Monday, July 23, 2007

Console Post of the Week: Microsoft

I've been thinking about the scenario of a full Xbox 360 recall over the last week, and I've come up with one more potential scenario.

A voluntary recall.

Don't want to send in your 360 for replacement? Fine--you still get the three-year warranty, and if it fails, it can be replaced then. If you do want to get it replaced with one of the "newly unshitty" units, though, you contact Microsoft, they check the serial number, and if your unit qualifies, it gets replaced.

The best part of this scheme, for Microsoft, is that they don't have to contact customers--we get the choice, but we also have the responsibility of contacting them.

Plus, if they announced this at, say, the end of January, it wouldn't cut into their holiday sales, and nobody's going to buy a console in February, anyway. If they get flooded with replacement requests in February, it's no big deal.

That seems workable compared to any other scenario I can envision.

The one downside in a profitability sense is if they just wait for consoles to fail, the BOM is steadily moving downward on the unit, so replacing one in June is cheaper than replacing one in February, for example.

As class action suits rain down on Microsoft, I think some kind of recall becomes a more likely possibility. And a voluntary recall would blunt critics quite effectively.

If you want to know some of the potential numbers involved, take a look at this outstanding analysis by DQ reader Skip Key, based on information revealed in the Microsoft conference call:

We know that $1.057B of that operating loss for the quarter was the one-time charge due to the warranty. So without that, the operating loss would have been $142m, which would put them well on the way to profitability, I'd think.

We also got further information on the $1.057b charge. It was about a 50/50 split on charges based on the existing warranty, and forward-looking charges, which they'd already announced. But the interesting thing was that 35% of that number is for 'existing inventory writeoffs'. And one of the analysts asked about that in the q&a section of the webcast. The inventory writeoffs are for returned 360s that can't be refurbed profitably. So that gives us approximately $370m in returned 360s that are essentially going to be junked.

So this gives us a pretty hard baseline for RROD (red ring of death) returns. Let's say that a core 360 without drive, controllers, etc. costs $250-$300. I bet it's less than that, but that's a good round number. So basically you'd be writing off any unit that cost more to refurb than that. At $300/per, that covers 1.23m xboxen. At $250/per, that covers 1.48m 360's. So using these figures somewhere between 10.6% and 12.7% of all shipped 360s were returned under warranty with this problem, junked so badly they couldn't be refurbed. What percentage of all the returns for this is that? Half? If it's half, we're right in the 21-26% return rate that I figured they had, and within spitting distance of the anecdotal 30%.

The main other interesting thing was the forward looking statements. For every other division they were projecting revenue increases for FY08 that varied by 1-2%. IE, they'd project a 12-13% revenue increase. But for the Entertainment division they projected a range of 10-19%, which is huge.

This got brought up in the analyst Q&A. And it was confirmed that the wide range on the guidance was because they didn't want to tip their hand on 360 pricing strategy. So lets take that 9% range and shrink it to 7%. As I figure that, they basically see that 7% range as the difference between different price cuts. 7% of revenues on this year is ~$426m. At this point, I doubt that they do a price cut before Thanksgiving, but I'd be shocked if they don't do one then. If the cut is in November, that leaves 8 months in the fiscal year. My guess is that this $426m range represents a $100 range on possible price cuts, which would imply that they expect to sell about 4.2m units from November through June. So what level of price cuts would give them 500k/month sales? I don't think $50 would do it. $100 might. So this probably represents that they plan to cut between $50-150 or between $100-200 on the price, probably after Halo 3, but before Christmas.

Let's take a closer look. Key points:

--$370 million of the $1.057b charge is for returned units that are just going to be scrapped instead of refurbed.

--I think the $250 BOM estimate is probably closer to correct (at this point) than $300, so the 1.48M (let's just round up to 1.5 million) number is probably the best estimate. So Microsoft is scrapping, potentially, a million and a half units.
--Microsoft gave a wide range on guidance for the Entertainment division so that future action on prices couldn't be divined by the guidance. So they are planning a pricing action, but the range makes it impossible to nail down with any certainly.
--Skip mentioned that a $100 price cut might get the 360 to 500k units a month in sales. I think it might double holiday sales, but expecting a permanent doubling of the weekly sales rate is less likely.

Here's the most interesting part of the analysis, and it's quite fun to speculate here. If the BOM for the 360 now costs $300 (which I think is fairly close), then it means that (Skip's calculation) about 1.25 million units will be scrapped, based on the $375M that's set up for "existing inventory writeoffs."

If the installed base is around 10 million units at this point, that projects to a staggering failure rate. Remember, they're not scrapping all returns, just the ones that can't be refurbished profitably (red ring of death units). Skip notes that even with a very high writeoff rate of 50% of returns, for example, it would mean that the overall failure rate (on a customer basis, not a unit basis) would be at least 25%.

Incredible. That's so far past "nightmare" that it doesn't even slow down as it passes nightmare.

Tomorrow, Nintendo.

Site Meter