This is insightful, paidcontent.org has added a twitter summary of the article you are reading at the head of the article. This allows you to tweet it without the 140 character dance that summarising an article usually involves. It will encourage lazy tweets however, as some people will not bother reading the article in full, preferring to simply tweet it to their followers instead. I wonder if it will even be possible to track a combination of time on site (ie, time theoretically spent reading the article) alongside the nature of the tweet that followed so the apparent reading of the article?
A user spends a mere 30 seconds on a long article followed by an automated tweet that simply repeats the offered twitter summary,
A user has spent time on the page, possibly with page clicks and other metrics of interaction tracked, followed by a personalised tweet on the subject.
Scenario 2 would be more ‘valuable’
You could set this up comparatively easy in Google Analytics, with the virtual page views feature (to track the interaction portion of this scenario) combined with that person’s twitter feed. Assuming they reference your article URL, you should be able to pull in this data into excel and, frankly, have a party with it all.
You would have to factor in average reading times in relation to the number of words of the article but that’s about as complicated as it would get. My knowledge of Google Analytics is entry-level at this point but even a newbie like me could set this up… I wonder if any publishers are thinking about their data in this way? We could even devise a ranking system for the tweets themselves, factoring in the time on page and tweet content metrics I mentioned above, multiplied by that person’s twitter ‘influence’. Of course, the issue of influence in twitter is itself something of a thorny subject, but, perhaps worth investigating all the same.
[Time on Site (probably factoring in other signs of interaction yet to be defined) *
unique tweet (this will have to be a binary number, 5 for a unique tweet, 1 for an automated tweet)
* user’s twitter influence]
Project! I had better start talking to the staff Data Scientist whiz kid.