A very strange set of events occurred in the late 80's; events that make painfully clear the value that standards can bring to the market.
Between 1992 (introduction of Windows 3.0, and 486 processors) and 1996 about 100 million desktop computers were sold, most based on compatible technology. Oddly, this same capability existed with Macintosh on a 68000 since 1984; and UNIX systems on a variety of processors since 1988. (I pick 88, since that was the year that windowing "standards" became available for UNIX systems.)
What was needed to capture the applications market were Windowing, reasonable address space, multi-tasking, and retail/shrink-wrap distribution. By 1988 these were possible for both Mac and UNIX. Why did the market not explode for desktops until 1992? The answer was simple --- no clear standards were existent prior to that. Clearly Mac was a standard, no ambiguity about that, but it was also not "open" with competing alternatives, and the premium price kept folks going to DOS. UNIX folks got into the "GUI" wars, promoting the concept of dis-unity, and discouraging both customer investment and ISV commitment.
The applications folks played with DOS, and waited ... Eventually Windows was good enough, and the bandwagon started rolling.
100 Million units .... Apple could have gotten 25% at least, maybe 70% if they had opened up their architecture and sold the OS. Figure average selling price of $4k (over time with add-ons), that's 100 million x 4k or 400 billion dollars; Apple could have gotten a bit of the action, maybe $40 billion or so. The same picture can be drawn for the UNIX vendors, probably sharing 20-50% of the market, but with 5-6k per unit (higher end systems) which would leave them fighting over a 500 billion dollar market; most spent the last 7 years downsizing rather than expanding at that rate. (HP, the current market leader (1997) at 15% share or so has total revenues of 18 Billion, much of which is from printers and servers as opposed to desktops. .... the failure to establish these standards probably has a 10 billion dollar per year impact on HP's business.)
You can argue with the numbers, that's easy. But, if you agree that standardization on either Mac or UNIX/GUI platforms in the 1988 time-frame would have shifted some percentage more of the desktops to those platforms, then the only question is what percentage, and what ASP do you associate with this. The rest is arithmetic, and the 100 million multiplier makes even small changes very evident.
Oddly, the consumer's of these 100 Million units also have lost money in the deal; Mac has always been easy to use; and UNIX reliable (as well as implicitly network-able via TCP/IP, which didn't get integrated into the Windows world until 1995 or so) ... either of these approaches would have provided better utilization (why are you waiting to print, or format a disk?), less frustration (just ask General Protection Fault), and a better investment in applications (which would still be running, instead of replacements on the installment plan). I will let someone else figure out the loses to the consumer side, but they might just match or exceed those on the vendor side.
Confirmation of this "order of magnitude" impact comes from an article in ASTM's Standardization News, by HP's Brian Unter (12/96). This article states "IEEE POSIX standards played a critical role in HP's multi-billion dollar HPUX computer business. Thirty percent of sales would have been lost without POSIX compliance. Another 35% of sales where heavily influenced by HP use of POSIX related standards." A rough projection of the UNIX market from 1988 to 1998 based on IDC's numbers gives us about $250 Billion, and HP says that 30% or more of that market is standards driven. This translates into a $75 Billion impact over 10 years. Numbers consistent with the apparent "lost opportunity" from the GUI fiasco.