 |
 |
 |
 |
| Vol.
8, No. 5, July 2009 |
To The Editor
| Subscribe | Back
Issues |
KDPaine.com | Masthead |
KDPaine's Blog | Reprint
Information | Search |
|||
by Chris Near, Director of Research, KDPaine & Partners
Whether you're an at-home startup or a global Fortune 500 company, brand monitoring is becoming a necessity. With Twitter quickly becoming one of the most popular media platforms in the world there is a lot of interest in seeing how one's company or brand is being mentioned there. The easiest tool to use for this is an online Twitter search engine. New technologies are making these tools more common and convenient, but, with so many options available, deciding which is best is not an easy task. For this article I selected five Twitter search engines, based on a very scientific process -- the first five new ones that came up in a Google search. I've analyzed each one in terms of accuracy, reach, and overall functionality. The short answer: If all you want to do is gather keyword or brand mentions from Twitter, then don't bother with these, just use Twitter's internal search engine at search.Twitter.com. Here are the details: BingTweets is Microsoft's mashup of their new Bing search engine and Twitter search. They claim that this search engine enables users to see "deeper" with real-time information into the hottest topics on Twitter and the web. Besides simultaneously pulling up a scrolling list of current tweets along with related Bing search results, the site also offers a window showing the most popular Twitter topics of the day:
BingTweets is far from exhaustive and comes off more as a side-note, with the true intent being to showcase Bing's search abilities. The site is good for getting a brief glimpse into the link between current tweets and Bing search results. It's got no advanced search options, sparse results, and only the ability to search back to midnight of the current day. Conclusion: BingTweets clearly is not designed to be a research tool, but is rather a simple way for Microsoft to jump on the Twitter bandwagon and gain some notoriety for their new Bing search engine. The newest trend in Twitter search engines is to search Twitter and other social media simultaneously. Scoopler keyword searches index Twitter, Flickr, Digg, Delicious, and others. The goal is to give users the most relevant results, updated in real time among a variety of sources. The Scoopler homepage shows you a history of your recent searches as well as a list of "hot" topics and lets you break it all down by video, links, images, or all content:
Sounds great in theory, but in practice the results are disappointing. A search using the word "@kdpaine," pulled up 10 tweets from the past two and a half days. The same search in search.Twitter.com, pulled up over 45 tweets from the past 24 hours alone. It had no advanced search option that I could find. Conclusion: The ability to search a number of social media types at once is nice, but if you're only pulling up a fraction of the relevant results, then what's the point? CrowdEye was created by former Microsoft search engine team leader Ken Moss and wife Becca. At first glance I really liked what this site offered. It lets you search tweets in real time within the past 3 days. You search by key term and it pulls up the related tweets as well as other online links, much like BingTweets (see above). It also builds an instant bar chart detailing the volume of tweets by hour containing your keyword. The site has a filter function and below the bar chart it shows a word cloud with the most popular Twitter topics:
This site was my favorite until I decided to compare their search results with those of other sites. Again I put in the keyword "@kdpaine." My bar chart showed that there were 10 tweets in the last 24 hours with that keyword. Scoopler only gave me 10 in the past 2 and a half days, so CrowdEye outperformed them. But it still wasn't anywhere close to search.Twitter.com. CrowdEye has a feedback link so I emailed them to see why my results were so varied. To his credit, Ken Moss got back to me almost immediately. Here's what he said:
Conclusion: This site has a lot of potential and I look forward to using it once their search stream is more comprehensive. This is one to keep your eye on. TweetFeel is the simplest looking of all the search engines. The front page reminds me of Google's homepage with just a box to enter your keyword and below that a few links to hot Twitter topics:
Of the five sites I researched for this article, this was the only one to offer sentiment analysis. (For more on Twitter sentiment analyzers, see my recent Measurement Standard article "5 Twitter Sentiment Analyzers Reviewed" for reviews of Twitter Search, Social Mention, Twendz, Twitter Sentiment, and Twitrratr.) I entered my keyword, "@kdpaine," and it came back with 18 results marked green or red depending on sentiment. For search that's OK, compared to the others reviewed here, but not close to search.Twitter.com. Interestingly, there is only positive and negative. No in-between for this site. The site doesn't show the date or time of the post so you have no idea how recent or old they are without opening the link. Every search also came with a disclaimer saying that these results may be incomplete due to too much traffic. A little worrisome; makes me wonder how I could ever be sure any of my searches were complete. One last note: The TweetFeel sentiment analysis is borderline ridiculous. First, they give everything either a positive or negative label. Can't some posts simply be neutral? Second, their sentiment analysis was highly inaccurate. In the small sample I looked at, more than half of them were wrongly categorized. For example, this tweet was labeled negative: "@kdpaine Wow, the Durham garden is exquisite! & your Berlin garden will be too (love me some walls!) What happens when it rains? Good work!" Conclusion: TweetFeel would be better off taking out the sentiment function until they work out the bugs. Comparing Topsy to the above search engines really isn't fair. It is not a search engine designed to search tweets, but rather it searches your keywords and ranks the results based on the number of tweets related to your subject, i.e., it is a traditional search engine powered by tweets. Topsy results are based on things that people link to when they are talking about your search term(s). Results can be sorted by month, week, day, hour, or all.
Conclusion: Topsy is targeted more towards searching all relevant conversations through links and keywords rather than searching streams of tweets. Although you can look at all of the related tweets and from there you can look at the links all of those tweets posted. Topsy also ranks Twitterers of high influence. Overall Conclusions These tools represent some very interesting technology being developed, and new ways to monitor your brand within Twitter and other media types. But if your purpose is strictly to gather keyword or brand mentions from Twitter, then I'm not sure that any of these tools would currently be that helpful since none of them came close to matching Twitter's internal search engine at search.Twitter.com. However, if you aren't as worried about accuracy and are focusing more on finding out overall tones and themes in Twitter and other media types, then sites like CrowdEye, BingTweets, and Scoopler could be of use. And
now, a word from our sponsor:
|
 |
Recent new articles:
Here is the book that gets people hired!
|
|
BingTweets
|
177 Main Street, Berlin, NH 03570 |
|||
