continue reading hover preload topbar hover preload widget hover preload

November 16 2011 Posted by: Mark Fidelman in: Enterprise 2.0

How to Improve Social Business Adoption as Told by IBM (Part 1)

I typically like to provide context around discussions with industry thought leaders and Executives of the Fortune 500.  Yet some discussions should simply be left untouched and published as-is. It’s not a short article, but the insights around IBM’s journey to becoming a Social Business is remarkable and a must read for those that are in charge of the Social Business program at their companies. 

Key Learnings:

  • When IBM experiments with new Social Technologies, 10,000+ employees try them out
  • The largest experiment,  Beehive, our Social Voodoo. had nearly 90,000 internal early adopters
  • IBM Communities can contain between 12 people and 50,000 people
  • The Social Maturity adoption plan at IBM is itself evolving
  • There are no reliable benchmarking metrics available for Social Business
  • IBM measures a lot of engagement activity beyond adoption metrics


Key Questions Answered

  • How important is adoption in becoming a Social Business?
  • How to track adoption and engagement metrics.
  • What impact is it having on IBM’s culture?
  • Why Big Data will be important in becoming a Social Business.

#1 Fidelman: Rawn Shah, please tell us about yourself and your background

#1 Shah: I’ve been at IBM for probably 11-12 years now and in the early years I was a contractor. Before that, I was a journalist and I wrote for a number of different technology magazines like JavaWorld and LinuxWorld.  

My most recent book is, “Social Networking for Business”, which I started around 2006 when I was the Program Manager for the DeveloperWork’s community at IBM. The Developer site has about 9 million individuals and I worked as a Program Manager both on the content side as well as the community management’s side.

After that I joined this internal team where we are focused on trying to involve all the IBMers in all the different ways that they work in Social, including working in a Social collaborative environment.

Interestingly,w hen people jump on in IBM for a new project it’s in the tens of thousands. It’s not like a dozen people try it out. I think one of the largest experiments had about 100,000 people join in.

#2 Fidelman: Can you expand more on how IBM leverages communities?

#2 Shah: Those community scaled at, I think they’re about 20-30,000 communities. Some of those communities, I think the biggest one I’ve seen has about 50,000 members in it. Other communities go all the way down to like less than a half dozen.

For example, the IBM Software group community, which is under Steve Mills, our Senior VP. He’s the Group Executive of all the brands that we have; everything you see from the collaboration to the database systems to the Rational products to the Business Manager – everything that comes under there. They created this community to move away from standard portals of communicating information to everyone. So only from having a webpage to actually having a community that people can participate in.

Steve Mills now does quarterly updates and kind of regular updates onto the community himself and so does his Executives; his Executives do the same.

#3 Fidelman: Does IBM follow a Social Business Maturity Model?

#3 Shah: I gave a presentation at the Dachis Social Business Summit back in March on Social Business maturity. What does maturity actually look like?

So one of the topics was… People come to maturity and they think of it as large numbers of people using it. We already have large numbers of people. That’s an easy thing for us. What percentage of the company is using it? We look at that as an adoption maturity. You are getting enough of the organization participating in the environment. So that is one particular type of maturity, just in terms of adoption.

As we have grown over the last four years we’ve also been evolving our program. How do you actually run a program to do adoption? How do you run a program to engage people in Social Business? That program itself is achieving maturity. What kind of staffing do you need? What kind of topics do you cover? How do you present it? How do you enroll volunteers? What do you do with the volunteers? All of those is program aspect. So that’s a second aspect of maturity.

The deeper one is where you start looking across the fold of what is the quality of how people commit to each other? What is the quality of how people learn from each other? How do they organize into groups? How do you organize across the company into Social groups of different sizes with different purposes?

#4 Fidelman: How do you track the answers to each of these questions?

#4 Shah: All of those are different levels; they have their own independent levels of maturity. So you might be very well connected to organization and you might likely also be doing a lot of Social learning from that, but if you don’t have the infrastructure to support decision making or organization then you might not be as advanced with that aspect.

So these are what we call capabilities. Capabilities are things that individuals can do. So you can actually go out and share to provide that Social learning aspect. You can go out there and connect to people. There’s a second part to it which is quality. So if capabilities are verbs: to do something, then qualities are adjectives. You do something in an open way, a transparent way. You do something in an agile way. You can connect to the very slow process, go through a chain of command until you eventually connect to someone else or you can just jump out there and access it directly. That’s a very different nature of transparency. It’s also a different nature in agility.

#5 Fidelman: How is that impacting IBM’s culture?

#5 Shah: You have that level of comfort in an organization where anyone can approach anyone without repercussions of any sort. So organizations view that in different ways. Sometimes it’s not necessary that everyone have that level of transparency but imagine an organization like the Intelligence Agencies, right, and the Defense Industry. Sometimes you don’t want everyone to know who everyone else is in the organization. You don’t want them to know what’s going on in other parts. That’s part of your culture; that’s part of the design for your organization. At the same time you can still allow some bits of collaboration.

So the key qualities we look at are the same things you might have heard from Jeff Schick. We look for agility, nimbleness; we look for transparency, how visible do you make it? We look for engagement. How many people are participating? Not just in the environment but how many people are actually engaged with the company in this form? That harks adoption; that harks responsiveness – everything that you actually want out of your collaborative environment itself.

#6 Fidelman: I think the troubling aspect, and what most people want to know, is how are you actually measuring these things? It’s one thing to say you’re transparent but is there a benchmark? How do you measure transparency for instance?

#6 Shah: That was actually a very, very difficult aspect of that. The measurements are hard. We look at measurement in two ways. There’s metrics categories and then there are collection methods. So metrics categories – there are a number of different varieties of those. Things that describe the structure of network, things that describe the quality or influence of people there: brand awareness, individual reputations, things like that.

Then there are methods of collecting the data. You can do that from activity logs, right? There’s a data log from some service somewhere. You can do that from surveys that you put out there and you get opinions. You can do that from interviews and focus groups and data log visualizations.

Let’s put an axis there, two main axes of these things. One is that metrics are behavioral meaning it’s something that someone has done or attitudinal in Social analytics. Attitudinal means it is opinion; it is views; it is things like ratings. These are all attitudes towards the information itself. Both kinds can either be qualitative or quantitative. Quantitative metrics I think we all understand. Qualitative metrics are more on the variable scale.

#7 Fidelman: Can you provide an example?

#7 Shah: Let’s pick a case study. A case study is much more detailed but a success story is a simple anecdote about how somebody uses something and what they use it for and why it was successful or useful to them. So that is actually a very qualitative piece of information though it may have some quantitative data points for it. That is one type of metric with two different types of data collection methods behind it. A final slice that we do is demographics. It is all of the above and then you slice it by the different types of demographics that you’re interested in.

So some of the analytics work that I do for our group is primarily adoption and we’re looking at demographics across the company of what does adoption look like in China versus U.S.? What does adoption look like in Software group versus our Consultant services?

#8 Fidelman: How do you know that it’s going well? I mean you could see quarter after quarter increases but do you really know you’re being successful unless you’re benchmarking against another company or by placing a stake in the ground that you’ve placed in the future?

#8 Shah: That’s the benchmark standard issue. For us, we can use the demographic within organization to see which group is ahead of which other group. There’s no single hallmark of what is best but it’s a running count of who’s in the lead or who’s behind. That’s useful information as it tells us, “Hey, we need to put more effort into adoption in this part of the world, in this type of job role, this division, things of that nature.”

If you want to compare against other companies, other organizations… So inside our company we can compare between team, business units and job roles. So they can see if Sales are doing better at collaboration than Marketing or Developers or any of them. There are different Executives who are interested in that view.

So against other companies is more challenging. What we need to do is run a service with other companies where you ask a number of survey questions which you have identified as key indicators of employee engagement. Then, run that survey with many, many different companies. Armed with that information, they can now start creating an industry benchmark. We have had hundreds of sets of data from companies from this industry. This is where you stand relative to the industry. This is where you stand relative to your geography.

So if you really want to tell across industry, you need a sort of neutral fashion which is not based on the number of people, which is not based on the type of industry that you’re in necessarily. Right now tech industry folks are really high into collaboration and all of those things because they have the tools. It’s part of what they do. Collaboration, when you’re talking about Social Business, is more than just the tools itself. It’s the attitude; it is the view of how people in the organization work with each other. In that sense collaboration has existed forever, before we had computers. How willing are people to talk to each other in an organization? How willing are they to share information? We just look at it in the terms of Software tools making it faster, accelerating that medium itself.

It certainly opens it up and we call it democratizing the ability to exchange information. So now think of that in Social Software in that same context within a company.

You benchmark parts of an organization or even entire organizations and then basically you have to evolve it over time. You don’t say, “This company is a standard, everything else is measured against it.” It’s a moving target. It’s a constantly growing, evolving, changing thing. Does that make sense?

#9 Fidelman: Have you published any benchmarking studies yourself or anyone within IBM that I could link to or point to?

#9 Shah: We have published different kinds of information. So this information itself is new. We are still working on it. When we are ready with the full form, we’ll certainly go ahead and publish that. We have published other kinds of data which is the, like I was saying with the metrics, we have different views of metrics. Departments need different kinds of metrics so this is, again, demographics. The Marketing, we want a different view of metrics than HR or Sales.

There are two ways of looking at it. So first, we look at it from a programmatic way. So programmatic tells us about our program obviously but it also tells us about our primary constituents – and the different Sales groups. There are tens of thousands of people in there and we want to understand how the Sales group is doing versus another group like Development. We give that view to our Sales Executives.

Now the second way of looking at that is rather than just focus on adoption,we’re looking at the capabilities of tasks. How are people connecting? How are people finding experts or expertise? We’re looking for expertise not necessarily just people. So it’s knowledge, it’s people, it’s all of the above.

Specifically,  How is that task, that’s one of those core tasks fit in the collective collaboration? How well is that going for the organization? Are people finding it easy? What are the different dynamics of that? Is it easier to do? Is it faster to do? Is it universally available? Is it actually responsive? Are they getting desired results? Is it integrated into the workflow? Is that something that people find easily that they don’t have to go think about it and jump off to a different point to go do it? How much of it becomes part of the process?

#10 Fidelman: Are you aggregating metrics across Connections, Notes, maybe mobile, any kind of Social stream actually and doing some kind of a global analytic research?

#10 Shah: Yeah so we do that primarily for the Software tools rather than the other mediums like the folks, mobile devices, things of that nature. So I developed a tool where we can look at it across our internal Connections environments and number of other primary collaboration tools that we have. What we do is we pick a demographic that we’re interested in. So let’s say Sales. We look at that population and all the different places that they’re interacting. We don’t use individual information itself; that really isn’t very useful to us.

We are looking for groups; we are looking for demographic units. Those units could be anything from looking at how China is doing versus how the U.S. is doing. How are people in a Sales Engineer role versus the Client Representatives doing in collaboration? What am I looking for? I’m looking for adoption metrics. I’m looking for how regularity. That’s a primary metric that we look for. Further, do they do this on a weekly basis? Do they do this on a monthly or quarterly basis or once a year or not at all?

We consider a weekly basis that means they’re really active. They’re doing something. They could be doing a daily basis. At that point I don’t think it really matters. Weekly is enough to tell you if they are actually part of the system itself, the environment itself.

#11 Fidelman: What are you using to measure that? Like a specific tool?

#11 Shah: We have the internal tools that allow that. They’re the same tools that go into our products as well. So SaND, for example, is a Social search system. It’s also a database engine in itself, a database engine of every kind of interaction that’s going on in the environments. They get down to the specifics of are people commenting versus are they downloading versus are they tagging and sharing? Things like that. What kind of action they’re doing, where they’re doing it and who are they doing it with, when are they doing it. Then we rollout that information. We’re not interested in any single person because there’s just too many privacy issues with that. Any individual is not important to us.

From a program function and an Executive role – they want to know how the entire group is doing.

#12 Fidelman: You’re not interested in finding experts, specific experts on specific topics then.

#12 Shah: No, we’re not necessarily doing that for the purpose of that. The SaND tool can actually do that just through the search. So you can type in a keyword and get experts for that.  The analysis that I’m describing is what Managers and Executives are interested in. Sure, you could show me a list of the top 10 experts but that’s not really what I’m interested in. I’m interested in how the entire population is doing.

#13 Fidelman: This seems to be a big data problem. I can’t see the average company being able to aggregate all this information, especially if you’re talking about email. Aggregating all the Social feeds that we talked about and eventually bringing in mobile and then doing some kind of analysis on it. It seems like you’d have to have a pretty powerful engine in order to process that information. How do you do it with IBM?

#13 Shah: I’m glad you brought that up because these are all big data problems. For us it’s obviously a bigger data problem with 400,000 employees. Think of all the possible networks in there and all the possible non-content that they produce. Others may not be as much but it just depends on the scale.

Now processing that volume of data can take time. Part of it that really makes it simple for us is the categorization of activities; that’s what SaND does really well. SaND also does the indexing of that which makes it easy to search. There are tools built on top of that which helps you visualize that like you can actually visualize Social network. I think we had that for several years.

For us, the even bigger data problem is sentiment analysis…


To be continued in Part 2 next week

  • Leave a Reply

    Required fields are marked *