Sarj

Archive for the ‘Technology’ Category

Putting Your Money on Word of Mouth: Research on the social value of seeding programs

In Marketing, Technology on June 3, 2010 at 12:27 pm

This post was originally featured on http://blog.sociomantic.com, published on June 3rd, 2010. Since the website will be relaunched and the post removed, I have relocated the tutorial to my personal page so that the SNA and online marketing communities can continue to benefit from it.

When I attended the Marketing 2.0 Conference in Paris this March, I had the pleasure of seeing a presentation by Barak Libai, a marketing researcher from Tel Aviv University. Throughout the last decade, Libai his various researcher partners have been exploring the intersection of complex networks and marketing. His research includes topics such as innovation diffusion, word of mouth marketing, customer valuation & customer relationship management, the impact of “influence,” and the application of complex system methods in marketing research.

Last November, Libai and co-authors Eitan Muller and Renana Peres published “The Social Value of Word-of-Mouth Programs: Acceleration versus Acquisition.” Because of my experience with a company that used social network analysis to provide businesses with intelligence about their customer influencers and other aspects of their customer networks, I found the implications of this research fascinating, so I wanted to give a brief overview of the research findings and what they imply for the application of “customer network analysis”.

WOM FTW? (“Word of Mouth For the Win?”)

Word of mouth initiatives have long been a staple of the marketing diet, and in recent years so-called “influencer” targeting has become the go-to extension for the traditional WOM marketing models. Whether they’re labels as influentialsopinion leaders, or hubs, finding these influencers in a vast and complex network of customers is no small task; in light of the costs, managers often struggle to “achieve financial justification,” for this sort of marketing, though many are convinced of its benefits.

Prior to this paper, most of the research measuring the “social value” of influencers used soft metrics like message diffusion, conversation, or brand awareness; the monetary effects of WOM contagion have not been thoroughly defined in research. Libai and his partners are some of the first to attempt to answer the big question of ROI, to define the social value created by WOM in terms of customer equity. They define the impact of these WOM seeding programs based on the interaction of two factors: customer acquisition and customer acceleration (defined below).

Defining the Social Value of a Customer

In order to get the most accurate assessment of a customer’s potential WOM impact, the researchers define an individual’s “social value” based on the absence principle presented in the famous film “It’s a Wonderful Life”:

For those who might not have seen it, this classic story shows the effect on a town if one particular man had never been born (e.g., his wife is a never-married old maid). Libai et al. suggest that measuring a customer’s social value requires a similar method: measure the change in value over the entire customer network if this particular node were removed. To determine that impact, the researchers used “agent based models” (think Sims 3 for scientists) based on real-life social network structures.

So basically, based on this method, if Coca Cola’s marketing guy wants to determine how much social value I have within Coca Cola’s customer network, he’d make an “agent based model” of their customer network, then run a simulated WOM seeding program on the model, first WITH me in the network and then again WITHOUT me in the network. Then he’d determine the difference in how much money they’d make (for a given period of time) in each situation – that difference is my social value. It answers the question: how much are my connections worth to Coca Cola over the course of my time as a soda drinker?

If it isn’t already clear, this method of defining a customer’s social value has huge implications for companies trying to determine the value of their their WOM programs as well as social media marketing programs: finally, a method to determine which nodes (ahem, FOLKS) in their networks can have the biggest impact on their profitability, and just how big that impact can be.

Acquiring new customers, Accelerating adoption from would-be customers

There are two major ways that WOM campaigns can affect a product’s market share: acquisition and acceleration. Acquisition is how many new customers decide to adopt a product who would have otherwise adopted another brand or not adopted such a product at all without the WOM program. In the business world, time = money, so the sooner you a customer adopts your brand, the more time they have to spend money on your products – that’s the “acceleration” aspect. Acceleration is how much sooner people adopt a product due to the WOM program. Why does this matter? Adopting a product sooner gives a customer a higher lifetime value, so the sooner they adopt, the better! (For example, if I start buying Apple products at age 18, I am probably worth more to the company in the long run than if I start buying Apple for the first time at age 45).
The potential social value of a WOM program can then be determined in a similar manner — by comparing the “lifetime value” of the entire customer network with the program to the “lifetime value” of the network without the program (using agent-based models).

Sow your (marketing) seeds in fertile ground – but which ground is most fertile?

The WOM program type used in this study is what is commonly known as a seeding program, in which a company offers some sort of promotion (free product, discounts, service trials, etc.) to a “seed” group of people with the intention that these people will market for them via WOM. (Why? Previous studies have shown that customer loyalty is higher for customers acquired via “recommendation from a friend” than for those acquired via marketing).

One of the key questions surrounding seeding programs over the last decade has been whether or not it’s more effective for companies to seed via influencers or random people their customers networks. (This Fast Company article from 2008 gives a nice overview of the great influencer debate.) To investigate this question as well as the impact of competition on a program’s effectiveness, the researchers measured the (simulated) change in a brand’s customer equity in five different scenarios:

  • no seeding program
  • “random” seeding program without simultaneous competitor seeding program
  • “random” seeding program with simultaneous competitor seeding program
  • “influencer” seeding program without simultaneous competitor seeding program
  • “influencer” seeding program with simultaneous competitor seeding program

The agent-based models were created based on 12 real-life social graphs from a variety of sources. The size of the seeding program (i.e., what percentage of the total network was chosen as “seeds” for a product) was determined based on industry standards, varying the size from 0.5% to 5% of the potential market. The “influencers” were defined as the top 10% of network with the most connections (highest “degree”), and the set percentage of “seeders” was chosen at random from this group. Conversely, for the “random seeding” program, the “seeders” were chosen at random from the entire network. Each of the five combinations of parameters listed above was run 20 times to combat the probability of an unlikely single run. These results were then averaged for each network to draw the final conclusions.

Research Results

At the end of it all, the research showed that the influencer seeding programs yielded higher customer equity gains in both the single brand and competitive scenarios, but  a competitive seeding program greatly reduced the overall effectiveness of a seeding campaign, whether the seeds were random or influential. Still, the results of the random programs were pretty good, so the decision about which program type to adopt depends on how much the company is willing to invest in discovering their influencers. Because it often requires complex, high-demand computing, customer network analysis can be time consuming and expensive, so some companies opt to let external vendors do their CNA.

One interesting point is that in a competitive scenario (e.g., when Pepsi and Coca Cola are both running simultaneous seeding programs), when the acquisition effect of the companies’ seeding programs is in essence “canceled” by the other program, BOTH companies can still benefit from the acceleration of their product in the market – getting customers sooner gives each of those customers a higher lifetime value, and therefore increases the lifetime value of the network overall.

The research offered in six other major results which we will summarize:

  1. “The competitive program effect”: The social value of a seeding program is “considerably higher” when the firm faces a competitor than when they are the “sole player in the market.” This can be explained by the fact that in a market with no competition, everyone in the potential market will eventually adopt the available product/brand, so the gain is only from acceleration instead of a combination of acceleration and acquisition.
  2. “The brand strength effect”: “Weaker” brands have more social value to gain from a WOM seeding program than “stronger” brands.
  3. There is a threshold “seed size” for both influencer and random programs before the program actually decreases the social value of a company instead of increasing it. So more seeds does not necessarily equal a more effective program!
  4. Influencer seeding programs reach that “threshold” much more quickly than random programs, so they reach their “peak” social value potential at a much smaller seed size.
  5. While about 75% of a seeding program’s WOM potential can be achieved using a random seeder program, “targeting influentials can increase social value considerably” (p. 31).
  6. The role of acceleration in increasing social value is more prominent in influencer seeding programs than in random seeding programs.

The downside

One things that the research (self-admittedly) does not take into account is the cost of implementing these seeding programs. Each company would have to determine on a case-by-case basis what their social value gain would be less the costs of implementing that program, and “such calculations may demand network-specific analysis” for each company network and product. Many companies have only just begun to realize the benefits of representing their customer networks for the purposes of customer network analysis. It’s a process that can be both complex and costly. But there are SaaS (software as a service) companies out there who can take your existing customer names and put them into a network framework to help you identify the influencers and analyze your network to determine the ideal WOM campaigns for your purpose, product, and people.

For a nice review, check out this presentation Libai gave at WOM UK conference in April:

Using Netvizz & Gephi to Analyze a Facebook Network

In Technology on May 6, 2010 at 3:00 am

This post was originally featured on http://blog.sociomantic.com, published on May 6th, 2010. Since the website will be relaunched and the post removed, I have relocated the tutorial to my personal page so that the Gephi community can continue to benefit from it.

If a picture is worth a thousand words, then a graph must be worth a thousand spreadsheet rows, right?

A Facebook network rendered in Gephi

Okay, maybe not, but for practitioners and researchers alike, data visualization can reveal insights that aren’t always obvious from looking at the raw data, no matter how well organized it may be. When we’re talking about social network, data visualization takes the form of a “social graph,” and it can be a powerful tool to discover deeper meanings and applications behind the relationships and communities within a network.

Here you can see some social graphs of the French political blogosphere created by researcher Tim Highfield using an open-source network visualization software called Gephi. After exploring Tim’s amazing Flickr full of graphs and reading @kristtina’s recent introduction to Gephi, I wanted to try out some of these social graph visualizations myself.

The Alternatives

If you’re interested in something with less of a learning curve, there are lots of easy-to-use, mostly flash-based visualization apps for Facebook and Twitter. These are the ones I’m aware of:

Facebook:

Twitter:

The great thing about these apps is that they do most of the work for you. And a lot of them look pretty cool. The problem is that they don’t give you much room to explore. If you’re hoping to analyze your Facebook network with a little more depth — to discover community clusters and explore network science parameters like degree, betweenness, closeness, etc. –  I’d recommend using Netvizz and Gephi. A colleague told me about Netvizz some time ago — it’s a Facebook app that allows you to make a .gdf file out of your Facebook friends or the groups you’re in (.gdf is the file type reader by programs like GUESS and Gephi).

Two quick notes about Netvizz:

1)      Right now it can only analyze the friends of your your Facebook “profile” (for individuals) and the members of groups you’re in. Hopefully soon it will be able to provide .gdfs for “Page” fans as well so brands and companies can do Facebook social graph analysis using Gephi, too.

2)      The .gdf files for the Facebook groups are limited to 500 randomly selected nodes, no matter the size of the group. (Theoretically you could generate the random list .gdf enough times to discover all the nodes in the group and combine them into one all-encompassing file if you were looking to do some serious network crunching.)

Here are some of the networks I analyzed using the .gdf from Netvizz in Gephi:

Analyzing a friend's Facebook network -- You can see distinct community clusters

Facebook Group "Graphs & Social Networks" : A highly connected network!

Here’s a quick key to understanding these graphs:

  • Circle = Node = Facebook friend or group member
  • Line = Edge = Facebook connection (friendship)
  • Node size = Betweeness centrality (measure of how much a node connects otherwise disconnected communities)
  • Node color = randomly chosen colors used to represent the communities/clusters, determined here based on their modularity class via the Louvain method

Taking a Closer Look at Using Gephi

I think the most interesting network I analyzed was the one I posted an image of up at the top of this post. You can easily see the different communities to which my friend is connected identified in the graph, and it’s interesting to see which nodes have the most impact over multiple groups.

Since I took screen shots along the way, I made this slideshow to explain the steps I took to reach the final visualization.

Since I’m still learning I initially followed the Gephi Quick Start guide.  They have a file you can use to try out this process if you don’t want to use your Netvizz .gdf.

From an industry standpoint, studying social graphs like these over time can enable companies and brands to understand things such as:

  • Which individuals are connecting disparate communities within their customer base. (If this Facebook network was my customer base, I’d definitely want to make sure I am reaching out to our managing director Thomas Nicolai, who has many connections to multiple communities within the greater network.)
  • Over time and using methodologies to determine parameters like reputation and bandwidth, you can discover which individuals are gaining influence within particular clusters (e.g., someone who starts small might become more influential over time)

I hope you found this tutorial helpful! Please feel free to share the link to help others learn 🙂

Nightwish and French Politics

In Technology on April 20, 2010 at 12:00 pm

This post was originally featured on http://blog.sociomantic.com, published on April 20th, 2010. Since the website will be relaunched and the post removed, I have relocated the tutorial to my personal page so that the SNA community can continue to benefit from it.

If you’ve never heard of the operatic power-metal band Nightwish, I’d suggest you start below. Even if metal’s not your thing (can’t really say it’s mine), it’s worth watching just to know it’s out there…

I was first introduced to Nightwish in college, but I hadn’t thought about them in several years. I was reminded of this outrageous group when, a few afternoons ago, our managing director Tom sent me a link on Skype. Not such an unusual occurrence, no doubt, but this link was extra-awesome. Within a couple of minutes, I realized I wasn’t the only one who’d received the URL, because suddenly the air was full of exclamations.

For the next quarter hour, everyone in the office was poking around this interactive graph (see static representation below), which maps the relationships of similarity between the artists in the Last.fm Audioscrobbler API. (Last.fm is a website that allows you to log, or “scrobble,” all the music files that you play on your computer or mp3 player so you can see statistics about the listening habits of you and your friends.) The graph, created by Dr. Tamás Nepusz (a postdoctoral research fellow for the Royal Holloway University of London), demonstrates the network of relationships between musical artists (or groups) based on the artist “similarity” algorithms from Last.fm. The different colors represent the various musical genres, the size of the circles is proportional to the number of listeners for the artist’s “top track” on Last.fm, and the little lines in between show the strength of the similarity between two artists. And how do our pals Nightwish fit into all this? Well, that’s the band that Nepusz started his data collection from, and once reminded of their unique musical offering, I just couldn’t pass up the opportunity to share.

Nepusz's static representation of the Last.fm artists similarity network

I’ll let you read the details of his project over here, but here I want to use this awesome network map to help explain one of the research projects that my colleagues have been involved in.  (Of course, I understand if you need a minute to plug your favorite bands or your Last.fm username into Nepusz’s graph.)

The Name Game

First thing first, let’s use this Last.fm graph to help explain some of the network science terms that that we use when talking about these sorts of visualizations (in our case, a social graph).

  • Node or Vertex
    A node is a connection point in graph. In the Last.fm visualization, each artist/group is a node represented by a colored circle (with the artist name inside). In the social graph, each node is a different person in the network.
  • Edges
    The edges, represented as lines between the circles, show the actual connections between nodes – in this case between similar artists. Here, the darkness of the line indicates thebetweenness of the connection, which basically describes the potential of the edge in terms of information flow between different groups. (An artist with high “betweenness” might be one with high similarity to artists in both “rock” and “hip hop” genres – someone with a lot of what the music industry calls crossover potential.”) In a social graph visualization based on the web (like sociomantic’s), the edges are the links between people — so if my blog is listed on your blogroll, that’s an edge connecting us.

Like a social graph, this Last.fm graph visualizes the relationships between nodes that carry various weights. Here, the “weight” (size of the circle) is based on the number of listeners for an artist’s top track on Last.fm, but in a social graph, the weight might be based on measures like centrality (number of connections), betweenness (how much the node connects separate clusters), or influence (determined by a combination of many measures).

Connecting the Dots

I wanted to share this visualization because playing around with it can help someone who might be new to network graphs to get a grip on what I’m talking about. And while I hope you had as much fun with it as we did, what I really want to share is some info about a research project my colleagues have been working on with the guys over at the ARC Centre of Excellence for Creative Industries and Innovation at Australia’s Queensland University of Technology.

So what exactly are they studying?

Axel Bruns, one of the QUT researchers, briefly explains:

Building on our joint research into the Australian and French political blogospheres, we’ve embarked on a large-scale, three year research project to investigate the processes of online public communication in Australia and France as they unfold across major social media spaces including Twitter, YouTube, Flickr, and the wider blogosphere.

This particular project is only one part of their ongoing study of the flow of information between blogs and media websites that began in 2007.  (Those interested can find the full research collection here).

A static representation of the French political blogosphere in 2009, by Tim Highfield

Naturally, data collection is at the root of this research. Back in 2007, the research team compiled a list of known political blogs (at this time, only for Australia – they wouldn’t begin keeping tabs on the French blogs until 2009.) Using web-crawling technology,  the list of blogs was expanded based on the outbound links from these known blogs, then expanded again based on the new list — and so on, until there ceased to be new links or the links surfacing were irrelevant to the political blogosphere. The links used could be from the official blogroll, from links within individual blog posts, or from links listed in the comments.

After the initial blog list was compiled, they continuously crawled each and every new post in order to scrape for new linkages, so that by gathering links over the course of time they could begin to understand the blog network – which blogs were linking to which others, which blogs had the highest number of inbound links, which ”clusters” of blogs could be identified by their interlinkage, etc. The researchers will continue to scrape these blogs for further links until the end of the three year research period. They also gathered topical data from the blogs so they could get an understanding of how different events drove the flow of information within the network.

There have been many stages of analysis over the course of this research, but below you can see just a few examples of how these researchers are making sense of all the data gathered. In the following slideshow, researcher Tim Highfield offers us a nice overview his initial findings.

(Scroll down to see Axel Brun’s October 2009 presentation about the findings in the Australian blogosphere.)

Tim outlined some of his recent observations over on his blog. The images below are just a few of the network representations that Highfield has created by plugging the data gathered into visualization program called Gephi.

Although the rainbow-colored visualization I showed before the slideshow might be the nicest to look at, it also shows how imprecise mapping can leave us with little more than what Nepusz labeled as an “ugly hairball.” For the two visualizations above, Tim groomed the data a bit by doing things like eliminating nodes with less than two incoming links and coloring blogs with known political affiliations.

Like Neupsz’s last.fm visualization, Tim’s pictures show us how, with careful application and analysis, network graphs can be powerful learning tools for understanding the way information and influence moves within a network. From a business perspective, hopefully now it’s a little more clear why it might be important to use a customer network graph to better understand the “big circles” in your web of clients and prospects.

What do you think the researchers will find in this data? Do you think it’s most likely that the information flows from the political blogs to the official news sites, or vice versa?

Marketing 2.0 Conference 2010: A Review

In Marketing, Technology on March 24, 2010 at 12:40 pm

This post was originally featured on http://blog.sociomantic.com, published on March 23, 2010. Since the website will be relaunched and the post removed, I have relocated the tutorial to my personal page so that the SNA and online marketing communities can continue to benefit from it.

I’ve just returned from Paris, where we had the pleasure of attending the M2C at the ESCP Europe campus. There were lot of really wonderful presentations from many great brands and companies, so we wanted to share some of our biggest takeaways from the conference.

Manish Metha, Dell
Intimacy and Scalabilty: Using Social Media To Manage Your Brand

  • As a brand’s social media model grows, they often lose their ability to interact intimately with their customers. Metha suggests reconciling this negative relationship through “conversation clusters” — break down the conversations that are happening into distinct spheres of interest to better manage the interactions. Segmenting the conversations into different clusters will help provide more and better conversations.
  • For B2B, social media should be about connecting “subject matter experts,” engaging the blogosphere, participating in the conversation

Emmanuel Vivier, Vanksen
6 Rules to Fail With Social Media

  • “Marketing is like karma,” you have to give something if you want to get something.  You have to offer customers value to make your social media strategy work.
  • He offered the following update to Forrester’s 2007 “POST” strategy for social media marketing:
    • Profile: understand the activities of your audience
    • Objectives: Decide what you want to accomplish
    • Strategies: What is the creative idea? message?
    • Tools: What tools, formats, platforms?
    • Evaluate: What metrics and KPI?
      • My two cents: I found “evaluation” to be a big thing missing from a lot of the social media talk at the conference. It’s clear that companies should be using social media to engage customers, but it wasn’t always clear what they could show the boss at the end of the day. This is where social network analysis and related analysis of customer lifetime value comes in.

Steve Knox, Tremor (Proctor & Gamble)
Perspectives on Innovative Word-of-Mouth Advertising — P&G Learnings and Future Outlook for Brand Markers

  • Amplification of your brand without advocacy of your brand is a waste. (Don’t be caught with a “successful” viral marketing scheme that no one associates with your company.)
  • A brief lesson in cognitive theory: To save brainpower, our minds operate in terms of “schema” — patterns that determine our expectations of a situation. Our attention can be grabbed when a schema is interrupted (things happen in an unexpected way) OR when two schema are combined in a “conceptual blend.” Brands must work to make their marking “mildly, not wildly, incongruous” with our existing schema if they want to get our attention.

Pauline Ores, IBM
Influence Starts @ Home

  • As a B2B company, in order to influence, you must be influenced. Be knowledgeable, respond quickly, and don’t forget to react.
  • The potential insights you gain by “listening” to your customers on the social web can only make an impact on your business if you turn around and feed that information inside, then grow upon it internally. At IBM, they use a highly successful internal social networking to drive innovation and improvement. No matter how big your company is, it’s essential for the internal communication to react rapidly to the external communication in order to see results.
  • (Pauline, if you’re reading this — thanks for your insight and conversation! It was a pleasure to meet you and to hear about the interesting and innovative initiatives you’re helping to lead at IBM.)

Carlos Diaz, Blue Kiwi
Successfully Tie In Traditional Social Media Techniques Within Your Existing Multi-Channel Marketing Strategies

  • You need to have an engagement strategy, not just a social media marketing strategy. How will you get your customers and potentials to interact with your brand in a meaningful way?
  • Conversation from influencers is the most important for ROI. Find a way to filter out the “best” (most valuable) conversations and focus on these conversations.

Mike Butcher, TechCrunch Europe
Will the Next Wave Be Intention?

  • In Web 2.0, marketing has gone in waves: social marketing > viral marketing > location-based. Is “intention based” the next wave?
    • My two cents: One of the main downfalls of existing social media monitoring systems is that they often provide conversation (complaints, feedback, wishes) to which companies can react, but they are seldom empowering brands to proactively pursue their customers and potentials.

Barak Libai, Tel Aviv University
How Can We Assess the Real Value of Word-of-Mouth for Brands and Marketing

  • “Social Value” is the monetary value a customer adds due to their social interactions. There are two dimensions in which to calculate this model. The first by the value of the product that the person buys because of a social interaction. The second, and far more complex, dimension has to do with the value of the people who are influenced to buy the product by this person.
  • Determining customer lifetime value: In the social spectrum, we move from a measurement model of customer acquisition (now we have another customer because of marketing effort X) to a model of customer acceleration (we have another customer sooner because of marketing effort X)
  • There is no simple way to calculate these values — there are issues such as of the number of friends influencing a single person (influence from multiple nodes), the amount of chatter surrounding any influential conversation, etc.
  • Net Present Value (NPV) of cash streams: academics have learned that clusters of people (friends, people united by similar interests, etc.) tend to have similar lifetime values. So in order to determine how valuable any one customer may be, they could use what Libai called the “It’s a Wonderful Life” approach, in which they build a model to determine how customer lifetime value might be affected if a single person were removed from the bigger picture. This is the approach they’re taking now:
    • Collect data on real social networks
    • Run simulations on these model networks (think Sims 3)
    • Conduct experiments in these model networks (removing nodes — notice the difference between removing an influencer versus a non-influencer)
  • What have they learned? Influencers are definitely important, but random seeding is good, too

Sumaya Kazi, YoProCo (formerly of Sun Microsystems)

  • Big companies should think like startups when it comes to social media marketing. Startups benefit from having a united voice, and this is what big companies need. Startups have to use social media because they don’t have big budgets for marketing; big companies should take the frugal approach to social media marketing instead of waiting for big budgets to pass through the upper tiers.
  • Great ideas from Sumaya:
    • Designed a Sun Microsystems Facebook group that promises “bytes” of info on a regular basis
    • “Sun Social U” training courses for Sun Microsystems to teach employees about Sun’s social media guidelines
    • Shared a conference they attended by distributing 100 (or so) Flip video cameras and having users upload their content in the form of one minute video interviews of people from the conference

The big question at the end of the conference is how to harness the two-way-communication marketing trend. How can we make these all these conversations actionable? What do you think?

Below are some great resources folks have posted to document and comment on the conference. We’ll update this post as more materials become available:

Presentation by Michael Donnelly, Coca-Cola

Culture Buzz Video Aggregate

ESCP Europe’s recap video on Facebook

Hilarious video (“Send Us Your Reckons”) shared during the panel about the role of journalism in a Web 2.0 world