I think everyone agrees that we need better Twitter analytics. The problem as I see it is that there is not one analytic solution that covers what I am seeking to be able to track. Some solutions cover pieces, but I don’t see a comprehensive solution. Plus, there are even some basic requirements that are not even being met at all.
Here are a few ideas that I’ve been kicking around for awhile. For beginners and most people, the basic requirements should meet the majority of your Twitter analytic needs.
Basic requirements – I want to be able to track on a weekly (or other self-defined time period) the following information. I want this information available over time in both raw form that I can analyze in a table and graphically so I can view trends:
- Number of tweets, @messages and direct messages
- Number of new following and followers for my account
- Number of times URLs in my tweets have been clicked
- Number of times my brand, competitor, twitter username or any other search term I choose has been mentioned
Advanced requirements – I would also like to see these advanced requirements met after my basic requirements are fulfilled.
- Number of positive, neutral or negative tweets for my brand, company, or other search phrase
- Number of times that I have interacted with other tweeters
- Analysis of my twitter influence and social capital
The rest of this post describes my key requirements along with various tools that meet some of my analytic needs.
1. Trends and counts with user definable dates just like Google Analytics Dashboard – At the top of the list, I want to easily be able to track the number of tweets on my account by day over a specific period of time. This is especially important when multiple people are tweeting on the same account as I want to know the basic metrics on my accounts. I also want to be able to put in a brand name or phrase and get a trend line and counts of mentions. This functionality must have user definable dates just like Google Analytics. Finally, I want this information in both graphic and tabular formats.
Tweetstats accomplishes some of this requirement, but only presents the information on a monthly basis instead of letting me pick my dates and periods. TweetStats also pulls together some other interesting information such as Tweet density, tweets by day of week and hour of the day, @replies and re-tweets. You can see all of their information by looking at my TweetStats page – http://tweetstats.com/graphs/tomhumbarger.
2. Twitter tasks like CoTweet – CoTweet sets the standard for making it easy for users to take action on tweets whether it is @replying, re-tweeting, direct messaging, assigning to someone for follow-up or emailing the tweet. CoTweet is especially useful when multiple people are tweeting on the same account. Right now, CoTweet does not provide me with any metrics on actions taken. And of course, I will need metrics to track these actions as defined in my basic requirements.
3. Easily embedded and tracking of shortened URLs like Hootsuite – The value of Twitter for me is reading suggestions from my followers and providing my own links. If you are leaving a lot of links, you will want to see others think is really important by tracking clicks on these embedded links. While CoTweet combined with bit.ly provides some of this functionality, HootSuite does a better job with this functionality. HootSuite also provides a mini-app called Hootlet which easily lets you create a tweet directly from a URL without having to open a separate program.
Hootsuite could do a better job by letting me download my stats to Excel so I could do my own analyses.
4. Pivot table like analysis of Tweets including drilldown like Excel – I feel that many of my basic Twitter analytic needs could be met by using Excel pivot tables. The trick would be to get a Twitter stream extract that could be easily parsed in Excel. At a minimum, this stream would separate the date, time, twitter user and message into distinct columns. Twitter needs to help here by making it easier for users to export, store and analyze tweets in an off-line environment.
TweetScan is one option for archiving all of your Twitter stream to an HTML file. A TweetScan costs $1.99 per account and is actually free if you tweet about their service. There are also other ways to archive your Twitter stream as discussed in a blog post I found by Maureen Pennock, and from another tool I found called Twitter to PDR Multi-archival-Webapp. These tools look to be built on the cheap and may be your only option until a company with a bigger budget steps up to the plate.
5. Analysis of Twitter mood, influence and social capital like Twitalyzer or Twinfluence – At the bottom of my wish list is the ability to analyze the mood, influence and social capital of tweets or any Twitter user. Essentially, the capability to analyze information in this manner starts out with the ability to do textual mining on a Twitter stream. In previous blog posts, I have looked at both of these tools but do not feel that they are ready for primetime yet. As noted above, it would be important to be able to track this information on a historical basis so trends can be identified. While the point in time information from these tools is nice, but it doesn’t tell the whole story that I want to know.
Twitalyzer is a unique tool to evaluate the activity of any Twitter user and report on relative influence, signal-to-noise ratio, generosity, velocity, clout, and other useful measures of success in social media. For example, here is my Twitalyzer Analysis and my original blog post can be found here.
There are others who care about Twitter Analytics too. For example, I just bumped into an interesting Slideshare presentation today on Twitter Analytics from Dr. Stephen Dann of the Australian National University.
So, Twitter and Twitter analytic providers…we’re waiting for you to truly meet our Twitter analytic needs!