Opening Cross: Is Data Latency Monitoring Music to Your Ears?

Music and market data have much in common, as anyone who has scrutinized the sheets of an orchestral score and wondered how those dots, lines and squiggles translate into soaring melodies will know. Yet despite the myriad different styles of music is that the most commonly used tempo across genres is 120 beats per minute (aka allegro, or equivalent to a brisk walking pace). Supposedly its popularity is because the beat mimics the human heartbeat when excited—excitement being what the music hopes to achieve in the listener—rather than at rest. So, you might think, it shouldn’t be a hard beat to stick to.

But I’ve also been part of bands and orchestras either incapable or unwilling to stick to the same beat, with the result that one set of musicians finishes before another, then sits awkwardly in silence waiting for the others to catch up.

My point is that music isn’t about blasting through a song as quickly as possible, but sticking to the correct tempo. That’s how the nuances of the music are communicated, how listeners get emotionally swept up in its melody, and what gives it the power to tug at the heart-strings (and ultimately, the purse-strings) of listeners.

Think about notes on a page as data: you have the different levels of notes (or strike price), the duration of notes (trade volume), harmonies (other market activity), and the interplay between each note to create a melody (or a combination of individual trades to form a larger strategy).

Just as the impact of a piece of music depends on adhering to the right tempo, the effectiveness of a firm’s investment strategy depends on receiving data at the appropriate latency—and with a crucial additional factor of balancing that against obtaining it at the appropriate cost.

As low latency has gone from being a game-changing advantage to a benchmark from which it becomes ever-harder and more expensive to achieve incremental gains—at least, that will provide any lasting advantage over equally keen and tech-savvy rivals—firms are reconsidering the benefits of being fast against how much it costs them to achieve it, how long the advantage will last, and whether they can make enough money during that timeframe to justify the investment. And since most of the low-hanging fruit—actually, more like everything but the very top of the tree—has already been ruthlessly harvested, fewer worthwhile investment opportunities exist on the latency front.

“If I can get a millisecond advantage, it’s probably still worth $100 million. But the problem is, there isn’t really a millisecond to give anymore,” said Spot Trading’s Daniel Penley at last week’s Inside Market Data Chicago conference.

With the level of investment required to play at the cutting edge enough to bankrupt some firms, it’s no wonder that others are urging caution and advising traders to consider the needs of their strategies before rushing headlong into buying expensive components only required by the most ultra-low-latency strategies.

But in all cases, it is becoming clear that the ability to measure latency is just as important—if not more so—than raw speed itself, especially if your objective is to understand the full impact of latency on a trade, and where latency is introduced: factors that may actually determine what type of strategy you should not pursue, and which will be more effective for your firm, given the different elements that may be within or beyond your control. Plus, how can you know how fast you are without measuring your speed?

So if you’re thinking about investing in a new instrument, or new technologies to improve your data and trading latency, first buy a metronome—or in the case of data, tools from vendors such as Corvil, which was recently rolled out by interdealer broker Tradition, or from Apcon, which has released a new sub-10 nanosecond timestamping switch. These will not only keep you at the right tempo, but will help you understand exactly where you’re gaining or losing time, where you need to focus your efforts (and money), and ultimately, what strategies are feasible.