Community and social computing
rawn 100000R0P5 Tags:  social-computing scalability population social-software blueiq e20 size 1 Comment 7,484 Views
On returning from the recent Enterprise 2.0 Conference in Boston, I had time to reflect on the scaling issues that come up in social software adoption across an enterprise. In watching Gentry Underwood's excellent presentation on how they designed the social computing environment in IDEO, I tweeted to him that new issues start to pop up when you move from an enterprise social environment for 500 people to 200,000--or in IBM, nearly 400,000 people in 170 countries. This is not a bragging point, rather a one of frustration.
There certainly are other large or technological-oriented companies deploying social environments, but from my experience in hearing from others, no one has hit some of the scale issues that we have in IBM. Obviously we are talking about an enterprise's deployment rather than a social site like Facebook; they're very different issues for each.
For one, while employee profiles and directories are starting to become commonplace in other enterprises, we have had one for well over a decade in one form or another. We've already gone through the issues and practices others have found: (a) include everyone; (b) prepopulate with relevant contact, work info, projects, etc; (c) popularize it as the place to look up data; (d) integrate into or make it THE basis for contact info for other existing internal and extranet Web apps; (e) invite partners,contractors and suppliers; (f) tie to enterprise-wide LDAP and single-sign on; (g) integrate into common work processes and behaviors. In fact, the last I looked, we had nearly 600,000 profiles in our Bluepages (including employees, supplementals, contractors, bots, some partners and suppliers).
While the Bluepages system certianly popular, it is but one of several dozen commonly used social software tools, some of which in themselves have hundreds of thousands of unique users. We have thousands of smaller communities and wikis some of which have tens of thousands of members. The multiple tools comes out of our laissez-faire attitude to allow many software ideas to emerge, and through our user base test and advocate the best ideas.
The population size of this system isn't quite the issue, but I put some thought into what enterprise 2.0 deployment issues might appear with scale and came up with the following chart. I hope this can help other maturing e2.0 environments consider some of the issues they may be coming up agains
A. distribution of people across time zones
B. distribution of people across cultures / countries
C. distribution of people across physical locations
D. distribution of people across job categories or dissimilar job roles
E. projects people work on are very different in nature
F. distribution across access devices (desktops, laptops, mobile)
G. many separate (non-integrated) social tools, or different interfaces
H. many separate databases as information sources
I. many separate or isolated social instances
J. number/reach of boundary-spanners
PS: Thanks dW folks for migrating my blog over!
Per Gartner's request, I have removed this post. Gartner does not permit comparisons to prior years' research.
I tried the MIT Personas site just to see what it would come up for my name. Basically, it does a Web search for any name you give it and analyzes the text on a semantic level to find common themes or categories for content from you or about you (e.g., it might suggest topics like "sports", "legal", "social", "management", "online", etc.) How it does this is a little beyond my knowledge of semantic processing, as well as what practical use this is seems to escape me.
But it's fun. :)
Some issues with it, is that because it searches on people names across the Web, the content may mistake e.g., one "Luis Suarez" (our own blog-evangelist) with another "Luis Suarez" (soccer player). It helps if you have a globally unique name like "Rawn Shah" (so far). Another problem is that the results may vary depending on how it processes whatever search results it finds. So, in trying it out three times to see how close it was, it showed slightly different categories for me and sized these differently as well.
I can't figure out why the following categories appear since I rarely if ever talk about them: military, religious, religion, genealogy
Lee Provoost’s post, “Adopting Enterprise 2.0 in large organisations: Fiat or Ferrari?” talks about how people can start with smaller
cars like a Fiat and eventually upgrade if need be to a Ferrari, rather than
wait decades of riding public transportation until they save enough for their
top car. I don’t see the entire negative with public transportation but when it
comes to social software, this ignores a large problem: migrating from one
social software system to another is a lot more complicated than just replacing
the tool itself.
Having seen first-hand several generations of social tools
in our company, and trying to get people to migrate from one existing
environment to a new one, it takes a bit of work to get people to switch to the
new system. A better analogy than cars may be “transportation networks”. In the
US, think about asking people to stop driving and using trains instead.
First of all, change is hard: people become used to certain features and know how to work it quickly. Unless the new social tool has the exact same features handled exactly the same, it means new learning, often new terminology, and trial and error.
Social tools also don’t always make it easy to migrate from one system to another. So you may also need to reenter your profile information, preferences, and generally reinstate what you may have already had before.
People place a lot of content and context into their social
environment, and unless that is all migrated with them too, they may see it as “loosing
their standing in that social environment.” For some this is in the form of
rankings; for others useful or valuable content that they left behind. New
social environments don’t need to start at zero but more often than not, they
are not fully compatible with the old ones or provide different tools and
require different fields; thus, migration is not a simple prospect of
Finally, a new system should probably not only perform better,
but all them to interact in better ways. This also means new features to ask
people to try out. The power of the new social tool may be in those new
features but in maintaining the status quo, many users will keep using what
they know, until enough people have adopted the new features.
These are just a few basic reasons in adoption that make it
difficult to simply “level up” to a new social system. If you run a social
software system in your enterprise, you should certainly not treat it like just buying a new car.