A good article, but the thing I liked most about it was following the link to discover Dave Kellogg's superb blog post, Why Palantir Makes My Head Hurt:<p><a href="http://kellblog.com/2011/06/27/why-palantir-makes-my-head-hurt/" rel="nofollow">http://kellblog.com/2011/06/27/why-palantir-makes-my-head-hu...</a><p>which, among many other points written up in an entertaining manner, gives the better summary of how investors look at the difference between service and product companies:<p>>That last point [about whether Panatir sends in consultants or research engineers to its clients operations] is important. Why?<p>> - If field technical staff are engineers, then the associated revenue is presumably license fees and the cost is R&D.<p>> - If field technical staff are consultants, then the associated revenue is services and the cost is COGS.<p>>Why does this matter? Because most software company boards and investors see the world in a pretty black-and-white way:<p>> - License revenue is good. Services revenue is bad. (Largely because gross margins run 98% on the former and 20-30% on the latter).<p>> - R&D expense is investment and ergo good. Cost of goods sold is bad.<p>>Almost all Silicon Valley boards will want an emerging enterprise software company to run with a consulting business that’s no more than about 20% of total sales. In practice this means a company can have at most about 1.5 consultants (pre- and post-sales) per salesperson. Any work that can’t be done either as R&D investment or by that small consulting team needs to get handed off to partners.<p>So, do you want high growth or do you want to help your clients? The two do not always go together quite as well as you might have hoped. To rephrase the point in Dan Shapiro's terms, many of the most worthwhile businesses out there have reasonable, not outrageous margins.