Three start-ups attack Twitter Search

New companies are emerging to address the deficiencies in Twitter Search I covered last week.

This isn't so easy. The challenge for anyone trying to fix Twitter Search is that the service is extremely good at showing you what's happening on Twitter at any moment. If you add a ranking algorithm to Twitter search results, you risk burying that value. On the other hand, without some form of relevancy indicator or ranking, Twitter Search can get so noisy as to be unusable.

Here are three companies that apply some form of ranking to Twitter search results.

The newest--scheduled to launch at noon on Monday--is Twazzup. It solves the relevancy problem first by not messing with raw time-sorted Twitter search results. Those go to the main column. What shows up around them makes this tool great.

Twazzup doesn't mess with Twitter's time-sorted results. But the related links and lists around the main display add a lot of utility.

(Credit: Screenshot by Rafe Needleman/CNET)

First, you get a menu of related words and hashtags at the top of the results page. There's also an expandable "Popular Tweets" sidebar item that has the tweets on your keyword that are the most retweeted or linked to. There's also a list of "Trendmakers," people actively tweeting your keyword. And there are photos and popular links.

My preview of this product showed some performance hiccups, but the concept and presentation are very strong.

Tweefind gives you what looks like a standard Twitter Search (albeit in My Little Pony colors), but it layers a ranking algorithm into the display order.

Currently, it factors the number of followers and the number of retweets on an item in the ranking. Its results are better than the unfiltered Twitter Search, but I have the feeling, when looking at Tweefind results, that I might be missing items I'd like to see. That's the paradox with which all these tools have to deal. More on Mashable: Tweefind Applies Google Magic to Twitter Search.

Tweefind ranks results using a Google-like algorithm.

(Credit: Screenshot by Rafe Needleman/CNET)

Finally, there's Twitalyzer Search. It shows you raw, time-ordered Twitter results but adds two relevancy numbers to each tweet in the results list: The Influence (I don't understand exactly how this is calculated) and the number of followers the tweet's author has. It helps you find the items from the people to which other Twitter users are paying attention, and since the search order isn't modified, you don't feel like you might be missing other items.

Twitalyzer Search adds Influence and Followers data to standard Twitter Search results.

(Credit: Screenshot by Rafe Needleman/CNET)

I like Twazzup the most of these three products, since it presents a lot of useful information while maintaining the value of the raw Twitter search results. I do worry about the viability of all these experiments.

Twitter Search itself is going to get some sort of ranking technology eventually--possibly from one of these companies, possibly from Google or Microsoft--leaving the others stranded. Until then, try Twazzup.

[n3td3v]

It's only noisy if your search terms you use are too general. Try using terms that aren't so general.

[eltoro2827]

twitter is lame.

[EricTPeterson]

Rafe,

Thanks for including Twitalyzer SEARCH in your assessment of these alternative search engines for Twitter. The Search page on our site has a link to the definition of "influence" I have copied it into the bottom of this comment. In short, influence is the core summary metric at Twitalyzer, one that has helped nearly 70,000 Twitter users understand how they and their followers are using Twitter today. I'd be happy to discuss Twitalyzer in greater depth if you're interested.

Eric T. Peterson
Twitalyzer
http://twitalyzer.com

(The definition of "influence in Twitter" as measured by Twitalyzer)

As Twitter becomes increasingly important to online communication, the creators of the Twitalyzer believe that the need to measure the impact of our efforts in Twitter will increase a commensurate amount. While some believe that "popularity" is an appropriate measure of success, we disagree, eschewing this easily gamed metric in favor of something more robust, more fair, and more difficult to cheat.

The Twitalyzer solution is our measure of "influence in Twitter" calculated based on:

* Your relative reach in Twitter, measured by the number of followers you have
* Your relative authority, measured by the number of times you are "retweeted"
* Your relative generosity, measured by the number of times you "retweet" others
* Your relative clout, measured by the number of times you are referenced by others
* Your relative velocity, measured by the number of updates you publish over a seven day period

Each of these measures are weighted but otherwise the calculation is incredibly simple. We believe that what you get is a measure of success in Twitter that can be applied in a variety of ways.

There are a huge number of search apps for Twitter.

Look at http://twitdom.com/tag/search for over 30+ apps!!

[DanielBStern]

"I have the feeling, when looking at Tweefind results, that I might be missing items I'd like to see"

>> One way around this problem, a method which I haven't heard discussed or seen implemented anywhere, would be to allow the user to control the relevance threshold a Search Engine Result would have to pass in order to be displayed. This would also allow results to be displayed chronologically in the event that your query is 'time sensitive.'

The basic model would be a slider, or some other adjustable number display asking the user to filter results by relevance. The user, if they dont see what they are looking for at the 90% threshold could then say (through slider interaction) 'show me results that are above the 80% threshold. These could newly displayed results could possibly even be color coded differently, or have a relevance number displayed next to them

[tojosan]

Twazzup seems great. I enjoyed the feature you mentioned about highlighting certain folks, links and tweets. As well there were no dropped tweets.

Another plus is the tag suggestions at the top. Played with it and found it quite easy to get around overall and would recommend it.

[gculpin]

These engines are "OK", but depending on your needs, you might find MicroPlaza's tribal search engine much more useful and innovative, and yes I thoroughly believe it totally beats these other engines to the ground as it provides a search through the stories that people relay within Twitter, not within individual tweets themselves. It's basically "a search through Digg over Twitter".

E.g. You want to search for most talked about information sources regarding "Swine Flu" ? http://tr.im/jJMd
E.g. You want to search for most talked about articles from CNET.com ? http://tr.im/kJHU

Currently still under private invite though for more advanced functions such as OAuth pairing with your Twitter account to only get relevant stories from your network.
http://microplaza.com

[searchmotive]

I have thrown my hat into the ring of improved searched with my search service at searchmotive.com. I'd welcome feedback from anyone regarding what could be done to make the service more useful for all. Thanks!

- Jason Gregory
@jag959
searchmotive.com