First up, Collective Intellect's Media Intelligence Service is a combination of software and service to help institutional investors, hedge fund managers and others keep an eye on the blogosphere and make sure they're not missing news about companies in which they've invested or are about to invest. "A lot of information is breaking first in new media like finance message boards and blogs," explains Darren Kelly, VP of Collective Intellect. Portfolio managers often don’t see such information until it's been sanitized and officially delivered through a wire or news service. The trick, with 58 million blogs out there, is to separate the wheat from the chaff, the credible bloggers from "splog" and people posting family news. The company's artificial intelligence engine (partially based on technology purchased from Britain's MI5 spy unit) ranks sources for credibility -- it determines the top experts on a subject (often starting with a list provided by a customer), then analyzes who they link to and who links to them, and who those linkees link to, to about 15 levels. "When 42,000 people have links to a person's blog, that person is not producing splog and is not an idiot," Kelly says. Then the engine categorizes the information in all the credible blogs by ticker and theme. It also extracts sentiment from the text -- early on the company fed the AI engine a training set of 100,000 messages identified as bullish or bearish, so it would learn to make that distinction. Finally, a team of about a dozen finance majors and journalists evaluate the results and tweak them as necessary. Results are delivered in the form of email alerts and web-based reports.

"What's important in blog search is to establish the validity of the information, and to extract sentiment based on paragraphs," says Susan Feldman, Research Vice President of Content Technologies at IDC. Collective Intellect certainly has tools designed to do this, but the only way to know if it would work for you is to try a demo version. The service costs $5,000 per person per year, with volume discounts. Other blog search tools exist, but no other that we know of is specifically tuned for the stock market.

Second is Edgar Online's I-Metrix 2.0 tool for analyzing detailed, XBRL-formatted SEC Edgar data (gathered from quarterly, annual, and preliminary earnings reports as well as institutional investors' 13Fs), to spot trends, opportunities and risks. For instance, a peer analysis model displays a set of competing companies and their income, earnings, profitability ratios, liquidity ratios and such, with exceptional and below-average values highlighted accordingly in green or red, for a quick snapshot of company performance. Users can drill down to the original reports to verify data cells and they can view any combination of comparative report data (for instance, if a company revised its earnings statement 10 times, you could see all the versions side by side and know at a glance what exactly was changed). The data can be presented in Excel sheets and through a Web interface that includes current stock price data (from IDC) and can be drilled down for profiles of all major holders of a stock. A new screening feature lets users request companies with revenues in a particular range or by other criteria. The content was recently expanded to cover not only all U.S. public companies and private companies with U.S. debt (including small and mid-cap companies), but 1,500 companies from two China exchanges as well. In the future Edgar Online plans to add Japanese companies to the mix.

Third is Standard & Poor's new Backtester tool. This software, based on S&P's Research Insight service and CompuStat fundamental securities data on 45,000 U.S. companies, lets portfolio managers test their investment strategies by setting certain parameters (such as, if a company falls out of the top 25 in the S&P or the stock decreases 20% in value, sell it) and then simulating how these assumptions would have performed over the past year. In addition to tracking overall performance, the tool provides transparency into every individual trade within the simulation, including information on taxes and transaction fees. So far, Backtester only works with CompuStat data on U.S. securities. Later this year the company plans to offer global data. The software competes with backtesting tools from Factset and Clarifi, which work with data from multiple sources.