I’m talking, sans ideology, about “free” as in no-money-up-front. When business is already hurting, up-front software license fees hurt especially hard, and I just don’t believe that Enterprise Software, as currently priced, has much future, in the near term anyhow. [This is part of the Tough Times series.]
Capex and Other Endangered Species · When times are tough, cash is king. “Capital” is cash; accountants pretend otherwise, but real businesspeople know better. And the reality we have to face is that the axe is gonna fall—in fact, it’s already falling—on capital expenditures. There are a lot of middle managers all over the world who, when they see capital-outlay requests come across their desks, won’t even take the issue up the management chain because they know what the answer’s gonna be.
In that environment, imagine what happens when you walk in with a project proposal and the up-front Oracle or Tibco or CA licensing costs are pushing six figures. You’ll be walking right back out again. The drum that Jonathan has been pounding on ever since I got here about “Monetization at the point of value” is just getting louder and louder and it’s not going away.
Technology adoption has to be free. Technology deployment has to be real cheap. When you scale a system up and you’re getting business value for it, that’s when you can actually write a proposal for infrastructure and support that won’t get turfed by the nearest corporate-finance person.
Take-Ways · Two seem obvious: First, use Open Source. Second, if you’re doing a startup, and you were thinking of the Enterprise Software business model, think again. Because nobody’s going to be cutting those big software-license POs.
Contributions
Comment feed for ongoing:
From: Lennon (Oct 13 2008, at 16:04)
So, what's your take on the (decidedly non-trivial) segment of the OSS business community whose revenues depend on the sales of "Enterprise Edition" extensions to their free core product?
Personally, I think they're going to get hit pretty hard, since they lack both the name recognition of an Oracle or Siebel, and the zero-entry-cost allure of a truly free software stack.
[link]
From: John Truong (Oct 13 2008, at 16:35)
I'm voting green as you are in tomorrow's election. It might be worth noting to you that adopting free/libre open source software is actually part of the Green party's platform.
You can read it here: http://www.greenparty.ca/en/policy/visiongreen/partone#_Toc180047589
That's section B7 in case the URL doesn't work. Here's an excerpt, but it's worth reading the whole section on the site:
"Green Party MPs will:
- Ensure that all new software developed for or by government is based on open standards and encourage and support a nationwide transition to FLOSS in all critical government IT systems. This will make Canada’s IT infrastructure more secure and robust, lower administration and licensing costs and develop IT skills.
- Support the transition to FLOSS throughout the educational system."
[link]
From: JulesLt (Oct 13 2008, at 17:08)
I'd love to rip up our legacy code and architecture and start again, but frankly, even if I could guarantee a tenfold increase in productivity, and to be able to deliver the lot on a free open source stack, I'd still be looking at something like 50 man years of rewrite, plus about the same in testing.
With the mass of bank mergers going on right now, there's going to be an awful lot of work in integrating and consolidating systems - there will be some big POs signed to get that work through - but of course there are now less big players and that's going to mean consolidating onto fewer applications.
But I see that work as largely consolidating onto existing systems - right now that is going to be a bigger priority for a merged bank than any long-term IT led projects to port their legacy apps from proprietary Unix / database to open source. That's long-term strategy, whereas consolidation gives better short term gains. 5 figures for an extra Oracle Enterprise licence almost certainly pales to the cost of a project to migrate and test against Postgres, particularly to the type of enterprise that spends 6 months testing a point release.
Which is kind of the same conclusion - it's not a good time to be a startup trying to sell new Enterprise software into a market that is going to be tied up spending money with their surviving existing suppliers.
Now it strikes me that there is an obvious opportunity here for someone to develop some very sector specific open source apps, but you'd also be battling against a mentality that still pays for exclusive IP rights, that still sees that they can establish a business advantage through better software (this probably reflects the relatively poor quality and lack of commodidisation of big software).
However, there is another low-cost / low-risk model that I do see working in the enterprise, which is starting your project as SaaS with some form of shared risk/reward agreement (i.e. Salesforce).
It doesn't matter what the underlying infrastructure is because that's your cost, and if it takes off then you can afford to replace your legacy - because right now no one is going to invest in you doing so.
So I guess, same conclusion - it's not a good time to be a startup in Enterprise IT, but the position of Enterprise IT in the technology cycle suggests that there is plenty of work yet to be done.
[link]
From: robert (Oct 14 2008, at 09:15)
from any Econ 101 First Day:
Now class, always remember the Prime Directive; sunk costs are irrelevant to decision making.
What MBA types always ignore is, while the next fix/mod to 40 year old COBOL is cheap compared to a rebuild in RDBMS and Python, 10 more years of flogging that olde pig costs a lot more. But the MBA types figure they'll be CEO by then, and won't have to foot the costs directly.
Come to think of it, that short-term analysis explains much of what's wrong with Capitalism.
[link]
From: Andrew (Oct 14 2008, at 15:39)
What about the recurring maintenance costs and support contracts that companies like Sun and Red Hat are so fond of selling? Seems to me that the real cost to a business are the recurring costs, which is why hiring people is so expensive as you have that cost forever.
Upfront software license costs are no different from upfront hardware costs, you buy a certain capability at a certain price. Hardware is delivered in the form of atoms and software is delivered in the form of bits but from the perspective of the buyer it's all just capability.
[link]