Michael Brzozowski, the creator of Watercooler, the internal social media system for HP, recently left HP for Google.
Talents move around all the time, especially in the bay area where the industry is rife with interesting opportunities. However, in this case, the departure of Mr. Brzozowski has put the fate of the Watercooler system in question.
To understand why this is worth blogging, we need to first understand what the Watercooler system is about. Many of you may not know this, but Watercooler is a social media system that currently has 100,000 users! Brzozowski originally started Watercooler aggregate RSS feeds from across the company. Overtime, it has morphed into a social media aggregation platform that aggregates content from HP’s internal wikis, microblogs, various discussion forums, and social bookmarks. The system has a documented set of open APIs and supports a powerful and expressive set of content filters across different social media systems. It is also integrated with HP’s user directories.
Brzozowski wrote a nice paper on a study he conducted with Watercooler data. Published in Group 2009, the study revealed some interesting facts about social media usage inside HP. Perhaps one of the most concrete statistic to date arguing for the value of enterprise social networks, Brzozowski’s paper, points out that 69% of all Watercooler blog users subscribe to content generated by someone outside their business unit. This kind of cross-company instant collaboration is a huge benefit social media system provides its user community.
Unfortunately, though Watercooler can be considered a success from HP labs, it has not generated the kind of support from HP proper. Brzozowski has been trying for the last 2 years to get the system out of HP labs and into the hands of HP operations. But his efforts proved futile – HP operations were not interested, or at least not interested enough to take actions. After Brzozowsi’s departure, another researcher from HP labs took over the system. But this person is only doing it on a volunteer basis — he’s got his other core tasks. As we all know, researchers are not great maintaining production systems, especially one that requires such scale and performance. Now you might ask why HP would ignore a social media system that’s already got such a large user base? Do you know how many social media start ups would kill to have 100,000 users? Well, perhaps only HP can answer this question.
This whole thing came to its head a few weeks ago when some of HP’s executives were meeting with SalesForce. The latter mentioned Chatter, the new Social media system SFDC is launching at DreamForce this week. Chatter is a cool system, but is not nearly as developed or as widely used as Watercooler. Especially when you consider Watercooler had supported a documented API for users to modify for their own purposes, pro
The HP executives, after meeting with Salesforce, said about Chatter: “Hmm, that’s a good idea, we should have something like that.” [obviously this is a mock conversation, not the dude’s actual words]. Finally someone in HP said, “Well, we do have something like it, it’s called Watercooler”. The executive then said: “Really? Well, let’s take a look at that. Maybe we can make something out of it”.
As if on cue, Watercooler stopped working because the whole system had been running on one server (what? One server? You asked. Yep. You heard right, one server to support 100,00 users. That’s how Research Labs typically work). The researcher who had been supporting it after Brzozowski left was unable to get it up running again quickly.
HP labs had many top industry talents, but these people are now leaving the organization, for the reason that their work has not been properly respected and utilized. Last year, they lost one of their HP fellows, John Wilkes, to Google. In addition to the recent departure of Michael Brzozowski and Kevin Lai, a game theory specialist, Joe Pato, a noted Computer Security expert, though ostensibly still an HP person, has been spending most of his time at MIT. HP has come a long ways since the garage company days of Hewlett and Packard, but it seems like the company has lost some of its innovative spirit along the way. Yes, it’s difficult to remain innovative when you’ve got 30,000 employees. But people are the greatest asset of any organization, if you lose them, you lose the future of the company. This is why Google recently implemented measures of 10% payraise and bonuses to retain talents against the new-kids-on-the-block competitors like Facebook. Companies like HP should take notice. Innovations like Watercooler should have flourished instead of being left to flounder.
Chenxi Wang's blog 
This is typical of any medium/large organization, the do not encourage cross department interaction, its not part of their core process. Management are focused on current projects, and anything that does not immediately show benefits to those projects will be discounted as irrelevant. Executives forget the power that casual conversations between colleagues can bring to current and future projects. When something like this is presented they will always dismiss it because it has no apparent benefit the the bottom line.
Completely agree. In this case though, HP could have done something with this system, which is an entirely HP invention, rather than shelling out cash to acquire companies, as HP has done so aggressively recently.