How Analytics is Revolutionizing Audience Development

I was recently invited to speak at the IAEE meeting in Boston to shed some light on how analytics can be used to increase attendance at events. In recent years event producers have found it more difficult to attract attendees, due to the rise of the Internet, the growing inconvenience of travel, and an economic recession. As event producers have struggled against these forces, they have in many cases not yet taken advantage of analytic techniques such as data mining, CRM, web analytics, social network analysis, and test and learn strategies to grow attendance levels. My session explored how to apply analytic techniques to radically improve the results of audience development campaigns. I have used these techniques on over 100 conferences, trade shows, and special events to achieve significant increases in attendance.

The topics I covered included the following:

The Lifetime Value of a Customer – A discussion of how to determine the lifetime value of a conference attendee is followed by the an examination of the much more difficult question of how to determine the lifetime value of an exhibit attendee. These attendees usually attend at no charge, and usually generate revenue only indirectly by attracting exhibitors and sponsors. In addition, I review an example of how knowledge of the lifetime value of an attendee can be crucial in decision making.

Closed Loop Marketing – A closed loop marketing system allows event managers to measure the results of all the various components of their audience development programs.  With accurate measurement of program results, they can accurately gauge the ROI of marketing programs, run controlled tests to optimize ROI, and identify key leverage points.

Email Optimization – Email is the keystone of many audience development programs. It is vital to optimize the revenue and response generated by email marketing through a comprehensive testing program. Properly done, email optimization can improve response by 50% or more, and in some cases double or even triple response. The presentation provides examples of how to identify key email test elements, implement carefully designed tests, and analyze the results.

Customer Profiling – Using the information about attendees collected during the registration process, prospects can be targeted with increased accuracy, and the results of marketing programs can be markedly improved.

Predictive Modeling – Moving beyond simple customer profiling, models can be developed that accurately predict which customers are likely to respond to promotions, and which customers are likely to defect. A case study is included on how predictive modeling helped triple conference revenue.

Segmentation Analysis – A highly effective way to identify which customers will respond to which promotions. Event managers can create custom-tailored marketing messages that address the needs of each segment to increase response, lower the cost of customer acquisition, increase retention, and increase cross-sales, up-sales, and referrals. An example of how a segmented campaign increased response by 20% is reviewed.

Web Site Optimization – Small increases in conversion rates can have a dramatic increase in registrations. An example of how minimizing abandonment rates during the registration process helped increase registrations by 30% is discussed.

Social Media Optimization –  Analytics can help event producers amplify the results of their audience development campaign through the optimal use of social media. By mining social networks to identify influential customers and prospects, adding social media profiles to the CRM system, and using predictive modeling to target high probability prospects, an event increased attendance by 30%.

As more event producers take advantage of these analytics techniques, they’ll be able to attract more and better qualified attendees to their events. Face-to-face meetings, the original channel of social media, will remain a vital method of marketing. Here are the slides from the presentation:

Twitter at Events: Find Out What Attendees Really Think

A New Way to Segment Your Twitter Followers With Analytics

The Twitter list feature has made it possible to create groups of people and follow their twitter stream independently. It’s a new twist on a classic marketing technique – segmentation. In the case of Twitter, it’s an opportunity to listen to a group of followers with a common set of interests and learn what’s on their mind.

If you have a personal Twitter account, it’s easy enough to create the lists that might be most meaningful to you. For example, you might organize them by family, friends, golfers, wine enthusiasts, etc. When it comes to business, however, it’s a little less straightforward how to go about segmenting your followers. And if your business has thousands of followers, it can get pretty tedious trying to segment them manually.

Using Text Mining and Clustering to Segment Your Followers

There are several approaches to segmenting your followers on Twitter. The powerful approach I’ll illustrate here is to use classic text mining and clustering techniques to let the data you have regarding your followers organize itself into the most appropriate segments. For instance, you could use each follower’s Twitter bio as the data to be used in creating segments. I recently did this for a client that had over 3,200 followers on Twitter. An illumnating visual aid to use in this example is a word cloud. Here is the word cloud that resulted when combining all 3,200 bios.

As you can see, there are certain words that seem to stand out and suggest different segments, such as storage, marketing, virtualization, media, technology, etc. That seems straightforward enough, but not all of the key words I just listed necessarily appear only once in a bio.

For instance, how do you go about classifying the following bios?

• Solutions for data storage, protection, availability, virtualization and collaboration.
• I tweet about Internet marketing, social networking, business development, and technology.
• Strategy, positioning, PR and Social Media for tech, B2B, consumer. Focus on clean, semi, storage and start-ups.

Each of the examples includes more than one of the most frequently used words listed above. You can see that assigning a follower to a segment on the basis of a single world is not a simple matter. If you were trying to attempt the segmentation of these followers manually, you could easily spend a lot of time trying to decide the most appropriate segment in which to place them.

That’s where the power of text mining can save the day. Using text mining and clustering algorithms, it’s possible to classify the bios into segments not just based on the appearance of a single word, but on the frequency of the appearance of all the words in the bio, and their tendency to appear together in the same bio. It’s the same principle used to find relevant documents when you use a search engine. In that way, all of the information in a bio is used to create the segments of followers who are most alike.

As an example, let’s take a look at the resulting word cloud for the storage segment.

As you can see, the bios in this segment are indeed dominated by a single word: storage, and secondarily, data. The text mining and clustering algorithms combine to create a very pure segment of followers that are focused on storage issues. The beauty of the algorithms is that they are not limited to the presence of a single word. For instance, consider the following segment:

Once you see this segment, it makes perfect sense. While the words social and media appeared separately in the word cloud of all bios, it turns out, not surprisingly, that the two words appear very frequently together, and create another very pure and distinct segment of followers. While you might be saying, well, that one was obvious, would either of the following two segments be obvious from the original word cloud?

Putting it All Together to Listen to Your Customers

The algorithms are finding patterns in the data that may not be intuitively apparent. This is a great illustration of the ability of text mining and analytics – in this case clustering – providing considerable added value in finding unexpected and fruitful patterns in data.

And here’s something that’s a real time saver: once the segments have been defined, as new followers are added, a classification algorithm can be used to place each new follower in the best segment, making the process of determining which list to place a new follower automatic.

Now let’s get back to the original reason for placing your Twitter followers on a list: to listen to what a group of people with similar charcteristics are saying. So here’s the icing on the cake: after placing your followers on a list, you can then collect their tweets over a period of time and use the same methodology: text mining and clustering, to classify their tweets into common areas of interest or concern. When you consider that you can process thousands of tweets this way, and find novel and unexpected patterns  in their comments, you have the ultimate opportunity to really listen to your customers.

Summary

• The Twitter list feature has made it possible to listen to a group of followers with a common set of interests and learn what’s on their mind.
• It’s possible to use text mining and clustering algorithms to let the data you have regarding your followers organize itself into the most appropriate segments.
• The algorithms are finding patterns in the data that are not be intuitively apparent, providing considerable added value in finding unexpected and fruitful insights into the mindset of your customers.
• As new followers are added, a classification algorithm can be used to automatically place each new follower in the best segment.
• Text mining and clustering can then classify the tweets of individuals on the different lists into common areas of interest or concern, and find novel patterns  in their comments, providing the ultimate tool for listening to your customers.

How to Kick-Start Your Viral Marketing Campaign

In my last post, I reviewed the ability of some of the most well known Twitter users to extend their reach through viral marketing. One well-known approach to viral marketing is to focus your message on a small number of highly influential people, who will then help to start a word-of-mouth chain reaction that effectively broadcasts your message to a wide audience at a low cost.  Using this strategy requires that you can identify the most highly influential individuals in your target market. New research is now available to help facilitate the indentification process. Four researchers at the Max Planck Institute for Software Services recently published a landmark paper investigating how to measure and identify influence in social networks.

Measures of Influence

The researchers focused on Twitter users. With the cooperation of Twitter, they compiled a dataset used for the research that comprised more than 1.7 billion tweets among 54 million Twitter users containing nearly 2 billion follow links.

The researchers compared three different measures of user influence on Twitter:

• Indegree Influence, or the number of followers that a user has, an indicator of that user’s popularity.
• Retweet Influence, the number of retweets in the dataset containing a users’s name, a measure of their ability to propagate a message among their followers.
• Mention Influence, or the number if tweets containing a user’s name, indicating the ability of the user to initiate and maintain conversations with others.

The Million Follower Fallacy

One of the most interesting questions tackled by the study was to what degree the three measures of influence were correlated.  The researchers focused on the 6 million most active Twitter users, and ranked each one according to each of the three measures. They then examined the correlation between the rankings, shown in the following table:

Correlation ranges on a scale of -1 to 1; a perfect positive correlation is 1 (meaning that a high rank in one measure tends to occur along with a high rank in another measure); a perfect negative correlation is -1 (meaning that a high rank in one measure tends to occur with a low rank in another measure); no correlation is indicated by a score close to 0. All three measures of influence were positively correlated. However, ties in rank among the lowest ranked in the 6 million active Twitter users artificially generated the relatively high correlation seen in the column “All” in the above table. The researchers therefore  isolated the top 10% and top 1% of users based on their number of followers, and examined the correlations between the three measures of influence. The researchers reached the following conclusion:

After this filtering step, the top users showed a strong correlation in their retweet influence and mention influence…This means that, in general, users who get mentioned often get rewteeted often, and vice versa. Indegree, however, was not related to the other measures. We conclude that the most connected users are not necessarily the most influential when it comes to engaging one’s audience in conversations and having one’s messages spread.

This phenomenon has been dubbed “the million follower fallacy” and is one of the most important conclusions of the study: if your goal is to identify the users who are most likely to repeat your message to others in a viral marketing campaign, don’t look for the users in your target market with the most followers, look for the users with the most retweets and mentions.

As you move event further up the rankings, the overlap between the top 100 ranked users according to the three different measures of influence becomes smaller:

There were 233 distinct users that made the top 100 ranking in one or more of the three measures, and 67 that appeared on more than one of the top 100 lists.

Influence Across Different Topics

Another key question the team investigated was whether a user’s influence varied by different genres of topics. To address this question, they examined top-ranked influencers for three different topics which were among the most mentioned during 2009: the Iranian presidential elections, the H1N1 flu virus, and the death of Michael Jackson. Among the set of users who tweeted about any of these topics, only 2%, a set of 13,219 users, tweeted about all three topics, demonstrating the diversity of the topic genres.

Once again, the team looked at the correlations between the rankings on the topics, this time looking specifically at retweets and mentions among the most popular users, as measured by Indegree Influence.

Given the relatively high correlations among the most popular users in their ranks for retweets and mentions across these diverse topics, the researchers concluded that opinion leaders can hold sway over a wide variety of topics. This means that the opinion leaders can help spread a message outside their area of expertise. This is consistent with recent efforts to insert advertising links into popular user’s tweets. The fact that the degree of influence of users is a long-tail (power law) distribution also leads the authors to conclude that it is more economical to target top influentials to kick start a viral marketing campaign, rather than a massive number of less influential users.

How to Become Influential

The team next looked at three groups of users who tweeted about only one of the topics, to determine what factors make ordinary users more influential. These three groups of users had from 3 to 180 times fewer followers than the highest ranked influencers. In comparison to the top ranked influencers across all topics, this set of users that tweeted about only one of the topics saw their influence rise to a much greater degree over an eight month time period in 2009.

This lead the authors to conclude that through concerted effort, and focus on a single topic, users had the greatest chance of increasing their influence over time.

Summary

The key findings of the research are as follows:

• It is easier to kick-start a viral campaign by focusing on top influencers, rather than large numbers of individuals with a small degree of influence. This follows from the fact that all three measures of influence fall into a long-tail (power) distribution.
• If you are targeting individuals with a message with a view to extending your reach through viral marketing, the number of followers an individual has is less important than the number of times they retweet, and in turn, are retweeted. The number of followers a Twitter user has may indicate popularity, but this measure has a weak correlation to retweets and mentions.
• The most influential users can hold influence over a variety of topics, as measured by retweets and mentions.
• Ordinary Twitter users can best gain influence by focusing on a single topic, and consistently including links to useful and engaging content in their tweets, as opposed to focusing on conversations with other users.

Unanswered Questions

Like all good research, the analysis presented by the authors suggests further areas for investigation:

• Are the same influence patterns evident in business-oriented communications, as opposed to the general interest topics specifically investigated in this research?
• The research focused on the users with the most rewteets and mentions. The data was not normalized to account for the number of their followers. What patterns emerge when the data is adjusted in this manner?
• Are retweets driven by the influence of the user, the content of the tweet, or the content of the link? If it is driven by all three factors, what is the relative importance of each?
• How can the research be used to predict the outcomes of social media campaigns?

The researchers are currently working with Twitter to make their entire dataset available to researchers. While the dataset is not yet available, you can check their website for updates on their data sharing plan. If and when the dataset becomes available, we can look forward to further detailed investigation of user influence in social media.

« Previous Page — Next Page »