Khoros Atlas Logo

Are Your Facebook Fans Real Fans?

Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Dr Michael WuMichael Wu, Ph.D. is 927iC9C1FD6224627807Lithium's Principal Scientist of Analytics, digging into the complex dynamics of social interaction and group behavior in online communities and social networks.


Michael was voted a 2010 Influential Leader by CRM Magazine for his work on predictive social analytics and its application to Social CRM.He's a regular blogger on the Lithosphere's Building Community blog and previously wrote in the Analytic Science blog. You can follow him on Twitter or Google+.



Hi there! This is the third article in this mini-series on Facebook Engagement, where I develop the deeper levels of engagement metrics that quantifies the full spectrum of engagement. I apologize that these posts tend to be a bit more data intensive, but these eight levels of metrics are necessary for the development of the Facebook Engagement Index, which I will reveal soon. If you missed the earlier posts in this series, I recommend reviewing them before proceeding.


  1. Quantifying Facebook Engagement: More than Just Counting Fans and Likes
  2. Deeper Facebook Engagement: Dissecting Interactivity


Last time I covered the intermediate level (i.e. Level 3 to Level 5) engagement metrics, which are all metrics that quantifies how fans interact on the fan page.

  • Level 3: Thread Depth – amount of interaction
  • Level 4: Unique fans per conversation – with how many other fans?
  • Level 5: Average/median response time – dynamics of interaction

This time we will investigate the next two levels of the engagement spectrum (i.e. Level 6 and Level 7). We will also re-examine the meaning of fans and what it means to be a Facebook fan vs. a real fan in the traditional sense.


Are Your Fans Truly Loyal Fans?

If you are a true fan of say, a certain celebrity or sports star, would you continue to participate in conversations and/or activities about that person? My guess is “yes,” because that is what it takes to be a fan. You need to continue to participate in all the activities that other fans do in order to be a part of that fandom. And true fans often take pride in their participation. If you slow down or stop participating, then technically you are no longer a fan anymore.


fig06_Engagement Data_web.gifSo far, we’ve been looking at the statistics of the active fans who are participating in conversations. As a scientist, I am rather curious to see how many of them are truly loyal fans and actually return to the fan page and remain active over time. To answer this question, I computed the active fans’ re-engagement probability with the fan page. That is, the probability that a fan becomes active (i.e. posts something) more than once on the same fan page (see Figure 6).


According to the data, most Facebook fans are not very loyal to the fan pages. Only about 30% of the active fans re-engage with the fan page more than once (i.e. through posting). 70% of the active fans will post only once and never re-engage the fan page again!


So, most Facebook fans are not real fans in the traditional sense, because they are not particularly loyal. Facebook skewed the meaning of fans by calling any user, regardless of how passionate he is, a fan. If that is the case, we will need another term for those users who are true fans of a brand. They do exist, albeit in much smaller quantity. And we will call them the superfans.


Do Fans Develop Relationships on Fan Pages?

In the first article of this mini-series, I demonstrated the structural similarity between a fan page and a community. In my mini-series on cyber anthropology, we also established that community is where relationships are established and developed. Since a fan page is structurally a community, I wondered if it also functions like a community (i.e. enables the development of relationships).


In order to build relationships, people must engage each other actively and participate reciprocally in conversations. To see if the fans are doing this, I computed the active fans’ return probability to the same conversation. That is, the probability that a fan posts something more than once within the same thread of conversation (see Figure 7). Note that this is a very lenient criterion for having a conversation. In reality, it would probably take many more posts within a thread to have a meaningful conversation. And it would take many such threads of conversation to contribute significantly to the development of real relationship. But I am giving the “so-called” fans here the benefit of the doubt and see what the numbers tell us.


fig07_Engagement Data_web.gifTo my surprise, most fans do not participate in reciprocal conversations. The probability of a fan returning to the same conversation is very low, only about 9.6%. That means 90.4% of the active fans post only once in a thread and never return to that conversation again. This fan behavior would definitely hinder the development of any meaningful relationship on fan pages.


Therefore, even though a Facebook fan page is structurally like a community, it doesn’t function like one, because the fan page environment is not conducive for the development of relationships. I will speculate here about why this is the case, since I don’t have any hard data to prove it. I suspect one of the major reasons is due to the stream interaction style of the wall post. Streams are great for news feeds, status updates, or any content that is broadcasted for mass consumption. Users typically interact with the stream by watching its content, and they can interact with any piece of content at the moment when it flows by.


However, this type of stream interaction is terrible for carrying on a long conversation. Because as you are having a conversation, that conversation is also being pushed down the stream by new content. Eventually, the conversation will be pushed off the wall as the natural flow of the stream. Ironically, the more people participate, the faster the stream flows. So the busier the fan page is, the faster any conversation flows off the wall, and the harder it is for you to hold a meaningful conversation. This is precisely why it is so hard to develop any relationships on fan pages, because the streaming wall posts made it very difficult to have any meaningful conversation on the fan page.



Engagement_Depth_04_resize.gifToday I presented data on the Level 6, and Level 7 engagement metrics. The data suggests that Facebook fans are not really fans in the traditional sense because they do not actively re-engage the fan page. Despite that, Facebook still calls them a fan. Moreover, the Level 7 data also confirmed our suspicion in the previous post that fans typically only post once within any conversation. This makes it very hard to carry on a meaningful conversation with other fans and therefore hinders the development of any real relationships on fan pages.


With levels 6 and 7, we are almost done characterizing the full spectrum of engagement.

  • Level 0: Total fan counts
  • Level 1: Active fans
  • Level 2: Interactivity through comments
  • Level 3: Thread Depth – amount of interaction
  • Level 4: Unique fans per conversation – who you interact with
  • Level 5: Average response time – dynamics of interaction
  • Level 6: Re-engagement probability with the fan page
  • Level 7: Return probability to the same conversation
  • Level 8: Dare to speculate what this level might be?


Next time, we will cover the last and the deepest level of engagement metric. If you like to speculate, I’ll give you a hint. The eighth level of engagement deserves is own post because it is a network-based metric. The data there is interesting, but given what we’ve already learn about how fan pages work, it is not a big surprise. So stay tuned for the deepest level of engagement!


In the mean time, I welcome kudos, comments, critiques, suggestions, tweets, retweets, debates, or discussion of any kind.  🙂  So, let’s get some discussion going before I see you again next time.





About the Author
Dr. Michael Wu was the Chief Scientist at Lithium Technologies from 2008 until 2018, where he applied data-driven methodologies to investigate and understand the social web. Michael developed many predictive social analytics with actionable insights. His R&D work won him the recognition as a 2010 Influential Leader by CRM Magazine. His insights are made accessible through “The Science of Social,” and “The Science of Social 2”—two easy-reading e-books for business audience. Prior to industry, Michael received his Ph.D. from UC Berkeley’s Biophysics program, where he also received his triple major undergraduate degree in Applied Math, Physics, and Molecular & Cell Biology.
Olli Parviainen
Not applicable

Good post again! I particularly like that the metrics are transparent, unlike some other "metric" driven utilities.


One question though. You have focused on wallposts and comments, how about the likes of a comment or a wallpost? Liking is hallmark of Facebook and provides a good network metric opportunity. True, all of the "likes" don't carry a timestamp and are harder to assimilate to your previous posts' metrics.


Speculation for the 8th level: do people like comments reciprocically and how often they are friends?


-Olli (

Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Olli,


Thank you for stopping by and commenting.


I'm glad you like the metric and the transparency. I guess that is probably due to the more academic side of me.


You posed a great question. The data I used in this analysis consist of wall posts and their comments. So comments are included. Maybe I didn't make it explicit in this post, but if you followed the 2 post before this one. It will be obvious that I included them. However, I have not incorporated the data on "Likes" yet. That will come later. So stay tuned...


Good speculation on the 8th level. You definitely got the reciprocity there. Come back next time to find out wht is the deepest level of engagement metric.


Thanks again for stopping by and for asking an excellent question.


Rawn Shah
Not applicable

Hi Michael,


I see a lot of analysis of behavior but not much or any of attitude of the members. What you seem to be saying here is that fandom is based purely on activity alone, and not their mindset. It seems very left-brained, and leaving out the aspect of what makes fans fans: their feelings on the subject. 





Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Rawn,


Glad to see you here. Thank you for stopping by and commenting.


Yes, there are definitely a lot more analysis on user behavior than on user attitude. That is not surprising for 2 reasons.


A1. Behavior data are easier to obtain and these data are usually cleaner than data on attitude. Data on attitude are not only hard to obtain, they are also noisier and less accurate. That is because attitude data, such as sentiment, often need to be inferred. To get cleaner data on user attitude, survey is still the way to go, but it is not easy to survey web scale population.


A2. It is behavior that drives the economy. It is the actual action of purchasing, referring, or staying loyal by continually using the product that drives the ROI. Not what people feel about a product. That being said, I suspect that attitude is probably leading indicator of these action. However, due to A1, these leading indicators are not very accurate and therefore not very predictive and thus not very valuable yet.


And I do believe that fandom, at least in the traditional sense of the word, should be based on activity (the actions they took). That is what fans do. Fans (of some celebrity) are called fans because they not only like a celebrity, but are willing to go places and do crazy things for that celebrity to identify with him/her. Of course attitude matters, they should have positive sentiment while having these activity. But just having a positive sentiment and liking something is, IMHO, not sufficient to be a fan, at least in the traditional sense, which is defined by as “an enthusiastic devotee.”

Thank you so much for commenting on my blog. Hope to see you here more often.


Frequent Commentator
Frequent Commentator

Hi Michael,

This was an interesting read worth some contemplation. 



One thing I wanted to ask you was if you have any data that would factor in the quality of the fan pages (new posts, photos, videos etc.)? I am wondering if your "Re-engagement probability" would vary significantly with those factors.

I am not surprised as much with the 'Probability of Conversation' metric. Facebook in general is not a conversation platform. And one can argue that Facebook Fan pages are different but in terms of interactions I think they are similar (In this I agree with you that FB fan pages don't behave like a typical community).  In most cases, I visualize the FB Fan pages as a panormic view as seen when you are riding in a car along with other folks who are looking at the same view (riding in different cars). In most instances, the interaction spurt is a snapshot thought output and not a prelude to a conversation.


In your last post, we had an exchange about looking at the 'quantity' of interaction and the challenges of avoiding gaming and deciding what to measure. I agreed with your viewpoints (and still do) but still find myself going back to somehow categorizing the interactions. As an example, when I look at many of the fan page comments many fall into what I call  'interjections' & 'emotional ejaculationss' and relatively few are complete sentences. Something like:


Can't wait


Whoo hoo!

Love it.

That is good news...etc.


While these are "interactions", are these really interactions? My thoughts on this might be not be too popular :-). Btw, interestingly I think negative emotions generate better interactions that positive ones (statement made without any empirical analysis). 


Last couple of points. Do you have any computation on how often a "fan" both Likes & Comments on a post? From a business perspective, would be interesting to correlate this with the type of post etc.  Also, is there a way for businesses to leverage the number of friends a fan might have - in other words do we see any value of bringing that into the mix?

Was a good read Michael. 




Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

 Hello Ned, 


Thank you for the complement. Very glad to hear that you find this post an interesting read. 


You are totally right. Our data do show a lot of short interjections and short spur of emotions that does not have much content. And I totally agree that these comments should be weighted down and probably relatively less with respect to the interaction between fans. However, to do this at web scale, it would require some engineering. We would have to compute a new word-counts feature for every single messages (please excuse the machine learning jargon), before we can use it, because word count is not something that you get for free from Facebook. It is definitely possible! Maybe I will do that as an incremental improvement to the algorithm. 


The ability of negative comment to engage people is over rated. Quite a while ago, I’ve look at how negative comments spread. What I’ve found is that negative comments tend to create a bigger spike, but they are more transient. Positive comments typically do not trigger a big spike of activity, but last much longer. However, the data I look at was not statistically significant enough to conclude either way. I also have a theory that maybe it is the contrarian view that triggers a big, but transient spike of activity. So if the world is full of negative view, then the rare positive comments would behave much like what we see now for a negative comment. So it's possible that negative comments only trigger a big-transient spike simply because they are rare, and not because they are negative. 


We do have Facebook “Likes” data and it is not very difficult to compute the frequency of both “Likes” and comment. I am planning to integrate “Likes” into the algorithm for the next version of the Facebook Engagement Index. For version 1 that plan to roll out soon, it will only consist of commenting activities. As you can see, the 8 levels of engagements metrics is already pretty data intensive.  🙂 


However, I do not believe that the number-of-friends a fan has should be factor into the mix of these levels of engagement metrics. Number-of-friends is a connection metric (as opposed to an interaction metric. See No Game, No Gain: Realizing the ROI of Your Facebook Fans). That means friendships are maintained by Facebook indefinitely. It doesn’t require any action or effort from the user’s part to maintain that friendship on Facebook. For that reason, it is a poor metric for engagement, interaction, or pretty much anything time sensitive. On the other hand, frequency of communication with friend would be a much better measure. However, that is not publically accessible data and require opt-in permission from user in order for use to crawl their wall posts. And if not all users opt-in, we are automatically biasing against users who didn't opt-in. And we definitely do not want to do that. So, as you can see, this is a more challenging problem, and we probably can't resolve it in a comment here. 


Anyway, good discussion. Always enjoy these intellectual discussion with you. Hope to see you again next time.



Frequent Commentator
Frequent Commentator

Hi Mike,

Thanks for the response. One thing I want to clarify (for other readers especially) - 


I am not saying that we should ever use negative comments to engage people - and in fact am against it.  My point was more to the fact that while most comments are interjections consisting of a word or few words, negative emotions generated :"longer" comments. I guess this is similar to in real life when people keep talking when they are angry (more than when they agree with something).



On the number-of-friends, I agree with you - was just a thought :-). The reason I brought it up was that from an academic point of view I was wondering if the amount of interactions in some way correlated with the number of  friends the "seed" fan had. In other words, the liklihood that the fans who commented to a post were linked to each other through their network. Anyway, like I said this is just a conjecture.


Thanks again,



Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Ned,


Thank you for continuing the conversation.


Well, concerning the number-of-friends. Basically the answer to your question is there is a very weak correlation. Just look among your friends on Facebook, there are a full spectrum of friends from people who you never communicate with, to people you talk to every single day. In terms of math, we can write it like a conditional probability: p( communication | user is a friend ), which reads: the probability of communication given a user is a friend.


In reality, this conditional probability can be anywhere from close to 0 to nearly 1. So essentially there is no statistical dependence aside from the fact that 2 person has to be friends first before they can communicate or interact. Besides this dependency, which gives a very weak correlation, there are pretty much no other dependencies. And this statement is basically the content of No Game, No Gain: Realizing the ROI of Your Facebook Fans.


thanks again and hope to see you again next time.


Not applicable

Is 8. some sort of betweeness/brokerage/referral metric?

True fans I guess would like to recruit more fans or convert non fans to keep fanning the fire...


I guess what a page wants is a reputation - and a reputation is build on what a fan says to 3rd parties outside of the community. Bridging and bonding and all the Burt stuff.


Really enjoy these posts. Good job



Douglas Crets
Not applicable

I would bet that this conversation engagement is much stronger in Google+ and even though there is no business or brand presence in Google+, yet, I am sure this would be the case, because people in Google+ can customize their engagement settings, and choose with whom to engage and how. they are also truly being invited and engaged with themselves through reputation and through nurturing offline, as well as online in other platforms. 


Google+ benefits from the priming effect of having millions of users who are already experienced in social networking, and frustrated by the relationship settings available in Facebook. This should be looked at carefully.

Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Tom,


Thank you for taking the comment and for taking a guess on the last level.


It is a good guess, and fans do have the tendency to refer and convert more people into fans. However, that is not really an engagement metric anymore. But you got the idea. It is a metric from social network analysis (SNA), but not betweenness centrality per se. Well, I will definitely talk more about this in the next post of this series.


Concerning reputation, it is probably one of dozens of things that a fan pages want. And it heavily depends on the purpose of the fan page, which could be anything. And pages with different purpose may want very different things.


Thanks again for commenting, and thanks for being patient with me before I get to the last level. See you next time.


Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Douglas,


Glad to see you here. Thank you for the retweets on Twitter and thank you for commenting here.


From my limited interaction with Google+ (they really haven't been around that long), I think that is probably true, at least for now. People are definitely engaging with me more deeply on G+. However, I hate to speculate and would like to see some data before I make the final conclusion. It can also be just an novelty effect and things may slow down once all the hype slows down.


The ability to customize one's communication through circles is a plus. Because through circles, you can communicate with a more relevant set of audience. And it is definitely conceivable that the greater relevance lead to a higher probability of conversation (level 7). But i don't think it automatically guarantees deep engagement. I would love to get some data and and check it out.


Users are definitely more sophisticated now having been exposed to several popular social networks. Let's see how this play out in the long run.


Thank you again for the comment. I'm sure we'll have more opportunity to chat about this topic later. Thank you for your interest, and see you next time.


Not applicable

Hi Michael, great post and great series. I'm taking my time going through your findings and hope to engage you further on this topic. Can't wait for your final post.

Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Ricky,


Thank you for the nice affirmative comment.


Feel free to connect with me on Twitter, Linkedin, or G+. You can find links to all these on my Linkedin profile (by clicking on my photo at the beginning of the all my blogs).


I'm also in a process of collecting even more data to see if the result I obtain is indeed stable. So more data will come.


See you again next time.


Not applicable

thank for your insightful and very pedagogical posts


I couldn't agree more with your analysis on the low level of repeat visits as a good indicator of the low level of relationship-building. The Wall is the watercooler part of the Facebook page, but the meeting room where meaningful discussions occur is supposed to be the discussion tab. It is very seldom used however and I doubt there's much meaningful data to be found there (the failure of discussion forums in Facebook would in itself make for an interesting article but I digress...)


Facebook Groups on the contrary are much more conducive to community building and lasting conversations. Seems to me those Groups are actually created as 'breakout sessions' for groups of people with an existing relationship or common point of interest and the level of engagement there is a lot stronger. IMHO Google+ actually competes with Facebook Groups (the fastest growing segment of Facebook properties I believe), rather than traditional Facebook pages.

Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Stan,


Thank you for the comment. I'm very glad to heard that you find it insightful.


Your observation is very true. Although fan pages have discussion, it's not heavily used. Most of the participation on a fan page occur on the wall. Yet, the stream type of interaction on the wall is not conducive of relationship building.


It would be a very interesting discussion to see why Facebook fan pages are not engaging. I have some speculations and it has to do with the stream interaction, which make it very difficult to keep track of any conversation. In fact, the more popular a fan page is, the harder it is to keep track of any conversation, because new wall posts will push earlier conversation down the stream at a faster rate, and they will disappear off the wall faster. Also the lack of activity tracking is not helping either, because users cannot know how much they've contributed and brands cannot motivate user based on participation.


From personal uses of groups, what you said seem correct. They seem more conducive for conversation and are therefore better for building relationships. However, groups are not public, so I cannot crawl them and get the data to back it up.


However, as you mention groups are usually for discussion with people who already have relationship and a common interest among those people already with relationships. So I am not sure how well groups enable the formation of new relationships.


Google+ is definitely competing with Facebook, whether they admit it or not. Maybe I will write some thing on that too. They dynamics of these platforms are very different.


Thanks again for taking the time to comment. I hope to see you next time.


Frequent Commentator
Frequent Commentator

Hi Mike!

We have discussed your FEI posts under smm managers and I have a resukting question. 

How differ % of active and "returned" fans for big-size Pages (> 100k fans) from median level (3,45%). 

Tony Garver
Not applicable

Hi Mike,

I'm curious if you ever published #8?  


Also, I'm attending the EMBA Program at the University of Colorado and we are headed to South America next month.  Each of 6 teams is going to be working on a business hypothesis and my team is working on social media use by small and mid-size wineries.  Consequently, your research and information has been really useful to us.  Is there a way for us to connect via email so that I could tell you more about our project and get your thoughts on measuring our data appropriately?



Lithium Alumni (Retired) Lithium Alumni (Retired)
Lithium Alumni (Retired)

Hello Tony,


Thank you for the comment and I'm glad you find it relevant to your project.


Although I'd love to chat with you about your project, with my current schedule it is simply prohibitive. However, if you like to share your thoughts or project here, I'd be happy to comment when I find some spare time. I tried hard to respond to every comment although I do miss some from time to time.


Good luck with your project.

And I hope this blog will continue to be of resource to you and your team.

See you next time.