...and you can count your leads, but you can't count on on how much money your leads will turn into.  At this morning's Social Media Breakfast the topic was ROI in Social Media. Brian Halligan of HubSpot and Matt Cutler of Visible Measures gave great presentations about how they track social media efforts.  Andrew McAfee, however, dropped a bit of a bombshell in saying that ROI from enterprise software cannot actually be calculated. Not that enterprise software in not valuable mind you...just that its value is too complex to effectively calculate with models.

This was a very interesting line of thinking for me for a variety of reasons. 

First, with the financial meltdown there is a lot of conversation swirling around about whether there was enough good old fashioned experience and intuition governing the big financial firms or whether they had become subserviant to their models which, apparently, told them that lending to people who couldn't pay them back was a good idea.  I'm being a bit cheeky here and I can't begin to know all the factors that led to the crisis but it has started a conversation around the use of models in decision-making and management.

Second, I spent a number of years benchmarking the supply chain and product development performance of Fortune 500 technology companies.  It was a fascinating exercise and from it I brought away two primary lessons.  One, benchmarking can be incredibly useful.  Two, people misuse and misunderstand the use of metrics quite a bit.  The metrics won't tell you what to do they will just show you where you've been and how that compares, directionally, to others.  It is one piece of information to use in decision-making but it should not be used alone to make decisions.

Third, I've build enterprise software ROI models.  There are a lot of assumptions you build in and a lot of exceptions that you have to ignore (and unfortunately for the models...business is mostly about managing exceptions).  I've also build system dynamics models which are incredibly cool and handle behaviors better than your standard database models.  However, the important lessons that good models provide are not the results.  Good models help you test changes to assumptions and see the relative/directional implications.  They will not give you an accurate projection.

So...where does that leave us? For Andrew McAfee, he suggests measuring what you can but trust that business decision makers are wiser than to trust any enterprise software ROI analysis.  But here's what I think.  Business is about relationships and the better the relationships you have, the more stable and secure your business will be...but you may sacrifice growth because scaling good relationships is difficult. 

The real value of social software is that it helps develop and maintain closer relationships with more people than you could without it.  The business leaders that have really succeeded get the relationship angle...if they can see clearly how social media can increase their ability to maintain more relationships, the ROI will be obvious.  But not calculable.