I was searching for a post using Twitter search and it just came to my mind, what algorithm the search can use to give quality results. As of now, Twitter finds keywords in the text of the most recent Tweet. Often, the search results are repetitive or duplicate and not much relevant to the keywords. Jayaram, VP, Operations for Twitter (an ex-Googler) hopes to change all that by taking Twitter search beyond this, by looking at the pages people are linking to within their Tweets. A Twitter crawler will visit the links, index the page and then correlate that 3rd-party content with the Tweet itself. This should help make for more relevant Twitter search results. This also made me think of more ways the algorithm can be set and I started comparing the organic search engine algorithms with Twitter search, and I have come up with these few assumptions.
- Follower counts: I guess follower counts do not really matter as an important factor to be considered in the search algorithm. It really does not prove, if you have thousands of followers, you contribute quality tweets. And, if I have very few followers, it is not that my tweets are quality tweets.
- Synergy of Tweets: I think the synergy of your tweets and that of your followers and that of your followers’ followers, should be considered as a factor. (In organic SEO - To have links on websites which have similar content and the link building strategy).
- Groupies: Say, most of the employees of a corporate are on Tweeter and they follow one another, then this really forms a closed group with known people together exchange good, informative and quality tweets.
- Tweet and URL Synergy: The synergy between the tweet and the web page that is linked in the tweet should matter. Just the case in PPC ads – the text ad and the landing page should talk the same thing.
- Frequency of Tweets: You already know the more you tweet, the lower the tweet goes on your follower’s account and the lesser the visibility it gets. And the tweet gets diluted in the number of tweets you have.
- Age of Tweet: The latest the tweet, the important is the update. But, how the tweet-bot will identify the spammy tweet to the original tweet is a big question.
- Tweet Favorite: It will really add value to the tweet, if the tweet has been marked as favorite by other folks. Bookmarking?
- Subscribing RSS: Yes, your tweets will be really important if you people subscribe them through RSS. People really will like you to know your updates instantaneously.
- Keyword Density: The density of the keyword among all the tweets over a period of time can be considered as a factor. You may know a lot of information about the keyword and are tweeting about the same and the updates can be informative. Or should it consider the density of the keyword in that particular tweet?
- Age of Tweet account: It does not matter when you created an account in twitter and when you have started tweeting. There can be many accounts which have been created and yet no tweets but, the handle can have followers. Tweets from such handles can’t be quality ones.
- Re-tweets: Just to spam and get more clicks, there will be people who just re-tweet the same thing again and again. So, this has to be take care of and should not be considered for ranking. Rather it should be considered as spam. Also, linking the same url to different tweets should be spamming.
- IP Address: It should note the no. of handles or IDs being logged in from the same IP address. People may have multiple IDs and would be using to promote their website using them. Such IDs need to be penalized.
- Geo-targeted: Track the IPs of the users they are tweeting from or logging from. And show the results to the members who are searching from that geo areas. This way the results are geo-targeted.
- Demographic profile: Twitter can have an advantage here over Google. They can target the ads or websites as per the demographic profile – age, gender, place, income etc.
These are just my assumptions and thinking. And I would like you to come up with your points and opinion as well to make this post very informative.
But one thing that is most important is, Twitter can be an enormous source of traffic to a website if it is handled well in the social media more so in Twitter.
2 comments:
Interesting. I think I would want to consider ReTweets in the ranking. But to use that as a consideration of quality rather than spam, one could focus on the number of people who Retweeted an item, rather than the number of times something was reTweeted. If more people passed it along, then it may be something of value.
What are your thoughts regarding #hashtags? I will often compare a plain key word search with a similar search using a #hashtag. For example if you compare a search for "recipe" with "#recipe" you'll find far more relevant Tweets with the latter. In that case, users have self-identified their Tweet as referring to a recipe so you are more likely to find actually things you can cook rather than mere references to the word.
Very true Heidi. I think you are right. Thanks
Post a Comment