Over the last few weeks I’ve been thinking a lot about communities, and the different kinds of communities that are needed in a corporate environment today. Working through this in my head helped me put some recent discussions and announcements in context, so I’m hoping by sharing we can start a discussion on how to better drive and leverage these communities.
The first type of communities are the ones that are internal to a company. These constructs are used to help facilitate internal discussion, cross-functional communication and projects and even to get employees a place to vent, self-support or interact with each other. In my experience, these communities are generally located inside the firewall and are their success or failure is largely dependent on the tool used and how useful the implementation of that tool is to the discussions that are being had. Most large companies, and certainly the ones I have experience with, do a fairly good job of this kind of community. Interestingly, many of the most active communities inside Cisco, VMware and EMC have to do with supporting Apple laptops in a corporate environment, which invites a line of commentary that I’m not going to get into here. 
The challenge with these communities is that I’ve never found one where the participants have a sense of ownership in a broad sense. We participate in these communities, but we don’t belong to them in that we self-identify as a member of that community. It’s a resource, like a library, but not one that we see the need to spend a lot of time investing in. These communities end up being run by an off-shoot of the IT team, and while they are useful, we don’t necessarily miss them when we no longer have access to them.
The next type of communities are those that companies extend out to their customers. John Troyer who runs the social media platforms for VMware, gave a great presentation at the Kansas City Regional VMUG this week about the reach and scope of the communities that exist as resources for VMware customers. Obviously they have the incredibly popular VMTN Forums which average hundreds of posts a day, but it was interesting to me that John looks at non-VMware run sites like Experts-Exchange and others as extensions of the communities they have built organically. The feeling was never that VMware was competing with outside communities, but was looking to make sure that customers had all of the resources and information they needed available in whatever form they chose to consume.
EMC and Cisco also have such communities available, but the differences are noteworthy. Where the VMware communities and forums are literally the first two links at the top of the VMware website, EMC and Cisco position theirs as a part of their support services behind a dropdown at the top of the page. Both companies also provide a text-based link at the bottom of the page too, but the communities are not positioned front and center as social spaces the way they are at VMworld. While I haven’t been involved in this part of the Cisco strategy, I know that EMC has been working hard on enhancing their posture towards customer-based communities. Folks like Matt Brender and others are working hard to build a framework that can be leveraged, and Chad Sakac has long championed the “Everything X at EMC” forums as a way to interact with customers and partners.
One of the primary differences here is that in a best case scenario the customers own the communities. John Troyer surprised me when he acknowledged that there are no internal VMware FTEs dedicated to the VMTN forums; it’s completely run by the community, for the community. That kind of self-sustaining model shows that the participants really see lasting value in belonging, and are willing to give back.
Both of these kinds of communities are focused on the social aspects of connecting companies with their employees, customers and partners, but there’s another kind of community that is becoming increasingly critical in traditional enterprises: a community of code. Of course, these communities aren’t new, especially to the OSS world, but they are both new and terrifying to old-school hardware vendors. Dan Hushon shared with us the quote that OSS doesn’t succeed because it’s free; it succeeds because it’s transparent. That rings true, but understand that neither of those things (giving things away free or being transparent) are part of the DNA of large, multi-national companies. They just aren’t.
The need for this kind of community, in my opinion, one of the driving forces behind the announcement of The Pivotal Initiative by EMC and VMware. If you look at every part of the spin-out that has been publicly announced, each of them has a deep and built-in need to have a community of people that is specifically tied to the codebase or language involved. Greenplum needs the Chorus portion of the application, which allows for data scientists and BI specialists to collaborate and share on a project or company basis. SpringSource is, at its core, a community, and needs to have it to flourish. CloudFoundry, no matter what direction they decide to go in, need a closely knit and aligned set of customers and partners to allow that platform to fulfill its potential.
And those communities are inevitably going to need access to code.
The problem is that there are many logistical difficulties with giving away code inside enterprise structures that use a revenue model to relate to stock holders. You can’t, in those constructs, spend money to give something away and keep it away from the stockholders for long. Sooner or later, you have to generate a return on the investment that was made, or you need to close up shop. I have used this graphic before, from Nick Weaver’s fantastic overview of Razor at PuppetConf, but it’s such a succinct statement on the problem that it keeps finding its way into the conversation. Pivotal allows EMC and VMware to actually make strides towards opening up their portfolios to a larger degree of programmability without exposing their core business model. It’s a hedge, on a grand scale, to see how and where they can embrace open source without impacting the stock price.
While this was a large and significant announcement, don’t think it’ll be the end of them. The next shoe to drop was EMC becoming a corporate sponsor of OpenStack. How many people would have predicted that a year ago? My bet is that we’ll also see significant strides on the programmability of the EMC storage arrays, and the collapsing of appliance based storage features (VPLEX, RecoverPoint, Avamar, DataDomain) into the mainstream storage product lines (VNX, VMAX, Isilon) as software features. The use of vVols and other “object” based storage virtualization techniques will drive integration with VMware to new and interesting places.
And, my guess is that we’ll see Pivotal become a place where OSS and community-driven models can be tested and put into practice without running into legacy roadblocks of the process and people variety. Finally.