First, the obligatory boasting

It appears that you don't need to know much about software to understand the software business. Two quotes from various articles of mine under the pseudonym "streetlawyer" on OSDN member website kuro5hin:

Not only that, but arguably, the company itself is shying away from the ideas by which Raymond thought he would revolutionise the economy. The latest hope for VA's profitability is to get away from hardware, get away from consulting and to sell copies of the intranet version of Sourceforge. -- from "Eric Raymond Wealth Clock is Down"

This gives a burn rate of $31.1mn per quarter. Which gives VA three quarters to reach profitability, or it will have to cut core functions. Larry Augustin thinks they can make it by October 2002. That's five quarters, by my countin'. Something's gotta give. -- from "VA Linux 10Q Analyzed".

I'm not claiming any special insight into anything -- personally, I thought that the move away from hardware was pretty easy to read between the lines of the VA disclosures. But it's always good for the ego to have a prediction success.

Second, running the numbers

Well, we said that something had to give, and something gave. The publication of the full cashflow statement put it on the line (and incidentally showed that our burn rate estimate couldn't have been far off -- cashflow from operations was negative $62m over nine months). The problem at VA was an excess of costs over revenues, not helped at all by a $7m buildup of inventories and the only thing they could hope to do was to shrink, and fast, to husband the existing cash resources. Meanwhile, it didn't take a Hahvud MBA to realise that if you have to choose between shifting boxes in direct competition with Compaq and IBM, or providing services in a small but obscure software niche, you go with the niche unless you have cash to burn.

So, Larry Augustin, the eleven and a half million dollar man (that's how much cash he's realised by selling VA Linux stock since the IPO) cut the boxes, with 35% of the remaining staff to follow. Greater love hath no man ...

According to the statement, this will reduce the cash burn to $8mn/quarter. Does it add up?

It's hard to say. No, make that, it's impossible to say, because we don't have any separate disclosure of costs and revenues in the hardware business. I'll be charitable at this point and say that if the VA management had known they were going to do this a couple of weeks ago, they would probably not have decided that the best divisional split for their revenues was "OSDN" versus "Systems and Services". In the opinion of this analyst, if you can separate out a business line to close it down, you really ought to separate it out in your accounts. But hindsight is a wonderful thing.

So, a certain amount of triangulation is required. To start off our analysis, here's a synopsis of VA's cashflow statement, with a few explanatory notes (for nine months ended April 28 2001, US$m)

Revenues (FROM INCOME STATEMENT)                                      118
Costs (FROM INCOME STATEMENT)                                        (356)
(estimated)

Net loss -                                                           (235)
Add depreciation back in (SEE BELOW)                                   74
Add back in non-cash elements of loss 
(mainly stock compensation)                                           101

Add in cash from collection of a/c receivable                          17
Subtract cash spent on building inventories                            (7)
Subtract cash spent on prepayment of bills
(usage of money set aside in previous periods)                         (7)
Subtract increase in accounts payable
(revenues accounted for but not yet collected)                        (10)
Correct for other non-cash accounting items                             5

OPERATING CASH BURN                                                   (62)



In order to get from the $62m cash burn in the Q3 accounts to the projected ongoing $24m (ie: three quarters @ $8m/quarter), we need to find $38m. We can get $7m of that by undoing the inventory increase (we're assuming that the inventory can be sold at cost, which may be optimistic), but it's clear that the bulk of the work is going to have to be done by reducing the size of the net loss.

In order to reduce the size of the loss, VA is just going to reduce the size of the business; to work out by how much it intends to shrink, we just work backward from the bottom line; see how much of a loss they need to get rid of, then divide by the net margin on that kind of business.

The effect of the April round of layoffs should come through in the next quarter (Q3 was particularly bloody for LNUX, as the revenues had fallen but the costs hadn't had time to react). That ought to be worth $10m right off the bat, given that the charge made for severance pay of the San Diego bunch was around $3.4m, and they'd likely have been on three month terms. And other one-off items can probably account for another $5m. So the remaining hole is $15m; this is the loss attributable to hardware-related business. If we assume that the gross margin of 14.0% which VA earned at the time of its first SEC filing in 2000 (before Sourceforge etc were a significant part of the business), that would imply that the revenues of LNUX will shrink by $100m when it gets rid of hardware. So when Augustin says in the statement linked above that "VA expects its revenue to significantly decline with the elimination of the hardware segment", he ain't kidding -- we could be seeing an 80% decrease in the turnover of this company. Admittedly the numbers above are heavily dependent on assumptions (particularly, the margin assumption is worryingly arbitrary), but it could be that those fine upstanding citizens, the American Small Business Association, just welcomed another member to their ranks.

Thirdly, assessing the future

I've got a few ideas on a number of subjects which the press release touched upon. Breifly:

• I don't agree with Steve Gilliard that OSDN is a cash cow for VA. I tend to be of the opinion that while Slashdot was, verifiably, cash generative at the time of the Andover.net float, a lot has changed since then (the advertiser base of Slashdot was no less heavily weighted toward dot coms than the buyers of VA Linux servers, remember, and although it's true that a banner ad doesn't cost as much as a new router, it's also true that bankrupt companies don't advertise). The Register appears to agree on this one.
• As mentioned in the first article linked above, I can't help seeing the move toward SourceForge Onsite as a move towards selling software, for money. I don't understand how that fits into the Open Source business model -- in fact, I don't think it does.
• I'd be interested to know what is to become of the Open Source celebrities on the payroll at VA (the "UberGeeks". When do they jump ship (if at all), and what happens to the new software-driven business model when they do?
• And of course, the spectre of a takeover has to be on the horizon. At $3.6, VA is trading at half its stated book value -- 50% of the capitalisation of the company can be accounted for by cash in the bank. The period immediately after stabilisation of the cash burn rate (once the risk of a "black hole" has been removed) is often the period when troubled companies become attractive to predators.

But this article is already too long, and I'd like to get some opinions before proceeding. So for now, so long, and don't take any wooden nickels out there!