the NYSE and Euronext, we have ten data centers. That will be whittled down to four. We have two equity and two derivatives systems. We need one of each. But we still need to figure out the right number of platforms based on customer, cost and regulatory issues.

Can you have multiple platforms and still deliver the operational efficiencies that shareholders want?

Our ideal has always been to have multiple products trade on a single platform, but the question is whether that makes sense in this case. Can a single platform meet all the different needs our customers have? These multiple platforms may have to exist because of local requirements, as in one obvious example regarding latency. If a trader in Paris is very latency-sensitive and we send that transaction to the U.S. for processing, latency will clearly become a problem. But even if we stick with multiple platforms, we can get the efficiencies we need because we will still have fewer systems than we have today.

Will finishing the NYSE-Arca integration slow down the merger with Euronext?

As we work to complete the integration between NYSE and Arca, we’re factoring in where Euronext will fit. The deal is not done yet, but I’m reminding my staff all the time to keep Euronext in the back of their minds and to keep certain systems’ doors open for Euronext. We’re prepared. I don’t want to oversimplify it, but these are all trading systems. We’re not trying to integrate a trading system with a space shuttle.

Related