Filtering v. Speed: The Simple Stuff

FirstRain argues that the web isn't really real-time since the bulk of newspaper content, for example, gets uploaded at midnight Eastern time. Exceptions to that rule are numerous, but other services focus on bringing you information fast. FirstRain focuses on bringing you information from around the web that you wouldn't have found otherwise. Blogs, SEC filings and news releases are all indexed by the service.

An engagement with FirstRain begins with users identifying what other research services they already use, whether that's Bloomberg or Google Alerts. FirstRain will automatically filter out any search results that your other services are already going to deliver to you.

FirstRain then finds out what companies and industries you're interested in. Natural language processing and analysis of word proximity is at the heart of the company's intellectual property, their ontology will discover relevant content on the web about more than 10k different companies, 85% of the publicly traded companies in the US and some international but relevant firms. It will also filter out content from "left-wing wackos," they told me, and you regular readers know how we feel about left-wing wackos here at RWW.

Below: From a daily update sent to customers, available by email or RSS. Click for full-screen image.

Exotic Filtering

The FirstRain dashboard looks like a startpage, your different companies and topics each have a widget that displays feed previews when you hover over headline links.

Two widget types provide a great example of what kind of sub-topics FirstRain discovers. The company offers one widget that displays quotes from executives in the companies you're tracking. Another widget uses Natural Language Processing and word proximity to pull out stories that contain executives saying anything that amounts to "no comment." You can imagine what percentage of those stories might be of interest to financial traders.

The FirstRain Dashboard. Click for full-screen.

Trend Crunching

FirstRain also processes the results it finds to pull out and display a number of trends. The primary one is around executive job changes. The management turnover widget expands into a full page that graphs out recent trends in executives coming and going from any company. In the image below the darker units denote turnover news discovered directly from the company and the lighter units denote news discovered elsewhere - like one company announcing that an exec from this company has been hired.

The Management Turnover screen, just one of several trends analysis options. Click for full-screen.

Hover over any point in the graph and you can see a detailed view of just who changed jobs, with links to the full story describing the move. FirstRain can track the same trends in things like product inventory, all gleaned from official and unofficial stories on the web. Cross reference items like this and you've got a whole new ability to ask informed questions of any of these companies.

I Want

I love seeing what kinds of tools people who trade in information on high levels have access to. My research workflow, which I wrote about here a year and a half ago, could use a big upgrade. FirstRain is a real inspiration.

There's not much in the consumer space, or hell for less than $10k/year, that I've seen that could compare to this though. RSS filtering is really simple, Yahoo! Pipes is probably the least so. There's no algorithms there, though, just the ability to filter for words in various fields. You can work some serious magic with Dapper (our demo) and I'm excited to see what Toluu can do (our coverage). For company tracking the most comparable thing I've seen might be RivalMap, which I recommend to any startups working in competitive market verticals. FirstRain only tracks publicly traded companies, something for startups, like Crunchbase with smart, automatic updates would be great. Nobody in the web world is doing many of the far-out things that a service like FirstRain can do.

Part of the problem may be the widespread "it has to be free" sentiment we see in the consumer market. I may not feel like I can spend $10k on a product like this, but I would definitely pay a hell of a lot more for something that can do some of the same kinds of things than I would for, say a premium Flickr account. Developers, please, build research tools I can pay you for!

This is great inspiration to spend some time fine-tuning the research flow, but it's also reason to stay tuned to the bleeding edge of the internet. Once semantic markup is widespread, determining what companies are most relevant to a particular search result and what the real topic of articles are should be a lot easier. Many times enterprise technology learns from consumer technology - FirstRain's indexing of blogs is probably one example of that. In this case, though, I'd like to see the consumer market do some learning from high-end research tools aimed at the enterprise. The potential is really exciting.

Meanwhile, if you're a FirstRain customer - have a great time. Drop a line here in comments if you would and let us know how it's working for you. All of us have a vested interest in seeing the web research field move forward.