If you like this post, or any of my work, please, nominate me for a Shorty Award.
The linguistic analysis engine behind TweetPsych has given me a bunch of cool data points to analyze, so I’ve begun to look at various factors and their relationship with follower counts. Using a database of over 30,000 accounts that have been analyzed with TweetPsych, the first dimension I’ve looked at is “Social Behavior”.
The “Social Behavior” category includes inclusive language like “we” and “you”, as well as language that describes relationships and communication. As it turns out, accounts with more followers, tended to be using more social language.
Over the next week or two, I’ll be posting about the rest of the dimensions TweetPsych analyzes and how they’re related to follower numbers, so stay tuned.
If you like this post, or any of my work, please, nominate me for a Shorty Award.
If you liked this post, don't forget to subscribe to my RSS feed or my email newsletter so you never miss the science.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 7:31 pm
Awesome graph Dan! I talk about engagement a lot and using social media to primarily build relationships and this just goes further to show that it does make a difference.
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 8:32 pm
Let's do the reverse experiment and use lots of corporate speak and buzzwords and watch the chart take a nose dive!
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 25th, 2010 at 10:19 pm
Wow – thank you for introducing me to the RID. Having problems finding the complete list of the words but hey, it's 2010. It should be out there. Somewhere.
January 26th, 2010 at 12:45 am
[...] The “Social Behavior” category includes inclusive language like “we” and “you”, as well as language that describes relationships and communication. As it turns out, accounts with more followers, tended to be using more social language. via danzarrella.com [...]
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 26th, 2010 at 11:34 am
Hi Dan,
Does it break it out by nouns & verbs?
I'd be interested to see if 'commands', instructions & other such constructs have higher clickthrus.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 27th, 2010 at 5:26 pm
Dan, I noticed the same relationship between a high score on “social behavior” and the number of followers when I first tried your excellent TweetPsych tool. But then I noticed that (a) Twitter accounts writing about social media tended to score high on “social behavior” too (maybe the word “social” works as well or better than “we” and “you”) and (b) those Twitter accounts (there are a lot of them) also had disproportionately high followers.
If you exclude Twitter accounts talking about social media, I wonder how powerful the “social behavior” score is as a predictor of follower volume. TweetPsych is cool. A TweetPsych that allowed one to control for topic matter would be even cooler.
January 28th, 2010 at 10:18 am
[...] up on my last post using TweetPsych Data, I looked at a metric opposing social behavior: self-reference. This time the [...]
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 28th, 2010 at 10:04 am
Hello. I think your e-mail form dose not work.
How I can send you a private message?
My e-mail vladislavkorobov@gmail.com
January 29th, 2010 at 2:18 pm
[...] you risk alienating your base. Here’s a great article on how the helpful effect actually boosts credibility in the social media sphere (via today’s excellent SmartBrief on Social Media, one of the must-reads in my crowded [...]
January 29th, 2010 at 5:02 pm
[...] Data Shows That Social Behavior Gets More Followers | Dan Zarrella – [...]
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 6:47 pm
This is incredibly interesting. I've been using tools like Twitalyzer to get similar information, but it's not nearly as involved as this. Have you worked with those types of groups to try and further assess the connections?
I also wonder what the tendency is with regard to WHEN people follow and/or unfollow a person on Twitter. For example, do people tend to gain more followers when a conversation is taking place, or do people tend to unfollow at that point in time? Or do they tend to follow people that are clearly “social” but they are not being social right then – perhaps in an attempt to be noticed in-between the conversations?
The person may have, in general, more followers if they are more conversational and “social” but I wonder if they lose followers at times when they are because those followers were expecting something different from the person and/or they were hoping to be personally engaged within a certain amount of time. I would be very interested in learning more about the psychology in those areas.
January 29th, 2010 at 11:49 pm
Dan, you have provided salient points that are very appreciated. We ought to keep the conversation going for all of us to gain benefit from this powerful social medium.
@FeliciaGriffin1
January 30th, 2010 at 1:52 am
[...] social media is like being a good friend By kahickey A post on Dan Zarrella’s site made me think about using social media properly. Yes, it’s a conversation, but you have to [...]
January 30th, 2010 at 4:44 pm
Hi Dan!
An interesting case of internet coincidence as I've just blogged on the same issue, from a different perspective – the balance between the use of the first and second person when writing articles. Perhaps we could collaborate on this; the software I use allows written material to be broken down in a lot of different ways. I'm not going to spam your site with a direct link to the article; email me if you'd like to take this further
February 1st, 2010 at 6:42 pm
I'm not sure the spread supports this conclusion. Especially in the second half of the data.
February 5th, 2010 at 7:00 am
its a good article on social media behaviour. it shows good analysis
February 10th, 2010 at 1:50 pm
[...] Want more followers? Stop talking about yourself Statistical analysis of Twitter users shows that users who tweet about themselves are less successful than those who engage with their community, writes social-media analyst Dan Zarrella. Users who referred to themselves with words such as “I” or “me” tended to have fewer followers than people who tweeted using inclusive language such as “you” or “we,” Zarrella found. DanZarrella.com [...]
February 11th, 2010 at 12:50 am
we liked this. we hope it works.
from,
just us
February 11th, 2010 at 10:48 am
[...] commentsComments Continuing my series of TweetPsych based data points, this is based on analysis of over 100,000 accounts and [...]
February 11th, 2010 at 11:23 am
I can understand the way you love social networking. I am a great fan of yours from now.
February 12th, 2010 at 5:26 pm
[...] Continuing my series of TweetPsych based data points, this is based on analysis of over 100,000 accounts and looks at the “Negative Remarks” category. Negative remarks include things like sadness, aggression, negative emotions and feelings, and morbid comments. [...]
March 10th, 2010 at 8:09 am
[...] most recent study examines how social behavior affects relationships on Twitter and certain activities contribute to [...]
March 10th, 2010 at 8:13 am
[...] most recent study examines how social behavior affects relationships on Twitter and certain activities contribute to [...]
March 11th, 2010 at 1:05 am
[...] other people. Twitter analysts like Dan Zarrella have noted that talking about yourself is a less-successful Twitter strategy than talking about others; now you see this truism remains ingrained in users’ behavior (and [...]
March 12th, 2010 at 12:37 am
Neat! Tested your TweetPsych and it's definitely something I think anyone working in social media would be interested in.
April 12th, 2010 at 10:59 pm
Nitpicking: Data is the plural of datum. Therefore, a datum “shows” but data “show.”
May 21st, 2010 at 9:58 pm
I am skeptical. Brick and mortar stores naturally have more followers, e.g. Burburry and Gucci. These are brands that people love, but really their Facebook pages are just PR platforms, rarely updated, and not that engaging. I really think there has to be a separate measure for well known brands and new brands.
June 27th, 2010 at 3:04 pm
[...] Be sociable. Social language correlates with more followers, Zarrella found. So use words like “you” and “we” and write more about relationships and communication. [...]
July 20th, 2010 at 4:24 pm
[...] iemand me twee weken geleden op twitter op de volgende blog: en daarachter aan las ik nog meer psycho-analyse van het [...]
August 6th, 2010 at 2:08 pm
thanks for the tips – great information, and completely relevant.
August 6th, 2010 at 2:09 pm
Thanks for the great tips! I will definitely be able to apply these to my work.