During any CMS selection, it is fairly common to look at software products that span a wide range of prices — everywhere from free to several hundred of thousand dollars in up front licensing. Buyers invariably get confused as they consider vastly different pricing models and try to put those numbers in context of the whole project costs. They struggle to force different products into an "apples to apples" comparison. And, all the while, they are told by the vendors that each solution is the really the cheapest in the long run (bring on the tired free puppy analogies).
Content management software feels worse in this regard than other types of software selections. For one, there is no clear market leader in content management software. This means there is no single company to create a "gold standard" for feature/function/price. Secondly, it is not the case that the large vendors are more stable or offer more features than the mid-market and open source providers. In fact, it has been quite the opposite. The household brand names like Interwoven, Vignette, RedDot have all been in pre-acquisition balance-sheet-beautification mode over the past few years. In other words, they have been minimizing engineering investment and milking their customer base for revenue. Even during the Web 2.0 boom, what little engineering money was spent didn't make its way to web content management. With all of the high profile corporate acquisitions of Interwoven, Vignette, and RedDot, high end buyers didn't get the stability they were looking for either. So, the old adage "you get what you pay for" didn't hold up for the high end web content management marketplace. The middle tier of commercial software vendors have been delivering better products with less risk than the upper tier.
"You get what you pay for" didn't really hold up on the free end of the spectrum either. There are many examples of successful implementations built on top of open source software. These solutions were not entirely free. They cost money to implement and customize and many customers purchase support and maintenance but, then again, all content management software (commercial or open source) requires customization and integration work and maintenance can be 20% of licensing cost. My experience is that open source solutions are generally no harder or time consuming to implement than commercial software solutions (and I have implemented plenty of both). The way to reduce the cost of the entire solution is to minimize your customization/integration work by choosing a platform (commercial or open source) that best matches your requirements. It just may be that the closest match is open source.
When talking my clients through a selection, I am often put on the spot with a question like "what additional features do you get when you go with the more expensive products?" A similar question is "what will we get if we spend more?" These are hard questions to answer because there is no single feature that only appears in higher end products. Usually the best I can do counter with an analogous question "is a Porsche Cayman worth the $23,000 extra in price over a Subaru Impreza WRX?" Both have 265HP engines. Both have many of the same features (radio, air conditioning, speedometer, tachometer, airbags...) but they are very different cars. Obviously some people think the Porsche is worth the extra money. But I can't tell you whether it will be worth it to you because it is a personal feel thing. You get in the car and it feels right. You might even find that you like the less expensive car better because of the arrangement of the controls is more intuitive or the shape of the seats fits your back better. Of course, if you don't have $55,000 for a car, you can save yourself time by not looking at Porsches.
Open source software can be the same way. You might find a particular product whose feature list is very similar to a commercial application you are considering; it has versioning, workflow, a powerful content modeling framework, in-context editing, image manipulation, etc. But these features may have been implemented in a slightly different way — not better or worse, just different; and not because of the licensing model but because two different engineering teams implemented the features. I can't predict which alternative will feel right to the prospective user base. They need to experience the products to make the comparison.
The goal of my selection process is to present products that match my client's stated requirements. To use that car analogy a little more, the cars with the right size engine, the right number of seats, and below a certain price. I also look at industry filters like what is happening with the vendor (I have not recommended Vignette, Interwoven, or RedDot over the last three years because it was pretty obvious the direction they were going). Then I provide some exercises to help the client "test drive" each short listed solution and experience its characteristics and feel. We still do have discussions like "is this CMS worth the extra $100,000?" but only after the client gets a true feel for the differences in the products (and the vendors behind them) and maps these differences to the needs of the organization.