someone (someone@somewhere.com) on 2/7/09 wrote:
---------------------------

>>
>
>Using units is even worse because it presumes not
>only static prices but also prices are flat across units.
>

Units presumes nothing but units, it seems to me. Unless you are trying to determine the PE ratio of the company stock, which has no direct relationship to how well a specific product is doing in the market (presuming the company has many products). Unless I am mistaken, the success of the *product* is what is being discussed here, not the success of the *company*.

For example, let us consider a possible scenario...

IBM mainframes all sell for over $1M in 2000. They use proprietary hardware (CPU, memory, etc.). However, by 2009 IBM mainframes have incorporated numerous 'commodity' parts - NICs, PCI-E cards, Fiber Channel, etc. Now, many of these systems sell for $100K, and they cost substantially less to make than they did in 2000.

What will ASP tell you here? It won't tell you how popular or successful the product is. I don't get why ASP has any more meaning than total revenues or even units, unless something else remains static (cost, units, etc.)

It seems to me like using a single metric is like trying to draw a straight line from a knowing a single point - you might be able to draw the line, but what value is it?

At least with units you can say "They sold 1000 last year, and they sold 1000 this year. Therefore, they didn't grow their market". Sure, they might have sold *bigger* systems, but it would likely mean they sold them to the same customers. If the price of the systems doubled, it would look like they had 100% market growth, and if other systems only went up 20%, it would look like they were 'catching up' to the other systems in sales.

This is the problem I have with the mainframe market - they tell us how much they sold in dollars, but we don't know if there were new installations, or if they sold to the same customers. If the revenue is up, does that mean more customers or does it mean higher priced systems? If revenue is down, does that mean fewer customers, or lower priced machines? Since we *know* that there are now $100K IBM mainframes, where there used to be only $1M and up (with the average probably being several million), and IBM has been trying to reduce cost (and price) for a number of years - how do you compare that to a product where the system costs remain relatively the same (or slightly higher) year over year?

It seems like much of the discussion of what product/company is 'winning' and which is losing is based on a lot of nothing. It's been predicted that mainframes, and indeed, IBM, will be 'dead' for over 10 years. It's been predicted that Macs, and Apple, will be dead. That SPARC, and Sun, will be dead. That Itanium will be dead. That AMD will be dead. While eventually somebody will be correct - it's just a lot of people arguing to this point because they all see only one piece of the elephant (sometimes by choice!).

Regards,
Dean