Hashtag Analytics: Free Spreadsheets And A Booklet!

You knew it was coming!  You wanted a concise methodology for forecasting the future of your social media community.  And now you have it.

Hillstrom's Hashtag Analytics is a soup-to-nuts methodology for forecasting the future of a social media community on Twitter.

What do you get?

• A FREE dataset (in .csv format) that contains eight weeks of participant behavior in the #blogchat community, summarized at a participant/week level.

• A FREE spreadsheet that allows you to forecast the future trajectory of your social media community (you will have to write the programming code to get your data into the spreadsheet).

• 44 pages of text that outline the thought process behind forecasting the future trajectory of a social media community on Twitter.

You are unlikely to find anything of this nature from social media analytics experts, and if you do find something that allows you to forecast the future of a social media community on Twitter, you're going to pay an agency a hundred thousand dollars for the right to do the forecasting!

This booklet is available in three formats.

• Available in print from Amazon.com for $7.95.
• Available via Amazon Kindle for $2.99.
• Available via Barnes & Noble Nook for $2.99.

If, after you buy the book, you find that you want an expert to run a forecast for your social media community on Twitter, give me a holler, I'll be happy to perform the analysis for you!

Hashtag Analytics: Part 10 = Four Month Forecast

We know the probability of a #blogchat participant engaging again in the next four weeks.

We know the Digital Profile the #blogchat participant will migrate to if the participant engages.

We know how many new participants we'll have in the next four weeks, by Digital Profile.

This allows us to create a simulation, illustrating how the community will evolve over time!

Well, we have good news here ... the community was at 2,193 monthly participants, and is forecast to increase to 2,403 participants, then 2,485 participants, then 2,518 participants, then 2,532 participants over the next four months.  Remember, growth isn't coming from engagement rates ... growth is instead coming from new participants!


We can also forecast where key metrics are headed.  We know that the #blogchat community will grow by 16%, what will happen to tweets and other key metrics?

Well, this is a positive story!


What's happening is that the participants who are engaged are moving into more valuable Digital Profiles, Digital Profiles where participants are more likely to tweet and participate at high levels!




What Did We Learn?


We learned that the #blogchat community is a vibrant and successful community.


We learned that engagement rates are generally low, and that is perfectly acceptable.


We learned that a small number of participants generate most of the "oxygen" for this community ... we called them "Mega Participants".

We learned that kindness matters!!!  We learned that the simple act of thanking a first-time participant who retweets content yields an engagement rate that is up to ten times greater than observed when a first time participant is not acknowledged for a retweet of content.

We learned about Digital Profiles, descriptions of various participants that have predictive ability.

We learned that there is a common path that a participant takes as the participant goes from a first tweet to "Making A Statement" and participating at a high level.


We learned that participant growth will come from new participants, and that is perfectly acceptable (and is congruent with most of the e-commerce, retail, and catalog work I do).

We learned that the #blogchat community is growing at a 16% rate over four months.

We learned that total tweets within the #blogchat community is growing at a 40% rate, because many participants are moving into high-value Digital Profiles!!!!

We learned that, overall, Mack Collier and his #blogchat community is thriving and succeeding, a good thing!!!

We learned that we can predict the future ... not many social media analytics experts have a methodology for predicting tweet volume and participant volume ... we, however, have a methodology for doing this!!!


What's Next?

Tomorrow, we conclude our series with the introduction of a booklet that teaches us how to predict the future of a social media community on Twitter!

Hashtag Analytics: Part 9 = Seeing The Future

The beautiful thing about Digital Profiles is that we can see where participants are likely to migrate to, given the Digital Profile they reside in.

Let's take the Digital Profile a participant belonged in last month, and measure the probability of a participant engaging again in the next month.

As we've talked about all throughout this series, engagement rates are not high, and that's not a bad thing ... the #blogchat community does a sensational job of recruiting new participants!

If a participant engages again, we can look at the Digital Profiles that the participant is likely to migrate to.  This table illustrates counts by Digital Profile:

There is a logical path that a participant navigates ... the participant usually migrates to "Joining The Conversation" and then to "Shaping The Conversation".  Sometimes, the user migrates to "Making A Statement" and then to "Shaping The Conversation".

Regardless, participation and migration paths make sense.

If we know what engagement rates are, if we know what migration patterns are, and if we know what new participant counts are, we can create a simulation that allows us to forecast the future trajectory of the #blogchat community.

We'll finish our series tomorrow with a four-month forecast for the #blogchat community.

Hashtag Analytics: Part 8 = Digital Profiles

I'm going to spare you the geeky details behind Digital Profiles, there's a dozen or more posts on the topic on this blog.

Instead, we'll focus on the outcome of a Digital Profile project.  I looked at a four week period of time, using that data to classify each #blogchat user into one of eight Digital Profiles.  Each profile describes a subset of participants who possess similar behavior.

Here is the description of each of eight Digital Profiles.  Next, we'll analyze how each Digital Profile behaves in a future four week period of time.

Profile #1 = Shaping The Conversation:  These individuals are usually the most active within the #blogchat community.  What sets these users apart from other active participants is their preference for starting conversations.  When these folks participate, they average 9.4 tweets per week, and offer the most statements per week (3.4) of any other Digital Profile.  One might think of this audience as being the "thought leaders" of a #blogchat event.

Profile #2 = May Be Interested:  A subset of individuals who only participate in one of four events, but have the potential for being "conversationalists".  When they do participate, their questions are answered, they make statements, and they participate in conversations.

Profile #3 = Making A Statement:  These are active participants who generally don't participate in conversations.  Instead, these participants make statements, they have something to say, and they are comfortable sharing their point of view.

Profile #4 = Dipping A Toe:  This is an inactive Digital Profile.  Basically, these users issue one statement, and then disappear, they are “dipping a toe”!

Profile #5 = Joining The Conversation:  An active Digital Profile that likes to respond to conversations.  These individuals are likely to issue re-tweets, to amplify tweets, and to respond to tweets.

Profile #6 = One Topic Experts:  This Digital Profile is unique, in that users seem to participate in only one event during the course of the month.  When they participate, they tend to respond to statements, and they are very likely to re-tweet content from others.  When you have a topic where you are the expert, you want these folks to participate, because these folks will re-tweet your messages and offer oxygen to any conversation!

Profile #7 = Spreading The Word:  You want these folks in your #blogchat, because their sole purpose is to spread the word!  This profile represents users who are likely to re-tweet content.  These users are also likely to re-tweet content with links embedded in the tweet.

Profile #8 = The Ignored:  This is the saddest of all eight Digital Profiles.  These participants tend to issue one tweet, and most often, it is a re-tweet of content from somebody else.  Unfortunately, their re-tweet is not acknowledged, seemingly shutting down subsequent participation.
 

Hashtag Analytics: Part 7 = Why Newbies Matter

The number one complaint I get from just about anybody I discuss my work with is this:

• You only care about newbies.  Loyalty is where it is at, Kevin.

Alright.

Let's look at the #blogchat community.

From mid-September to mid-October, 1,631 participants yielded at least one tweet.


From mid-October to mid-November, only 32.6% of that audience tweeted at least once again using the #blogchat hashtag.

Do you understand the importance of that metric?


This is so eerily similar to the e-commerce, retail, and catalog data I analyze.  Those businesses retain about 38% of their twelve-month buyers, year-over-year, on average.

This metric is important, because the #blogchat community, on a monthly basis, loses 67.4% of participants.  The #blogchat community, then, must recruit 1,099 new participants over the next month, in order to keep the community at the same level of participation as last month.

1,631 participants, of which only 532 continue, requiring 1,099 new participants to fill the gap.

How did the #blogchat community do?


Try 1,924 new participants!!

This is the secret to success, folks.  It isn't about loyalty and extracting another three tweets out of mega participants.


No, it is all about finding new participants, newbies who can be developed to eventually become mega participants.

The #blogchat community has figured this out.  Good for them!

Hashtag Analytics: Part 6 = Tweet Volume

Let's look at the most recent week of data that I had access to.

I categorized all respondents based on their status during the prior four weeks.

• Mega Participants.
• Other Participants.
• New + Reactivated Participants

Now that we've segmented the users, let's look at key performance indicators in the most recent week.

Mega Participants speak for themselves, so to speak!  Twenty whopping tweets, and a total of 2,306 tweets ... nearly 40% of all volume coming from just 112 of 1,250ish participants.

Think about that one for a moment.

It is important to look at the type of tweets, and compare them to the audience who issues them.

Statements skew to infrequent participants.


Re-Tweets skew to infrequent participants.


Amplifications skew to other participants.


Conversations skew to mega participants.


Links skew to newbies.


Times re-tweeted skew to mega participants.


Times answered skew to mega participants.


It probably shouldn't come as a surprise that mega participants thrive on conversations, and are most likely to be re-tweeted or answered.


And it probably shouldn't come as a surprise that infrequent participants make statements and re-tweet other statements.


Most important is the disproportionate influence that mega participants have on the total conversation ... not good, bad, or otherwise ... just interesting.

Hashtag Analytics: Part 4 = One Week

We continue our analysis of the #blogchat community by looking at a segmentation scheme that includes recency (weeks since last participation) and weeks (number of weeks participated during past four weeks).

Last week, we looked at "Mega Participants", those with recency = 1 and weeks > 2.  Their engagement rate (probability of participating next week) was about 70%.


Now take a look at every row in the table where users only participated in one week.

What do you see?


Well, you see low engagement rates, don't you?

In other words, if the participant participates in only one week, regardless of recency, that person has a low chance of participating in the following week.


So that's not a good thing (though #blogchat data is so directionally similar to what classic direct marketers see in their RFM analytics that it is frightening).

What would any good direct marketer do when faced with an unresponsive audience?


Yup, that's right, they'd mine the data for the subset of participants who do want to participate again.

So we'll do that tomorrow!

Hashtag Analytics: Part 3 = Recency + Weeks

You thoroughly enjoyed the first two parts of our series ... among the most popular posts written this year.

So we'll continue.

In this case, I segment the user base by recency (weeks since last tweet) and weeks (number of weeks user participated in last four weeks).

"Engage" represents the percentage of the audience in each segment that participated in last week captured in my dataset.

What do you see?

Clearly, there are users that are highly "engaged".  In particular, there are two segments of users that are highly engaged.

• Recency = 1 Week, Participated in 3 of Past 4 Weeks.
• Recency = 1 Week, Participated in 4 of Past 4 Weeks.

This is going to be the criteria I use for what I call a "Mega Participant".  A Mega Participant is a participant who participated last week and participated in three of the past four #blogchat events.

In subsequent posts, we'll learn that these individuals are highly valuable, in terms of keeping the #blogchat community moving forward.


Notice, however, that "engagement" rates are not spectacularly high for other users.  I analyzed prior weeks (because there are different topics on different weeks each month) ... the same trend repeats, over and over and over again.


Also notice that prior Mega Participants only account for 9.6% of all participants in the week being analyzed.


In other words, the #blogchat ecosystem is a dynamic ecosystem, with users moving in and out every week.  And when that happens, recruiting new participants is of paramount importance!

Hashtag Analytics: Part 2 = Engagement

Today, we're going to explore a few introductory metrics, metrics that help us understand how the #blogchat community behaves.

I created a variable called "Engage".  This variable has a 1/0 value, a "1" if the user issued at least one tweet within the #blogchat community for the week ending November 11, 0 otherwise.

I then summarized activity for the prior four weeks, weeks ending November 4, October 28, October 21, and October 14.  I created two new variables, and transformed the remaining eight variables.

• New Variable = Recency ... defined as "weeks since last participation."  This variable can have a value of 1, 2, 3, or 4 ... one means the user last participated one week ago, four means that the user last participated four weeks ago.
• New Variable = Weeks ... a sum of the number of times the user participated in the last four weeks.  This variable can have a value of 1, 2, 3, or 4 ... four means that the user participated in each of the past four weeks, one means that the user only participated in one of the past four weeks.
• Statement, Re-Tweet, Amplify, Converse, Link, Tweets, RT, and ANSW are all "averaged" for the four weeks.  In other words, if a user participated in each of the four weeks, and had 2 statements, 9 statements, 4 statements, and 17 statements, I calculate an average ... 8 statements per week.  This process is repeated for each variable.

At this point, I have a dataset with eleven variables.

• Engage:  Did user engage the week of November 11 (1/0)?
• Recency:  Weeks since last participation (through November 4).
• Weeks:  Number of times user participated in past four weeks.
• Tweets:  Average number of tweets per week.
• Statements:  Average number of statements per week.
• Re-Tweets:  Average number of times re-tweeting other comments per week.
• Amplify:  Average number of times amplifying the comments of others, per week.
• Converse:  Average number of times conversing with others, per week.
• Links:  Average number of links mentioned, per week.
• RT:  Average number of times re-tweeted by others, per week.
• ANSW:  Average number of times answered by others, per week.

Finally, I transformed the last two variables, into 1/0 indicators.

• RT:  1 if user was ever re-tweeted, 0 otherwise.
• ANSW:  1 if user was ever answered, 0 otherwise.

This dataset provides us with a rich set of community dynamics, dynamics that can be analyzed in many interesting ways.

For instance, I measured "engagement", the percentage of users who participated the week ending November 11, based on "recency", the number of weeks since the user last participated.  Take a look at the findings:

Recency	Users	Engage	Totals	% Totals
1	496	38.7%	192	16.7%
2	530	17.7%	94	8.2%
3	342	11.7%	40	3.5%
4	206	9.7%	20	1.7%
99	807	100.0%	807	70.0%
			1,153

Take a look at the first row.  Users who last participated the week ending November 4 had a 38.7% chance of engaging (i.e. issuing at least one tweet with the #blogchat hashtag) during the week ending November 11.  496 users had a 38.7% engagement rate, yielding 192 users who participated during the week ending November 11.


This is a pretty low rate.  And look what happens as participation becomes more "distant" ... 17.7% of those who last participated two weeks ago engaged ... 11.7% of those who last participated three weeks ago engaged ... and 9.7% of those who last participated four weeks ago engaged.


In other words, it is really important to keep the user "engaged".  If the user takes a week off, or two weeks off, the user becomes less and less likely to engage in the future.


Look at the column labeled totals.  This is the number of users by segment.  In total, there were 1,153 users who participated in #blogchat during the week ending November 11.  Most important, 807 of the 1,153 users, 70% in total, had not participated in the past four weeks.


This is sooooooooo important.  The success of the #blogchat event for the week ending November 11 was largely due to the fact that 70% of those who tweeted something had not tweeted anything in the prior four weeks!


In Direct Marketing, we know that New Customer Acquisition means EVERYTHING.


In Social Media Communities, we have the first piece of evidence that New User Acquisition means A LOT!


Tomorrow, we'll continue to analyze engagement rates across various segments.

Hashtag Analytics: Part 1 = Data Acquisition

We're going to try something new here, something that is trendy in analytics these days.

We're going to analyze hashtag behavior on Twitter.

Our analysis will focus on the #blogchat community, a Sunday evening social media discussion hosted by Mack Collier on Twitter.  Every Sunday night, this community discusses various topics of interest to their community.  Folks communicate via the #blogchat hashtag, so that everybody can follow what folks are saying about each other.

I collected data from about eighteen weeks of #blogchat events.  After analyzing the behavior of the community, I reduced the dataset to five weeks of behavior.  I used four weeks as the segmentation period, with one week as a prediction period.  Later in the analysis period, we'll review eight weeks of data, using four weeks to predict the next four weeks.

Allow me to explain the variables I am tracking in my dataset.

The first variable is called "statement".  Here's what a statement looks like:

• MineThatData:  I am really looking forward to #blogchat tonight!

In other words, the individual is communicating a statement to the entire audience.


The second variable is called "re-tweet".  This is a huge form of social currency.  The person issuing the re-tweet is giving another individual credit for saying something clever, or is trying to gain attention in some way.

• MineThatData:  RT @mackcollier Bloggers really do a nice job of sharing interesting topics #blogchat.

The third variable is called "amplify".  This happens when a user adds to the statement offered by another individual.

• MineThatData:  And they have unique opinions. RT @mackcollier Bloggers really do a nice job of sharing interesting topics #blogchat.

The fourth variable is called "converse".  Here, one user is having a conversation with another user.

• MineThatData: @mackcollier Don't you think that bloggers could do a better job of being objective? #blogchat

The fifth variable is called "link".  Often, when a person makes a statement, the person links to another article to back up the statement.

• MineThatData: We covered this topic on my blog last month:  http://bit.ly/dk928d #blogchat

We sum each tweet a user issues, and create a sixth variable, called "tweets".

The first set of variables describe the actions a user might partake in.

The next two variables are very important.  These two variables account for feedback from other users in the community.


The seventh variable is called "RT".  Each time a user is "re-tweeted", I tally one for the user in the "RT" column.  In my earlier example, @minethatdata gets a value of "1" in the "re-tweet" variable.  @mackcollier gets a value of "1" in the "RT" variable, because his statement is re-tweeted.

• MineThatData:  RT @mackcollier Bloggers really do a nice job of sharing interesting topics #blogchat.

The eighth variable is called "ANSW".  This is an important variable, because it means that the user is engaged in a conversation, and that the other person in the conversation elected to answer the user.  In our earlier example, @minethatdata gets a value of "1" in the "converse" variable, while @mackcollier gets a value of "1" in the "ANSW" variable.

• MineThatData: @mackcollier Don't you think that bloggers could do a better job of being objective? #blogchat

My dataset is generated on a weekly basis.  Each week, I categorize all activity in the #blogchat community, for each user participating in the #blogchat community.

Let's assume that a user, called @user, issued the following tweets.

• @user:  @person But don't you think that brands should be "all-in" in Social Media? #blogchat 
• @user:  If big brands don't join the conversation, they're finished. #blogchat
• @user:  What do you think are the three most important things a large brand should do first? #blogchat.
• @user:  @person This link is very helpful. http://bit.ly/dkeo229kf #blogchat. 

If this is all of the activity I can find for @user, then @user has the following profile for this week:

• Statement = 2.
• Re-Tweet = 0.
• Amplify = 0.
• Converse = 2.
• Link = 2.
• Tweets = 4.
• RT = 0.
• ANSW = 0.

The data set has one row per @user / week combination.

And when you have data formatted in this manner, you can make magic happen!  Tomorrow, we begin to explore the magic behind the #blogchat community. 

An Analysis Of Social Media Users And Behavior

A Social Media Consultant named Mack Collier holds a weekly blog chat on Twitter. This Sunday evening forum gives users an opportunity to discuss various Social Media issues. Users review content via the #blogchat hashtag.

I took 71 page transcript from an April 2009 blog chat, coded each of 988 tweets among 131 users, and categorized each user into one of four Social Media user segments, based on behavior exhibited during the chat.

Why not give the analysis a read, and let me know your thoughts about the analysis and findings. There are many different ways to analyze Social Media activity, this is one attempt with potentially actionable findings for companies looking to participate.

Click Here For The 71 Page Blog Chat Transcript.

Click Here For My Analysis Of 988 Tweets Among 131 Users.