Do the heroic class systems provide a benefit to their vendors in terms of follow on sales?

This discussion erupted on the beowulf list today. I responded to a question on this, pointing out that prestige adds nothing to the bottom line. What matters is, not so curiously, the bottom line.
One author disagreed with me. His point was that prestige class systems translated into sales for the relevant vendors. I think his examples were stretches, and not applicable to HPC.
All of the public (and private) responses I have seen seem to support my thesis on this.
The argument came about over discounts, and NDAs for deals. Basically because of heavy discounting, the vendors wanted that information appropriately privileged, so that they could control who gets it. If one customer learns that another got a great deal, they are going to demand a similar or better deal. So if you go negative on margin to win a prestigious deal, my thesis was that this will not translate into additional sales, only lower margins.
Let me know what you think here. Go anonymous, and you can send email to joe _at_scalability _dot_ org if you wish, and I’ll post the comment anonymously.
I’ve seen vendors buy deals. This doesn’t last long. It drives vendors who do this more than once, out of business. Customers shouldn’t want to do this, as it destroys continuity, support, and decreases competition in the market. The next time you see a purchasing manager salivating over the prospect of getting an even better “deal”, just remember, there is a cost to every decision … and that cost could be the thing that drives that vendor under. Bad deals kill good companies. Ask LNXI. And many others.
So please do comment here. I am interested in your take on it.

2 thoughts on “Do the heroic class systems provide a benefit to their vendors in terms of follow on sales?”

  1. Reminds me of the merchant who said, “I lose a nickel on each sale, but I make it up in volume.”

  2. I think it depends a lot on who the vendor is. If you’re IBM, part of your operating budget goes to ensuring you’re at the top of the tree, because it’s a market obligation. I think those Top-25 kind of sales aren’t really sales, they’re offsets against significant marketing and R&D costs. IBM can justify that, amortising the “loss” over all the other engineering information they gather as part of the system, like hardware reliability/failure rates, thermal design, etc, etc. The actual loss per engineer headcount is probably pretty low and they’re still making a profit per headcount over their HPC business (insert magic $1MM revenue per employee benchmark here…). There’s only one IBM, one Cray, and a few others in that space who might have that kind of R&D vs. revenue model and who do most of their own hardware.
    For other vendors, I don’t think that works. I’ve worked in software companies where the revenue vs. margin sales race got out of hand, and once word got around customers that there was a certain price available, that was it (NDA? Hahaha, yeah, like that works, go sue a customer, word of that gets around faster than the low prices…). The price never goes up again.
    Unless you’ve got a channel to recover that lost margin somewhere or justify the expense through better technology that reduces your costs, you are risking losing money by doing it. “Look, we sold a big system, that might drive more sales” doesn’t count, that’s not a measurable revenue model. What’s the loss per headcount for the sale? Has it slightly reduced the overall profit margin of the company, or has it just made your positive margin a loss? That’s the real kicker for discounting to get sales.

Comments are closed.