In September I heard Marie Wallace, IBM Analytics Strategist, talk at the TED@IBM event about her work for her internal clients at IBM on identifying and measuring the impact of their employee-influencers. Her work is nothing short of amazing, especially given her European vantage point (she lives in Dublin) and its consequent view of privacy. We had a conversation recently about her work and what it means for the total realm of "employee advocacy."
Your role is to interpret employee data to provide a more complete picture of the networks that IBM employees have and the value of those networks, ultimately, to IBM business.
Yes! And there are many ways the business could derive value from this data. The solution that I am currently working on (and which I discussed at the TED talk and which you can also see at http://www.ibm.com) is a Social Reputation system where the purpose of the analytics is two-fold: (a) to help employees have greater visibility on how they are doing within the internal socialsphere and to use this insight to help them realize greater value from their organizational networks, and (b) to help management better understand how their employees collaborate to get business done, what that tells them about their organizations, and what can they do to foster a healthier and more collaborative workforce. The vision is to allow this data and insights to be leveraged across many parts of the business.
Can you tell me a bit more about the data points that you review?
Our first data source has been IBM Connections, which is the social and collaboration platform that we use inside IBM. We piece the data together into a large Enterprise Graph that connects everyone together. The types of "indicators" (actions) that are represented in the graph (and used in our analysis) are things like "Marie created Document A," "Jane liked Document A," "Jane shared Document A with Frank," "Frank commented on Document A," "Frank tagged Marie with Analytics," "Marie followed Frank," and so on. All the verbs that connect people are candidates to be added to this graph. We have lots of requests from our IBM employees to include more data sources outside of IBM Connections, such as Sametime, e-mail, and even our new IBM Verse.
And how are you using that data?
There are several projects across IBM that use IBM Connections data for doing everything from social search to content recommendations. However, for my project I am specifically interested in Social Reputation / Eminence, and therefore I am analyzing the data with a view to understanding four engagement measures: Activity (what you do), Reaction (how others respond to what you do), Eminence (how they respond to you), and Network (how effectively are you connected and what is your role in the network?).
Once we understand this, then we can make recommendations for how employees can be more effective and can provide management organizational-level insights around these engagement measures - i.e., to provide more relevant content-sharing for the employees? to determine conversions and predict outcomes? And so forth. The reason for building out this Enterprise Graph and providing analytics capabilities on top is to enable different parts of the business to use it in different ways. So HR might use it for designing learning programs and personalizing to individual needs; marketing might use it to identify employees who have the potential to become external company advocates.
Your location in the EU means that you are more aware than most IBM execs of privacy concerns for the employees. How have you navigated these concerns?
We've taken a really conservative approach to privacy to ensure that everyone feels comfortable. There are some basic principles (such as transparency, open dialogue, collaboration) where our Users actively participate in the evolution of the analytics system. However, the most important principle is that personal analysis is private and confidential to each employee. They get to choose if they want anyone to see their personal scores, and they are increasingly opting to do that.
Moving forward, we will likely investigate flexible approaches, such as brokering, that will make it easier for employees to share their analysis with different people and/or programs (such as advocacy, learning, mentoring, or even career development programs) if they so choose. The organizational analysis is de-identified so that it gives management the benefit of insight, but allows the employee to use our social and collaboration tools without the fear that Big Brother is watching them. This also removes the inclination to excessive gaming of the system.
What has worked? What hasn't?
Our approach has worked because we've been so employee-centric. We have a very active community of Users discussing where this solution should go, and we engage them in all the big ideas before we implement them. This has built what I believe to be a solid and trusting relationship, which is critical.
How much of IBM's employee base is in your purview?
As a technologist, my goal is to investigate new technological approaches to existing or new business problems; as Analytics Strategist for IBM's Software Group, I build analytics systems that IBM and our customers can use. IBM is frequently my first customer, and this is the case with this social reputation system. We launched the solution as a pilot inside IBM earlier this year, making it available to any IBMer who requested access. Since then it has gradually increased to over 3000 IBMers globally.
How much further do you see the program growing?
The potential is endless as there are so many insights that can be derived from analyzing human interactions. The biggest challenge will be in designing the business processes to most effectively leverage these types of insights, and to put in place the appropriate controls, such as privacy, to ensure that the insights are used appropriately. This is a big challenge for the entire analytics community when it comes to people-centric analysis.
What tools do you use for listening, for collaboration, for measurement?
Internally within IBM, our collaboration tool is IBM Connections (enterprise social network), Sametime (instant messaging), and Notes (e-mail). We have a number of solutions that analyze our enterprise social network. My solution has more of a focus around analyzing the network itself; however, we have another really nice solution called SocialPulse, which analyzes the content and provides heat maps to allow you to better understand organizational sentiment around topics.