Connecting to New Markets: Eight Costliest Network Mistakes

While technology has caught up with the needs of the market, many traders are still stuck in the mindset of years past and are losing money because of it.

Theres an adage that says one needs to spend money to make money. This is often repeated by people who are trying to sell something, but for the most part its true: no business infrastructure of value comes cheap.

For those trading firms looking to expand their business, the costs of trading in new asset classes can be astronomical. Conventional wisdom has held that achieving reliable, safe trading capabilities is a Herculean undertaking that should be left only to the largest, most well-funded players. But that way of thinking is changing.

While technology has caught up with the needs of the market, many traders are still stuck in the mindset of years past and are losing money because of it. Here are eight reasons many companies networks are costing them a lot of potential revenue.

Time to market exceeding 72 hours: A recent audit of a trading desk at a tier-one institution indicated a 44-day turn up period for each new client connection, because of approval, procurement and networking obstacles. The desks smallest client generated $1,000 of revenue a day. Thats a minimum loss of $44,000 in lost revenue per client per day.

Building more network than needed: According to recent industry figures, capacity on long-haul fibers remains 91 percent unused. Still, businesses pay for 100 percent of the bandwidth, over-spending by a factor of 10.The same audit of the tier-one institution desks revealed that 30 percent of its connections were completely inactive and bringing in no revenue. Yet these connections still resulted in monthly recurring. Without accurately identifying costs on a per-client connection basis, margins dont increase, yet spreads on assets continue to decline.

Lacking oversight of network connectivity: The complexity of networks, particularly in large institutions, makes it difficult to accurately identify the number and destinations of network connections. When new trading opportunities arise, existing infrastructures cannot always be isolated, resulting in delays in time to market and potential duplication of connectivity costs.

Running outdated compute facilities: In accordance with Moores Law, computing power doubles every 18 months, resulting in faster processing times at half of the cost. Average business depreciation, however, ranges between three and five years. This means businesses are running half as efficiently at double the cost when compared to competitors who regularly upgrade their facilities.

Inflexible IT budgets: As spreads in traditional markets decrease, more traders are looking to higher volatility assets to drive revenue. Yet the costs and risks involved in traditional entry into new markets can mean not enough IT funding is available to allocate to these efforts. Executing on these opportunities becomes financially unfeasible.

Building out new space in data centers: Low-latency requirements demand that businesses be co-located in the same data centers as the institutions they connect to. As markets shift and demand increases beyond American and European financial epicenters, building and maintaining a latency advantage requires huge expenditures and long-term investments.

Calling the IT department to spin up each new connection: When an opportunity exists, traders want to connect to it. Instead, most businesses are calling their IT department — after theyve received approval — who subsequently liaise with the destination institutions IT department, in order to begin setting up a connection. The ideal financial infrastructure should facilitate immediate trading activity, not mire it in unnecessary process and bureaucracy.

Entering into long-term service agreements. For trading firms, most network contracts represent long-term costs and commitments, even though revenue happens in the short-term and is never guaranteed. The industry needs to move towards a model where connectivity and computing costs are much more flexible and itemized to reflect only the services/venues being tapped at any given time. Firms should be able to add or delete services for any billing period based on how well their business is performing.

There are better ways to trade financial assets rather than the traditional buy or build methods. Too many institutions spend too much on building infrastructure that they dont have the capacity to maintain properly. And long-term commitments to large networks that go mostly unused are just not smart business.

Executing and maintaining op-tier trading capabilities is going to be expensive enough. Traders can get a leg up by leaving behind outdated practices that are costing them revenue.

Jacob Loveless is the CEO of Lucera and former head of high-frequency trading for Cantor Fitzgerald. He has worked for both HFT groups and exchanges for the past 10 years in nearly every electronic asset.