I don't see the distinction. Tacos cost more per pound than bulldozers, and cost more to produce per additional unit than enterprise software (assuming off-the-shelf and not custom-built).
A taco has just as large a long-term commitment as any durable good. The taco stand has to deal with more, in fact: if the shell is cracked or they forgot the meat, they have to deal with after-the-sale support. If I come down with e.coli and am hospitalized for two weeks they have an additional degree of commitment and exposure to deal with.
The problem I'm describing isn't with custom-built, built-to-spec goods (like some software), but rather commodity durable goods and materials. Whether I design my widget to be made out of stainless or titanium often depends on pricing, and determining that pricing carries unnecessary overhead which serves neither customer nor manufacturer. The opportunity cost of opening the dialog with the titanium company's salesperson becomes nontrivial, because now I have to wait a few days for them to call me back to tell me it costs more than my alternative. Salespeople should be taking down orders, driving traffic, cajoling hot leads, and negotiating discounts, not performing the work of a javascript in a significantly less capable manner.
Until someone does a double-blind study with identical products, my guarantee is based on anecdotal evidence (and as Bender once said, "I can guarantee you anything you want"). I think it would be fun to experiment with an app store listing with "call for price" as its cost. Would it really be so hard to predict a difference in sales?