What’s wrong with SIFMA?

The 2011 SIFMA technology event is over, and I find myself this week back on a plane to New York (via Boston). Melanie wrote a great piece on the incredible shrinking show. Here’s the problem. There are two schools of trading – Old and new. Old school remembers the days traders would come down from the desks after the close to peruse the new terminals. The brokerage folks would come to look at workstations and back-office systems and fancy new tools called CRM. New school is black box, quant, HFT, algo trading. SIFMA is run by old school. Technology was just different. Technology buyers were different. You could look at technology. Touch it. Move the mouse (or hit the right function combo for you 3270 folks).

“Here. Let me show you how this works”, said the average salesperson.

There were terminals and laptops on every display. The show organizers seem to think there will be a comeback someday the likes of the blue eye shadow and Chuck Taylors my daughter and her friends are bringing back from the 80s. Sorry folks. There is no comeback.

Howard, I like you, but you guys need some fresh blood organizing this conference. The new technology buyers can’t touch and smell and play with the new trading technology. How do you play with a Spread Networks or an Exegy or a Correlix?

The new buyer needs to learn. Instead of the topical discussions on that stage by all the big vendors, teach. Put me up there with Sybase, Streambase, and Progress and let’s talk about CEP. We’ll discuss the use cases. We’ll talk about challenges. Roadmaps. Take questions from the audience. If Jeff, Mark, or Dr. Bates try selling their products or picking apart their competitors on stage instead of in the booth, trust the moderator to drop the hammer.

Let’s talk about hardware versus software in feed handlers. Let’s discuss fiber options. Colocation and proximity hosting. Storing market data. Algo chassis. Pre-trade risk controls. Messaging middleware. FIX networks and connectivity.

You want to bring back the decision makers? You want them walking the floor? Give them a reason to show up. If a CIO/CTO could spend two days a year keeping up with the latest technology, I would advise every one of them to do so. I talk to a lot of them and few are current on the vendor community capabilities. The little upstarts are never going to get an audience with that executive if they cold call in. Put them on a stage talking about emerging technology where the CIO feels safe and un-accosted in the audience and he or she may learn enough to want to wander over to the booth.

Build the show around the new buyer. The path you’re on, I wouldn’t advise a single client to spend their marketing dollars exhibiting. Throw in a panel slot on their relevant technology with their booth and we’re talking a different game. When I started going to SIFMA technology events you had a better acronym and were the show vendors thought they had to be at so people knew they were a player. Now? Not so much. Right SunGard? Broadridge? Thomson Reuters?

Sentiment Analysis for the Average Jane or Joe

It has been awhile since my last post.  Sorry about that.  Lots and lots of stuff going on that I can’t write about, so I’ve been content deprived.  I’ve also been in the media a fair amount these days because semantic analysis is hitting its stride (three years after my first report on the topic).  I wrote a piece in this blog awhile ago talking about how institutional trading technology has a way of migrating to the retail market.  For semantic analysis, that seems to be happening much earlier than I expected.  Most trading firms are still in exploratory phases with unstructured data in black box trading.  But, the world is already opening up through online channels to retail trading people.  I’ll give you three bookmarks, and if you know of more, please comment on this article with the site to share with the rest of us.

1. StockTwits
2. TweetTrader
3. Flame Index

The first two track Twitter for stock sentiment, which is what everyone wants to talk about these days.  The third site tracks news sites and trends news sentiment for particular instruments.  I also had a chance to talk to John Coulter over at Titan Trading about their arrangement with Penson.  It looks like Penson is going to use Titan’s analytics to offer up unique algos to their institutional clients.  Truth be told, I’m disappointed.  Penson clears for some of the fun active trader sites and could be a leader in offering up unique tools to that community as well with very little, if any, extra effort.  Ah well.  To be king for a day.

To all you research providers, discount brokers, and personal financial management providers, you might take a look at some of these tools to deliver something different to your clients.  I mean, isn’t the whole social media thing about knowing and being known?  What investor isn’t going to be interested in what other people think about their investments at an aggregate sentiment level?  Anyway, I just thought those sites were kind of cool and I should share with my friends.

Trusted Sources Can Be Dangerous

The Associate Press reported on a press release indicating General Electric would pay US $3.2 billion in back taxes. The problem with the report is the press release was fake.  GE opened at $20.05.  At 11:09, they were up to $20.08.  By 11:59 GE was down over 1% to $19.87.  Now, unlike Apple and United Airlines, both victims of news-reacting algorithms in 2008, GE did not experience the rapid drop false news can have on a stock.  That said, the many firms we speak with exploring trade triggers on news and other unstructured data beware.  You can be gamed.

In this case, you would have been gamed by some corporate-hating ideologues called the Yes Men, who apparently don’t realize that the average American’s retirement savings are in mutual funds, which, you guessed it, hold stocks.  So beating up on GE doesn’t hurt GE management. It hurts GE shareholders, many of whom are the proletariat masses those idiots hope to champion.

I digress.  The point of this exercise is to show my unstructured data friends that even the AP will let you down, so watch those trade triggers.  In the slow latency world of reference data management, firms bring in multiple feeds.  Three is popular.  Scrubbing logic assumes a price is correct if two of three match.  You might have to think about applying similar logic.

Imagine a more nefarious exercise where someone understands how algorithms consume news and apply sentiment analysis.  There are a myriad of ways, including issuing false press releases or untrue statements on Twitter for you Twitter fund groupies, to affect stock sentiment.  Maybe someone only needs GE to move down 1% in 50 minutes to make money.  Keep that in mind as more firms employ low latency trading strategies based in part on news.  At that point, penny stock spam campaigns will be rookie league compared to what’s possible at the institutional level.

NASDAQ OMX NYSE Euronext Exchange Technology Perspective

I spend a lot of time with the technology arms of both these exchanges, so I figured while everyone else debates what a potential name might look like, I’d opine on what a marriage might look like from a technology offering perspective.  In short, while there is some overlap, there are some highly complementary offerings.  While trading volumes and listings may be down, the real money potential from a growth perspective would come out of the market data and technology areas.  NASDAQ OMX keeps those groups separate right now.  NYSE Euronext has them combined.

On the exchange technology side, NASDAQ OMX has the lead.  Over 70 exchanges run on NASDAQ OMX technology and HFT shops I’ve spoken with like the matching engine.  That said, NYSE Euronext is no slouch and maintains a small cadre of exchanges running their technology as well.  A marriage would provide ownership of a significant portion of global trading done on a single provider’s technology.

Risk management maintains the biggest overlap now that NASDAQ OMX owns FTEN and NYSE Euronext offers RMG.  This is also becoming a popular area for market entrants right now.  Celoxica, Fixnetix, Mantara, ULLINK, SunGard, BonTrade, etc. are all vying for pre-trade market share.  I’m not sure how RMG and FTEN would come together.  That said, NASDAQ OMX also owns SMARTS, which controls half of the commercial market for trade surveillance at exchanges and regulators and has its sights set on the broker/dealer community.  That market is owned by Actimize, with SunGard and Oracle (Mantas) following.  Combining strong pre-trade capabilities with market surveillance provides a compelling differentiator and could be interesting.

For market data products, NYSE Euronext and NASDAQ OMX are quite complementary.  NASDAQ OMX has made some forays into providing retail access to both real-time and historical market data, while NYSE Euronext has been focused on creating more commercial institutional products like their consolidated tape offering and their Wombat ticker plant / feed handlers.  Combining the market data teams is perhaps the most exciting potential product area.  Fritz and I looked at the exchange data market potential last year and found it a hot revenue area.

Lastly, NYSE Euronext is much further along building commercial product for broker/dealers.  They have the processes in place, the management structure and coordination, and the experience.  They could help NASDAQ OMX, which has just begun moving into that area.  They have new datacenters with excess capacity, consulting capabilities through a partnership with CSTK, and the SFTI network.  Their marriage could upgrade global trading capabilities given their presence and position in the exchange technology market.  That said, it could also bring with it a slew of M&A as the combined exchange goes after more of the broker/dealer trading stack.

March Madness and Galleon

So… The prosecutor for Paris Hilton’s cocaine charges gets arrested for, well, cocaine possession. Senator Claire McCaskill, who sponsors a bill to fire federal employees who do not pay their taxes, gets caught not paying $287,273 in taxes. The SEC’s David Becker, in charge of recovering money for victims of Bernie Madoff, has a mother with a $2 million investment in Madoff.  As I sit here watching my bracket for March Madness, I’m wondering how many regulators and prosecutors for Galleon are watching their own brackets.  I’m wondering if any of them did their homework. How much effort did they put into achieving a competitive advantage?  How many Vegas odds makers are looking at travel schedules, injury reports, and scouting reports to determine odds and/or rankings?  How many Galleon prosecutors are making decisions based on those rankings?

A Moody’s analyst tells a junior analyst about m&a.  The junior analyst tells someone at Galleon.  And?  Galleon got the injury report and made the pick.  Pardon me for waxing on about expert networks and people doing homework. Certainly, there were some shady activities in that firm.  But really, where is the line between diligence and insider trading?  If a regulator’s brother was best friends with a pro scout for an NBA team, would the regulator pass up the opportunity for expert feedback on their picks if they were at the same party?

The attack on information flow is disturbing.  I wonder how long it will be before the SEC wants to see Google queries.  How long before they want to see if an analyst scraped Facebook postings of employees?  How long before you have to save your Twitter hashtag searches?  Ah well, I hope they do well on their pool. I’m running middle of the pack, but being a total homer with Wisconsin to win it.  I have inside information that it’s cold as hell in the winter and there’s nothing to do there but be indoors playing basketball or out on the ice on some lake waiting for ice fishing rig flags to pop up because I lived there.  Please, don’t arrest me if they win.

I Suck at Online Influence

I lead a panel at SXSW on black box trading on news last Sunday.  Prior to our panel, I spent the day in panels on qualifying and measuring online influencers.  In simplistic terms, sites like www.klout.com provide metrics on your overall online influence.  The panels offered an impressive array of statistics and visualizations tracing metrics on online influence from the innocuous (Bentley cars) to the false (Nelson Mandela is not dead).  My particular score… Well, not so good.

First off, let me say this.  Capital markets friends, we would be embarrassed to put slides up after these whiz kids.  They do data visualization in ways that are both cool and easy to understand.  We typically choose one or the other and rarely choose cool.

Second, they’re one track mind geniuses.  It took more effort to talk to some of these companies about their potential outside social media marketing than it was worth.  None of them attended my panel on black box trading on news, even the news aggregation companies.  Ah well.  It’s not my job to help them.  It’s my job to help you.

Saving the best for last, you need these people.  They are your future.  They can understand things about the companies you cover that you don’t know yet.  Do you know who is going to influence the opinion of the latest razor?  Who is going to impact perceptions of the next consumer electronics product?  These people know.  They can determine things they don’t know matter yet.  Who works for which companies?  Are those people happy at work?

As an asset manager, they could also become your worst nightmare.  Let’s say I know the 100 most influential people online about a company.  If you wanted to affect that company’s stock price, how much effort would it really be to pay some of those people to say very specific things about that company?  Think of it as a penny stock spam campaign on steroids.

I like talking about the potential of machine readable content in trading.  Signal generation.  Risk management.  News is an obvious first step, but are reporters the only trusted sources of information anymore?  In fact, how do you know reporters even tell the truth?  Tracking social media influencers could be the next wave of quantitative analytics.  We’re still in our infancy with news, but get ready to account for both news and influencer opinions about news as dual components in your signal generation.