Corporate Social Media -- Lessons Learned

If you're a regular reader of this blog (or an EMC employee) you probably know I'm running a corporate social media initiative at EMC, among other things to keep myself out of trouble.

Over a year ago, I became dazzled by the transformational potential of social media as a business tool. I was lost, as many people are, in a babble of buzzwords, trying to figure out what to do.

Thankfully, I regained my composure, figured out a reasonable plan, and started hammering away.  Not only are we making delightful progress, but we've learned some lessons that aren't really talked about in the usual stuff you read.

I'm writing a more detailed blog on our journey, since I know that others will follow at some point, and -- like an explorer probing somewhat new territory -- I think my notes will be useful.

The Pitch

Social media techniques (blogs, wikis, forums, etc.) can fundamentally change the way people work.  This is true for company-oriented stuff, but is especially true as we work with customers, partners, and other folks in our ecosystem.

Last summer, we realized we had to start working towards getting good at this stuff.  For a variety of reasons, I was volunteered to lead the initiative, among the other stuff I do.

Some Basic Assumptions

Some of the best work we did was at the outset -- we studied the problem, listened to everyone, and made some fundamental assumptions that we thought were absolutely key to our success.

The first assumption was that social media was almost entirely based on people's skills, attitudes and behaviors, and had very little to do with platform, features, etc.  Success at social media means that people are communicative, transparent, respectful, engaging, comfortable with the environment, etc. 

As a result, we didn't spend too much time arguing about technology, despite the natural tendency to do so.

The second assumption was that this was a business thing, not an IT thing.  Yes, we needed help from IT, but if the business didn't own the strategy and the outcome, this wasn't going to work.  So we built two teams of people to tackle the problem: one on the business side, on on the IT side.

The third assumption was simple but essential: it had to be fun.  People are social because it's fun, not because their boss told them to do it.  We had to create an environment where people wanted to do this stuff, and not encumbered by process, policy or expectations.

The fourth assumption was that it made sense to practice this stuff in a relatively safe environment, behind the firewall.  Doing this would give us a two-fer: not only would we see benefits from internal improvements, but we'd be building the skills and behaviors we'd need outside the firewall.

And, trust me, social media proficiency is all about people.

What Happened

We put up our first pilot platform in Sept 2007. 

It's interesting to note that we didn't have to do an extended paralysis-by-analysis exercise on this.  We thought this would be a good thing, it wasn't too expensive, and we knew we had to learn.  The approval process wasn't a big deal, as far as such things go.  I think that says a bit about EMC's culture -- the willingness to try new things and make a few mistakes along the way -- another reason I like working here.

Like most first platforms, it was initially unstable and more buggy than I would have liked.  Thankfully, we didn't do a big launch, we went instead for viral marketing for several reasons that proved to be good decisions.

Lots of people came by to check out what we were doing on the intranet, but only a few stayed.  Those that did loved the platform -- they were merrily blogging, discussing, etc.  But there wasn't a lot of these people at the outset.

By November, the platform had stablized, most of the annoying bugs were fixed, and we had some genuine communities up and running.  Participation had grown, but it was small, incremental growth.  We knew we had something cool, but we were concerned that it wasn't catching on like we had hoped.

Something happened in early January. 

I don't know if it was the psychology of the new year, or someone said something at a big meeting, or people just got more comfortable with what we were doing, but the darn thing just exploded in terms of participation.

New communities are springing up on a daily basis.  It's a big, brawling, cross-EMC conversation-fest.  The platform's RSS feed is so overwhelming I've just turned it off.  We don't have to look hard for business-value-creating examples, they're everywhere.  People are blogging, discussing, wiki-ing, etc. with abandon.

We've created a monster -- a successful, value-creating monster, but a monster none-the-less.  And, to be fair, that was our goal.

So, Where Are The Challenges Ahead?

We're now starting to see content -- lots of it.  Yes, it's generated via wikis and other means, but you look at some of the stuff, and you see that it has definite business value.

The challenge is that the platforms that are good at the social aspects of community interaction aren't the same ones that are good at managing content on an enterprise basis.  The same is probably true the other way around.

We're going to need industrial-strenght content management, and fast.  We're going to want a common backbone that connects content and workflow to other parts of the business (e.g. other web portals and repositories).  We're going to want a capability to tag information more granularly to determine who can see what, especially as we venture outside the company.  We're going to want the ability to archive and ILM, since not everything that's created is especially useful, but a surprising amount is.

We're working with our vendor to see if we can build some simple hooks into Documentum.  Yes, of course we want to use our own products, but -- even if EMC didn't have Documentum -- we'd have to buy something very much like it and do the integration ourself.

Is this need for enterprise-wide content management unique to social media applications?

No, not really.  At one level, it's just another use case.  Same as our email, or our bazillion web portals, or our gajillion ad-hoc repositories.  Important content needs to be identified, leveraged and managed. 

Our social media platform is just another content generator, albeit a prolific one.  And, if current trends continue, we're going to see a signifacant majority of high-value content being created in this environment before too long.

A second challenge is rather subtle, but important, and that's hosting infrastructure.

At the outset, everybody thinks in terms of a big, sprawling uber-community, which implies a single instance of the platform software, usually running on a dedicated server, often with a clustered failover.

But, if you think about it logically, successful communities work because they target very precise, passionate interests from their community members.  If you want to discuss CLARiiON issues, you're unlikely to go visit the Documentum forum, as an example.

This leads to a world of many communities, especially as we venture outside the firewall.  In EMC's case, I can see dozens, maybe many more.  Each with their own platform requirements that aren't entirely homogenous.

Now, you do the math.  Imagine 30 or more dedicated servers, each with their own hot failover, each with their own self-contained database, etc. sitting in a hosting facility somewhere.  That's really expensive.  So expensive, in fact, that there's no way that's going to happen in most corporate environments.

What's worse is that the resources are almost entirely wasted.  Given our experience on our internal platform, plus some experience outside the firewall, these are NOT busy servers in the least. 

What's obvious here is that the right model is to take a single server, run VMware ESX, and load it up with 10 or more community platforms.  All of these communities start small.  Some don't take off, but some do, and if one does, move it over to a more powerful environment.  Duh.

Now, when we set up our platform behind the firewall, that's exactly what we did.  The EMC IT guys are very comfortable with VMware.  And, of course, with me as the business sponsor, I thought it a good idea as well.

But when we go talk to our software vendors, or our hosting partners, you'd think we asked them to do unnatural acts.  I was shocked by the lack of familiarity and comfort most people had with a virtualized environment.  I guess maybe I'm a little too close to the action, but it was stunning.  I'm finding myself explaining the basic concepts of server virtualization (circa 2004) in 2007.

A third challenge is the security model.

Imagine a person who wants to participate in a bunch of communities.  Maybe they're outside of EMC.  Our current model is a distinct security domain for each community.  And, if we're trying to incentivize people to become active participants, who wants to keep track of a whole bunch of username / password combinations?

Farther out, we're going to use some of our own technology to solve this problem.  We'd like a world where people identify themselves once (and once only!) and then use that identity to drive what kinds of information they can access.

Going with parts of the Documentum and RSA portfolio will give us a couple of interesting capabilities.  We'll be able to flexibly tag any piece of information in the environment.  We'll be able to wrap especially sensitive stuff in DRM if needed.

We'll be able to adjust a community member's priviliges very dynamically using an object-oriented model.  And, most importantly, we'll have the ability to audit who's seen what.

Now, bringing together all those technologies into a workable whole for my particular project probably won't happen overnight.

But, speaking honestly, I also know I don't need this yet -- it's still a ways off on the planning horizon.  But it's nice to know we've got the capabilities in-house to really manage this stuff at an extremely sophisticated level should the need arise.

This Is A Big Discussion With Customers

You probably know that I spend a fair amount of time with customers and partners, and usually I have to explain the different things I do at EMC.

Maybe you'd be surprised (or maybe not) on just how often when I discuss that I'm working on this topic, the senior person in the room does a real-time redirect on the proposed agenda.   

More than a few senior IT people are seriously thinking about knowledge workers, new collaborative models, and the new shape of IT.  Many of our partners in the service provider and outsourcing world want to get into this space as well. 

And they really want to talk about it.

So, I'm spending more than a few cycles sharing our experiences and our learnings.  Yes, I'm sure a few of EMC die-hard competitors will learn a trick or two from this, but that's unavoidable.  What's better is that EMC can play two roles helping customers and partners be successful at this stuff -- not only as a provider of products and services, but as a case study in how to take a large, sprawling corporation, make it proficient at social media, and see the results.

But, Let's Not Forget, This Stuff Is Fun -- Or Should Be!

So much of what we all have to do in IT isn't particularly fun.  Making technology work.  Making budgets work.  Making vendor stuff work.  Making people understand how important all this stuff is.

I have to say, all of us associated with this social media initiative are having fun.  Not only that, we're delighting users every day -- we're providing an environment that's fun to use, meets some important unmet needs for them and the business, and is kinda cool to boot.

Now, if we can figure out how to make the rest of IT just as fun, we'll have a winner!

Comments

[Lee White]

Chuck

It was great reading this, as I spent much of 2007 trying to do the same thing with my former employer, GSK. Unfortunately I kept running into the IT Machine that dictated that the only acceptable solutions must include the terms "IBM" or "Microsoft". I hear that they are making some headway since I left, but it will still be a slow journey.

As for me I am completely aligned with just about everything you highlighted in this post. Looks like you are creating that elusive case study everyone is looking for. I know I will keep tracking your outcomes so I can use you as an example of how this stuff CAN work.

Thanks.

Lee

[Chuck Hollis]

I think early on that we realized that if it was a business outcome we wanted, we'd need to present ourselves as a business unit, and engage IT on that basis.

We were lucky working with EMC's IT group in that they worked with us in some non-traditional ways to get to the right outcome, rather than some arbitrary standard.

I just hope others will be as fortunate as we were in this regard.