Dennis Howlett had a follow-up discussion with Jeff Whitehead and shares more insight into the power of precision persistent search.
Yesterday I had a conversation with Jeff Whitehead, CEO of RealTimeMatrix. Jeff is a propeller head but he keeps it just dumbed down enough for me to get it. His company has built a really smart real-time preference based search and deliver tool. With current search a la Google, once you get half way down the page of search results, it falls apart and the cruff level rises dramatically. Anyone disagree?
RTM overcomes that by serving a persistent query that is able to correlate your preferences across multiple descriptive dimensions. On a test of 20,000 search terms run across a million search items, RTM turned up 13 responses. All were highly relevant to the original query. RTMs accuracy levels are an order of magnitude better than Google. In the corporate world, that matters - hugely. Relevant information is not only more valuable but the times savings in research are enormous.
As I’ve written before, RTM has teamed up with Attensa, which understands corporate demands for RSS readers very well. However, RTM isn’t limited to the deep pockets of large enterprise. It is developing very smart widgets that you can use to deliver highly personalised work related tasks. RTM will immediately straddle the market so that SMBs will find the offering affordable. And useful. They are a little way off :”Q1″ is the word, but they’ll be an exciting step forward with much more to come.
TrackBack
TrackBack URL for this entry:
http://www.attensa.com/cgi-bin/mt/mt-tb.cgi/262
Comments (3)
So RTM takes a user's query and then runs multiple versions of it, search terms, against the item data. It then compares the results of each run and displays the results that appear the most?
Sounds interesting and in a way it mimics, and automates, what power users of Google already do (try variants of their original query for the best results.) My only concern is 1 million items is a fraction of the web. I can see why it is an enterprise product at this stage. Hopefully they can scale it up to the web though.
There are some subtleties here:
Google, Yahoo, MSN, Technorati and other search engines use indexing to "catalog" web and blog content and rank content based on popularity of information that has already been published. Honing your static search by entering variations of your search parameters only returns a snapshot view of web content available for that moment the search is excuted.
The Real Time Matrix doesn't use indexing in its search technology. It uses matching and correlation techniqes to return results as they are published in specific "streams or rivers of content." Information on your search is constantly updated as new material matching your search parameters is published online.
The concept is to consistently return high quality matches on new information from specialty information channels. Business and other large organizations spend lots of money for access to critical dynamic information provided by premium content channels including Moreover, Factiva, Lexis-Nexis, Westlaw and other publishers and aggregators. RTM and Attensa are providing tools to deliver highly relevant content matching precisely honed search terms. It's more about matching than the popularity indexing used by traditional search engines.
Paul, Great questions. Allow me to clarify:
- As Scott points out, RTM actually "persists" your preferences in memory, then correlates them with the live data that is streaming across the web. This, non-index approach to matching enables us to look at the "entire" stream of content and not miss anything. Furthermore, the process eliminates latency so the matched content is re-broadcast to you the instant it is observed.
- As for the volume, we are currently processing just under 1 million new, live content items per day which are of "quality." What this means is that of the many millions of URL's being observed, we are currently processing news and blog items that pass an initial aggregation test. To our knowledge, we have the most comprehensive news and blog sources available and are now preparing to add listings to the mix.
Subscribe to this blog's feed
