On 5/9/07, Ray Saintonge saintonge@telus.net wrote:
Anthony wrote:
On 5/9/07, William Pietri william@scissor.com wrote:
Anthony wrote:
I've suggested before that a lease of servers would make a lot more sense than all those capital expenditures, and this is a good example of why that's true.
Could you say more about that?
When I've looked at server leasing before I could never make the cost numbers work out, as the server lease arrangements were either a) from high-end companies that charged a premium for their gear, or b) were through private lease-what-you-want companies that charged too much for custom deals. But I haven't done mass gear purchases in a while, so I am probably out of touch with how you whippersnappers do it these days. :-)
You could never make the cost numbers work out compared to what? A lease is pretty much always going to cost more in the long run compared to an outright purchase price, as you're essentially paying to borrow money from someone. But, especially as interest rates are currently low, this type of purchase is perfectly suited to a corporation like Wikimedia which is experiencing such dramatic growth.
Leasing companies do not simply vary their leasing rates with the current market interest rates. Also note that for a tax free company the tax deductibility benefit from lease payments is not there.
I'm not sure what you mean by the first sentence. Interest rates aren't the only factor in a lease, but they are a factor. As for the tax benefits of leasing, right, that isn't a factor in the WMF decision.
Let's say the major fundraisers come twice a year. Let's say a lease can be had for 1/30 the purchase cost per month over 30 months.
Where are you ever going to find a 30-month lease that simply divides the purchase price by 30. The ones that do that tend to have a stiff buy out at the end of the contract. The server will more often than not have a useful life that exceeds the 30 months.
Here comes the answer:
That's a really high estimate, based on Dell and rounding the cost up, and surely the WMF can do better.
I looked it up on Dell, and then rounded the cost up. The buyout option is for fair market value. If you return the equipment at the end of the lease, you owe nothing. I assumed that the WMF would not want to exercise that option anyway. 30 month old hardware is obsolete.
If you think my numbers are way off, please tell me what numbers you think are more reasonable.
Let's assume fundraisers of $40,000, $100,000, $167,000, $300,000, and $500,000 (taken from the financials, rounded, and estimating in order to break up into semi-annual figures). Assume capital expenditures of $35,000, $55,000, $85,000, $150,000, and $275,000 (same methodology).
The effects of fundraising should only have an indirect effect on plans. It's far more useful to have a hardware plan that provides regular replacement of equipment that is no longer useful. If we have 30 servers operational at a given time purchasing one replacement server per month would be cheaper than the monthly lease payments on 30 servers.
Actually, it'd be about equal. But this isn't at all a realistic scenario. Hardware needs are growing, not static, the servers are not going to fail at such an even rate, and hardware is going to probably last longer than 30 months on average. The first two points favor leasing, the last favors buying.
With a lease, you spend $7000, $18000, $35000, $65000, and $120000 each half year. With the extra cash flow you can easily hire a couple of extra staff members plus pay a few consultants for "one-time" things like security audits. From what I've seen of MediaWiki I have little doubt the code contains serious security flaws, and I think we all know that the system has numerous DOS attack points.
The downside of a lease - there's no capital left over at any stage of the game. But considering that Wikipedia's value is currently about 99.9% goodwill anyway, I wouldn't call that much of a problem.
Goodwill remains an intangible asset, and that intangible asset needs to be supported by very tangible hardware.. To do a proper analysis of the situation we need to know the cash price for the piece of equipment, the monthly lease payment, any setup fees for the lease, and the contract buyout amount at the end of the lease.
As I said before, the contract buyout amount is irrelevant, because I'm assuming the option won't be exercised. If it turns out to the WMF's advantage to exercise the option, then that just swings the situation even *more* in favor of leasing. As for the cash price for the equipment, I took that from the financial statements. The monthly lease payments were estimated as I explained above. The setup fees on such a large contract would be negligible, likely less than half a percent.
As for goodwill being an intangible asset which needs to be supported by hardware right now, that's precisely why leasing is such an attractive option for the WMF.
Anthony