Previous  |  Next


Jan 2

Tim O'Reilly

Tim O'Reilly

Predicting Financial Market Performance with Real Time Web Data

A few days ago, I wrote that one lesson from Wall Street suggests that Google and other Web 2.0 giants will increasingly start "trading for their own account." The same history of Wall Street firms suggests that there will be many new opportunities for specialized information services that supplement the services that are no longer provided by the search engines themselves.

Let me begin by repeating part of the "unorthodox history of Wall Street" according to Bill Janeway, as reported in the Release 2.0 issue on Web 2.0 and Wall Street:

Thirty years ago, Bill points out that the price of a trade was regulated by the exchanges: it cost approximately 22 cents to trade a share on the New York Stock Exchange. Unable to compete on price, firms competed on the quality of their investment research, and brokers' relations to clients were based on the information and insights they could provide.

Once the exchanges no longer regulated the price of a trade, prices fell over time to current levels of a fraction of a cent per share, or for large trades, effectively zero. As a result, sell-side firms could no longer afford to do fundamental research. Two things happened: independent research firms grew up that charge directly for research, and more importantly, firms began to trade against their clients for their own account, such that now, the direct investment activities of a firm like Goldman Sachs dwarf their activities on behalf of outside customers.

In that issue of Release 2.0, I pointed out how search engine marketing firms are gradually morphing into the Gartners and IDCs of the search engine economy. Not only that, companies that mine web statistics, from Comscore to Hitwise have been scoring significant paydays. (Comscore's stock is up 41% since its IPO, and Hitwise was recently sold to Experian for $240 million. Hmm...I wonder when Amazon will think to spin Alexa back out as an independent company?) In short, there is a new economy of specialized trend tracking and data mining companies that is growing up in the shadow of the search engines and other Web 2.0 applications.

In an interesting coincidence, many of these players are also becoming of interest to Wall Street because the data they gather may turn out to be a mother-lode of information on the future performance of individual companies and the economy. We'll be featuring a number of these companies at the Money:Tech Conference in New York next month, including:

  • DataUnison, whose analysis products were originally developed to help eBay powersellers, is now becoming more widely used by anyone who wants an inside peek at aftermarket demand for products. This, of course, can tell you whether a new product introduction was a winner or a dud, and may thus shed light on the upcoming financials of said company (and its suppliers).

  • Altos Research, a company that originally developed real estate market analytics for brokers, home sellers and buyers, has discovered that their data is equally interesting to financial institutions. From their website:
    "For investors and traders, Altos Research publishes the Altos Real Time Price Index(TM). The Index monitors properties in the same metropolitan statistical areas (MSA) as those captured in the S&P Case Shiller Housing Price Index (CSI) and the Radar Logic RPX upon which the Chicago Mercantile Exchange (CME) bases its housing futures markets. The result is an analysis and index value that is highly correlated to the CSI value released 12 weeks later. [Italics mine]
    altos vs. case-schiller
  • Connotate, unlike the two companies above, provides a generalized platform for data mining topics on the internet, from news sources to blogs to gossip on mailing lists and forums, as well as regulatory filings and other internal and external company data.

The point: the real-time nature of internet data means that investors no longer need to rely on either official government data or quarterly company filings, which are delayed by months. I've written before that real time is a big part of what makes Google different from most corporate back office data mining. These companies provide real-time data that was never accessible before.

As both Wall Street firms and Web 2.0 giants leverage their scale and access to information to extract more of the total value from the market, innovative companies are democratizing the tools of analysis and insight.

tags:   | comments: 7   | Sphere It


1 TrackBacks

TrackBack URL for this entry:

» Getting Some Attention! from Altos Research Real Estate Insights

We've had some really nice attention in the press and blogosphere the past week or so. Here's a quick summary so you can see what people are saying about Altos.O'Reilly Radar: Tim O'Reilly highlights our real-time data products for Wall Street and the h Read More

Jim S   [01.02.08 10:20 AM]

I've been wondering for a while ( if it would be possible to use the search term / expression of interest event stream as a leading indicator for security pricing. Maybe the best place to start a hedge fund right now would be inside Google.

Alex Tolley   [01.02.08 12:56 PM]

I doubt that the Altos info offers anything that hasn't already been arbitraged away in the futures and options markets, otherwise we are really talking about $20 bills in sidewalk gutters.

This does not detract from your larger point that information is becoming more democratized and having an impact on firms.

My favorite story in this regard is when I worked for Dow Jones Markets, in the late 1990's. They had a proprietary data feed to distribute price and page data across the US and indeed, even globally to some extent. At the time, we were just seeing the early, 20 minute delayed pricing via Yahoo. The senior execs did not believe that the internet could ever compete with their expensive, proprietary communications system and pooh-poohed any thoughts about harnessing the web. Within a few years, Dow Jones divested itself of the division, partly in the face of burgeoning low price competition from players using the internet to distribute prices.

However, the real problem for firms is being able to charge for research in the face of low cost alternatives. Whilst you are correct that independent analysis services do exist, there is a plethora of also-rans that barely survive. The reason is quite simple. The reasrch that is done in the brokerage houses, bad as it is, is effectively gratis to the fund management firms. They drown in "free" information, mostly useless. The research is maintained, not because it is good, but because it generates "stories" for the brokers to call customers. Even if the the trades are break even in value, the trade volumes drive price changes that proprietary trading can capitalize on. Note that Goldman Sachs just reported outstanding profits despite their huge sub-prime business. Were they trading against their buyers?

So while there will be opportunities to make money on information asymmetries, I think these periods will become ever shorter for any given "product" and that the logic of costless information replication will drive down the value to zero.

The value will accrue not to the information, but to the temporary attention of the viewer. Thus the logic of using advertising to create value, rather than the information content itself.

The parallels to brokerage research driving sales stories is fairly self evident.

bex   [01.02.08 04:18 PM]

I sense danger...

There's a HUGE advantage in having these statistics delayed by months or more... it has a stabilizing effect on the economy. By the time you get the info, insiders have had it for a while now, and your enthusiasm will be tempered.

If I get this info immediately, it helps me... but what happens if EVERYONE gets it?

You think the stock market is chaotic now, just wait until 10,000 simultaneous SELL orders get executed in the same nanosecond. A casual diversification would lead to instant market panic.

Unless we also get instant feedback on WHY people made a specific sale

Alex Tolley   [01.02.08 07:10 PM]

bex - I disagree with you. The danger is when there are a few players all making the same decision. In reality, players mostly make different decisions based on data. Occasionally this fails when they all use the same decision making model - a good example being portfolio insurance in teh 1987 stiockmarket crash.

P-Air   [01.02.08 11:59 PM]

Two other companies I'd include on this list of innovative uses of real-time Web/Internet data would be Majestic Research and Skygrid, the latter of which I believe is presenting at the Money:Tech conference.

bex   [01.04.08 07:59 AM]

Alex - every benefit in a dynamic system has unintended negative consequences.

I predict that as real-time data becomes more available, stock markets will fluctuate much more wildly...

There are alternatives... wise investors might watch the real-time data, anticipate an irrational swing, and bet *against* it. Even wiser investors may time their trades to bet *against the wise*

Of course, I have zero data to back this up... but I'd be willing to bet that the more real-time data we get, the more wildly the market will fluctuate.

Cody   [01.09.08 11:41 AM]

I believe that more traders and more information will REDUCE fluctuations. In a perfectly efficient market, there would be equal pressure on both buy side and sell side causing the price to stabilize at a security's "true value". The primary reason we do not have a perfectly efficient market today is information asymmetry. As more data points from more sources become available, information asymmetry will be reduced and markets will continue to stabilize.

Post A Comment:

 (please be patient, comments may take awhile to post)

Remember Me?

Subscribe to this Site

Radar RSS feed