Digging Insight From Your Twitter Community Like A Miner

A huge thanks to everyone who shared their favorite Twitter tools on my last post about my mistake in unfollowing a ton of people on Twitter. You all provided a lot of great recommendations and it's appreciated.

I checked out every tool I could access and found all of them lacking the deep insight I'm looking for.

So let me clarify in this post just what kind of information I'm looking for and how I'm starting to get it. I'll keep the geekery to a minimum for the technophobes reading this.

The Main Problem To Tackle

Here's the “problem” I'm trying to solve:

How do you continue to get value from a continuously growing Twitter community that produces tens of thousands of tweets, without having to read a lot of the tweets?

When I say “value” I mean information that includes (but is not limited to):

  1. The top peeps in my community
  2. Influence of my community members on their communities
  3. Who is connected to who
  4. Who is talking with who the most
  5. What people are talking about now
  6. What are the trends in the conversations
  7. The correlation between all the numbers

A lot of solutions – lists,, (which looks very cool – thanks Kristi!) – though semi-automated (lists can be automatically created with apps) still require a lot of reading. And while some applications like help you find other people to connect with, they don't take you far enough down the rabbit hole.

How Deep Is The Rabbit Hole?

There's a very good reason I haven't been able to find a tool that exists to provide the numbers I want – they aren't that easy to get, and it requires a lot of number crunching. However based on your response to my last post it sounds like something that is definitely desired. So we'll see what I can come up with.

But one thing at a time. First we need to…

Start With Better Questions

It's great to know things like the number of friends and followers my followers have, how many times they're listed, and how chatty they are. That scratches the surface.

Here's part of the questions I'm asking (and finding answers to) about my Twitter community:

  1. Basics: # of friends, # of followers, # of friends not following back, # of followers I'm not following
  2. Comparison with another Twitter user: # of mutual friends
  3. Who is talking with who the most
  4. Who is my community is friends with who
  5. Cliques: # of cliques, average clique size, maximum clique size (largest set of common friendships)
  6. What are people in my ย community talking about now
  7. What are each of the people in my community talking about the most
  8. On a per-member basis, who does the peep retweet most often?

And there's more.

But before we go too far down the rabbit hole here's a sample of the data I've gathered.

Let's Go To The Data!

I'll save you the geekery of how I got these numbers but needless to say much reading, programming, trial and error was involved. Having said that let's take a look!

The Basics

  • Following 1,262
  • 582 of the 1,262 I'm following are not following me back
  • I'm not following 8,214 of the 8,894 that are following me (which I am in the process of remedying now)

Top 10 Followers From A Sample

User Follower Count
TheEllenShow 9,577,318
zappos 1,980,354
yokoono 1,848,044
threadless 1,768,705
TOMS 1,155,706
GuyKawasaki 587,723
ElNacionalWeb 545,830
AnthonyGemma 521,675
wakooz 509,752
StartupPro 450,966

The average number of followers that my followers have is: 8,984. Just a few ๐Ÿ™‚

As for the top 10 people following me, I'll let you guess what the probability of any of them actually retweeting any of my tweets is. I'm not holding my breath waiting for it to happen ๐Ÿ˜›

Hello Pam Moore!

Just for fun I ran the same analysis on Pam's Twitter community, which is quite a bit larger than mine. I'm not going to share that data (Pam you can contact me for it) however I can say that of the 10,000 users I analyzed from her community, I found out two things:

  1. We have 361 friends in common
  2. We have 2,217 followers in common

Coincidence? Perhaps not as we both talk about marketing.

How Connected Is My Community?

A partial analysis tells me that:

  1. There are 1259ย cliques in my Twitter community
  2. The average clique size is 2
  3. The max clique size is 3
  4. There are 359 max cliques

So what does that mean? It means that I have a lot of followers that aren't connected with each other.

How Influential Am I?

Ah the million dollar question that companies are attempting to answer. Well I'm not going to say I've found some “magic number” to say if you or I are influential on Twitter, but let's look at one number – how often we are retweeted.

In looking at the latest 3,173 of my tweets, 206 were retweeted at least once. That's (perhaps a paltry) 6.49%. How does Pam match up with her 73,373 followers? Much better than I!

Pam had 1,154 of 2,965 of her latest tweets retweeted at least once. That's a tasty 38.9% retweet ratio.

I'll leave it to you to (perhaps guess?) analyze whether or not her large follower count directly correlates to her retweeted percentage, or if it's because everything she tweets is JUST THAT GOOD!

Ok, So Where'm I Going?

Step 1: ask better questions.

Step 2: create codes to provide answers to said questions.

Step 3: write a blog post telling you fine people about what I've found so far

Step 4: begin building an online tool that will tell you even more information

Step 5: continue to keep my family fed and clothed (and daughter in school) by getting paid for said tool, once I show it's value

Any Questions?

What do you think of the information I've gleaned from my Twitter community? What do you want to know that todays tools aren't telling you?

Inquiring minds want to know!

Let's chat about it in the comments below. See you there!


  1. Wow, you have put a lot of effort into collecting this data and organizing them into useful information ๐Ÿ˜€

    Thanks for sharing the data with me ๐Ÿ™‚ I guess I need to ask these questions myself.

    Today’s twitter tools don’t offer much information to us (an exception is Bufferapp which is a great way to find your click through patters and the best times to tweet).

    Thanks for the awesome article, Robert,

    Jeevan Jacob John

  2. While twitter is not in my repertoire, I am following your analysis with great interest. And like Adam, I can see the huge potential in expanding this to FB and other social media outlets.
    Who says you donโ€™t need math after school?

  3. Great stats Robert!

    I did read about the unfollowing done on Twitter and thought as to why, but I think just like so many others have done it’s always great to start with a clean chit and take the measures you didn’t take before.

    There are many tools that can be used, as you so rightly mentioned, or unless you open up each profile and check it out before really going ahead and following them back.

    Nonetheless, I still doubt that people really read the posts they tweet about- most of them don’t. So, one can’t really say if they are truly tweeting your posts because they’re liking it or because they feel kind of obligated to do so, or returning a favor.

    Thanks for sharing

    BTW- Nice look to the blog, clean and white ๐Ÿ™‚

    • Great to hear from you Harleena! As for your analysis I tend to agree. So much to read, only so many hours in a day. I know a lot of people that mostly skim content.

      Thx for the kudos on the new design too! Much appreciated.

  4. That is some really awesome analytics. And if you create a tool to analyze all that stuff easily, especially if you can eventually make it so it analyzes facebook fans of a page, and Google+ people who have circled you or your business page. Man will that tool be worth a lot.

Speak Your Mind