Late last week Socialtext's Michael Idinopulos wrote a post with some interesting advice for anyone looking to start a social software implementation in the enterprise: skip the pilot. His argument was that since the new breed of enterprise 2.0 tools are about human interaction, something which changes dramatically at scale, then small pilots were not a useful barometer of future success or failure.

Idinopulos admitted that pilots are great for traditional IT, which revolves around a set of actions that do not change much whether it's 10 or 10,000 people (think billing or adding leads in a CRM). We agree that in any kind of collaboration, the shift from 10 to 10,0000 causes dramatic change. But that leaves an open question: do you still use pilots for your wikis, blogs, and other social software implementations, or are they a waste of time?

In one sense it's a little strange for Idinopulos to question the need for pilots, since Socialtext's switch to freemium was pitched to us with pilots as one important use case. It makes us wonder whether the "free 50" version has seen less use as a pilot than the company expected.

But whether you're looking at a big social platform like Socialtext, or something much simpler, whether or not you spend time and energy on a pilot is an important issue. We generally feel that the premise Idinopulos presents is correct; interaction among workers in a tiny subset of your organization isn't a fair test.

But where does that leave us? Are pilots a necessary evil to support purchasing decisions, or a waste of time that gives an inaccurate picture of how enterprise 2.0 really plays out? We'd like to hear whether you love or loathe pilots in the enterprise.

Photo by totalAldo