Danny Sullivan kicked off the session by defining real-time search in his view. He doesn’t consider new articles found on sites like Google News, Technorati or Digg to be “real-time” because the content isn’t necessarily brand new (meaning it wasn’t created, posted and available online within seconds). Danny considers being able to see a funny sign, snap a photo and upload it for instant availability; microblogging, etc., to be real-time because the creation, publishing and “findability” of the content is virtually instant. (The panelists didn’t always agree with this definition - this is clearly still a very emerging concept.)

While Twitter is the current poster child for real-time search, many of the speakers listed disadvantages to searching Twitter for breaking news or hot topics, such as:

You’re only searching content on Twitter, not all content that’s available on the web. (A panelist from Google hit this point hard.)

The ranking algorithm is chronological, with the newest post (not necessarily the most relevant) at the top.

Twitter can be a firehose - tens of thousands of tweets can be posted in a short amount of time for breaking news and there’s no good way to sort through or summarize them. Ironically, the more posts on a particular topic, the harder it is to tell what’s really happening - Twitter doesn’t scale well and its search function works better for less popular topics and hashtags.

Twitter is prone to spam, particularly when a certain hashtag begins being used frequently - spammers attach it to their tweets on other topics and link to their own sites.

Older tweets either disappear or it’s cumbersome to go back in time to see what the discussion used to be on a certain topic. It’s hard to see how the conversation changes or trends over time.

Tweets are limited to 140 characters, so they don’t include a lot of information. Many link to other web pages, so you have to go to multiple sites to read and absorb the content.

Fortunately, several third-party companies are being launched to sort through Twitter’s data and make it more useful. I’m not sure I got all the details right, so please double check these sites for details:

OneRiot: I didn’t get a lot of detail about how their system works, but the goal is to consolidate real-time search results in such a user-friendly way that your mom would find it valuable, as would people who don’t know what microblogging is.

CrowdEye: This presenter gave the best demo. This site displays a bar chart showing how often topics are tweeted over time - click and drill down and see what was being said months ago. Sort results by relevance so that the tweets from the most trusted/authoritative Twitterers appear at the top (they’re rated by number of followers). It crawls the links included in tweets to put the most-cited web pages in order. There are personalized suggestions for people you should follow based on people you’re following now, and several other enticing features.

Collecta: Here, the real-time query stream is analyzed so you can see top searches (and more, I believe, but I wasn’t totally clear on what).

Topsy: This site indexes every tweet about a particular link, which sounds like it’d be good to determine how viral a piece of new content has gone. It ranks web pages by the number of tweets and allows you to go back in time.

I haven’t had a chance to play with these new tools yet, but I’m looking forward to doing so, not only from a consumer point of view, but to figure out how marketers can appropriately capitalize upon real-time search. There’s no question that searchers aren’t only going to search engines when they have questions anymore and it’s important to ensure your brand is visible everywhere that people are going to seek answers.