What has been downloaded
As was told previously, Maurizio decided to download the last 100 posts for each of the 480 top accounts and used them to compare users. In retrospective this was probably not the best strategy to compare user data, but it was useful to guess how the Author rank may works.
what’s wrong with only 100 posts?
It’s difficult to do an accurate analysis when the original data is really sparse.
- some people did not have 100 posts.
- Five people had published the last post over 800 days before the data collection
- Two of users need only two days to write 100 posts.
- 101 people wrote less than 100 posts
- 7 people never wrote a post
It’s worth nothing again that the data collected is only the public part which the users decide to share. Even the 7 users with no public data are probably publishing something to private circles. One of those for example, publish the posts, but after a day delete them!
A grand total of 42.327 posts were collected
- Posts with only text: 3.667
- Posts with Link: 13.952
- Posts with Photo: 16.984
- Posts with Album: 2.632
- Posts with Video: 4.130
- Posts with Hangout: 47
- Posts with Audio: 34
- Posts with Event: 115
- Posts with Circle: 656
- Posts with Community: 24
- Posts with Docs: 24
- Posts with Check in: 2
- Posts with Maps: 6
- Posts with Market: 1
- Posts with Page: 17
- Posts with Place: 25
- Posts with Play: 1
- Posts with Reviews: 10
This categorization is not available through the api and has been devised using a customized tool.
Only 93 users write on average more than 160 characters per post.
This kind of data may explain the radical changes happened during the last months on Google+, with larger images and the disappearing description from an external link.
Google+ is now an hybrid between pinterest and twitter (which is probably influencing the usage of short texts).
All the posts generated 902.000 +1, 245.000 comments and 105.000 shares. On average there are 3,5 +1 for each comment and 9 +1 for each share.
We may say that the +1 is the most used social signal: it’s quick and effortless. It’s still useful because the post which has been plussed, will be shown in the circles of the person who has given the +1. There also an accumulation of social signals (+1, comments, shares): the receiver will be shown more often inside the stream of the giver.
Digging the data we are able to see that the most successful posts are “post with photos” and “post with album”.
We can also see interesting results regarding post with circles, but the number is artificially inflated, because it’s a group of people industrializing the process and putting strict rules too. The average shares is unnatural and the level of engagement become zero after the share is completed.
We consider this activity similar to low quality link building, where the number of followers is considered important like an high pagerank. Fortunately it’s quite easier to track this activities.
According to the table, the category with the highest level of engagement is the “VIPs”: this makes sense because it’s composed of a very limited number of people with a huge number of (fanatic) followers. One in particular is Beppe Grillo, which is now a politician with hundred thousands of followers. But considering the supposed amount of followers, those numbers are not really impressive.
The categories of photography and images have good results too: the similarities with pinterest are striking. Google+ has become an image-centric social network (expecially with the ease of upload and post process due to Android smartphones). Most of the people want easily-digested contents, funny pictures to share with friends or high-quality photos.
note i’m skipping a part about post by time. Sorry Maurizio, i think it’s boring. Let’s just summarize it by saying tht people usually publish contents between 12am and 4pm and with a with another spike around 8pm. If you want to see the graph, click here
the most successful post
In terms of engagement it’s a post by Elia Locardi, with just 3 words of text and an image.
If i don’t decide to strangle myself with my mouse cord, i’ll translate another part regarding a deeper analysis of the author graph.