September 17, 1998
ON THE HILL
An Investment Opinion
3Dfx and Market Efficiency
Shareowners of 3Dfx Interactive (Nasdaq: TDFX), designer of the popular Voodoo2 3D and new Banshee 2D/3D computer graphics chips, experienced another jolt this week after the former highflyer pre-announced a third quarter operating loss due to an unexpectedly severe seasonal slowdown. According to CEO Greg Ballard, "There appears to be more inventory on hand at the retailers for Voodoo2 than originally forecast by our leading add-in board distributors." The stock plunged as much as $2 13/16 to a low of $8 Tuesday before recovering a bit. Investors apparently looked at the firm's book value of $7 a share ($5.55 in cash) as of June 30, considered its well-regarded technology, and concluded that 3Dfx offers value even in a competitive niche known for its short product cycles.
Although consensus third quarter analyst estimates before the warning still called for the company to earn $0.47 per share, much of the bad news was already in the stock. Since hitting an all-time high of $35 in April following a blowout first quarter, 3Dfx had been drifting lower and lower. The decline picked up steam after the July 15 news that second quarter earnings beat estimates by just 6%. Inscrutable as it may seem, Mr. Market has been projecting second half problems for 3Dfx. The stock price itself was providing a "sell" recommendation even if the analysts weren't.
For Foolish fundamentalists, listening to a stock's gyrations for advice seems both foreign and inane. Yet 3Dfx suggests there are times when the market's apparent irrationality is surprisingly rational, so that a chart accurately points to something fundamental.
How can this be? There's a story of a mutual fund manager who asked one of his assistants to call up the top analysts following a company to find out why the stock was moving. The assistant said, "But I'm the best analyst covering it. No one knows more about this company than I do." The fund manager said, "That may be true, but every analyst covering this company knows at least something that you don't know. So collectively, they know a whole lot more than you do. And what they know is what I need to know."
The market speaks in the collective voice of all such market participants. That doesn't mean you should take up technical analysis and start trading based on candlestick charts. It does mean that when there's no apparent reason for a stock's move, it's best to assume that there's meaningful news even if you can't discern it right away. The stock market is not perfectly efficient. Yet in any specific case, you're better off assuming that it is until you can pinpoint why it's not.
Another way to approach the problem, and one that's especially appropriate for a fast-growing, high-risk smallcap such as 3Dfx, is to follow speculator George Soros's method. Soros always wanted to know the flaw in an investment thesis so that when he "played" the thesis, he would be prepared for the telltale signs of trouble and get a jump on others following the trend. In more fundamental terms, this would involve determining the most likely things that could go wrong with an investment and then watching for the signs.
In the case of 3Dfx, I can do little more than guess what those signs were. Certainly Intel's entrance into the 2D/3D chip market was one background concern suggesting just how competitive this niche is. The general inventory glut and falling prices throughout the PC and components markets was no doubt another issue. A spate of insider sales amidst the stock's decline didn't help. Yet it's also clear that investors and graphics board manufacturers had high hopes for Voodoo2. One general problem with new product launches, though, is that sales can be inflated early on by the pure necessity of filling the retail channel. Sales into the channel aren't the same as sell-through. A miscalculation about the strength of end-user demand can rapidly create an inventory glut.
A reasonable place to look for news about 3Dfx's business, then, would have been Diamond Multimedia Systems (Nasdaq: DIMD), a firm that uses the company's Voodoo2 chip in its Monster II graphics card. Diamond has been 3Dfx's biggest customer, accounting for 37% of FY97 sales, 41% ($20.5 million) of 3Dfx's first quarter sales, and 42% ($24.6 million) of second quarter sales. Though Diamond's overall revenues include many lines that don't employ 3Dfx products, graphics and multimedia accelerator subsystems make up most of Diamond's revenues, including 82% of its $172 million second quarter sales. Most of Diamond's 117% sales increase for the first six months of 1998 was due to the Voodoo2-based Monster 3D II.
Also worth noting is that Diamond has repeatedly disappointed investors since going public in April 1995 at $17 a share. Like other firms in this competitive industry, it has a history of encountering excess channel inventory and then taking a sales and margins hit to sell it off. This should have been troubling because Diamond revealed in its Q2 press release on July 9 that Monster II sales were quite strong but "fell short of our expectations toward the end of the quarter by approximately $20 million." Given that Diamond aims to maintain channel inventory amounting to "one to three months of customer demand," it's possible Diamond had more than three months worth of Voodoo2-based Monster II boards in the channel at the end of June. That's a significant amount considering the short product life cycles in this business.
Why keep that much channel inventory? For one thing, Diamond has a less than optimal handle on its operations. The company is in the process of implementing a new enterprise resource planning (ERP) software system, but it won't be operational until the end of the year. A more basic problem is that both 3Dfx and Diamond are a long way from the retail customer. 3Dfx sells to Diamond (and others), which then sells to mass merchandisers, retail distributors, and PC manufacturers. As Diamond mentions among its risk factors, "the Company typically has little or no direct visibility into end user customer demand." In short, news of slower retail sales might take a while to get back to 3Dfx.
That meant that Diamond's inventories were worth watching (even though they reflect non-3Dfx products too). Investors might have been concerned by any unusually large increase in finished goods since it would suggest a serious oversupply. Diamond's revenues usually follow a seasonal pattern (strong 4Q, weaker 1Q, still weaker 2Q and 3Q), but product cycles exert a powerful countervailing force. Since Voodoo2 launched in February, these sequential comparisons are useful (dollars in millions).
Inventory 12/97 3/98 Change 6/98 Change
Raw Material $29.9 $35.6 19% $32.4 -8.8%
Work in Progress $40.3 $45.5 13% $41.3 -9.1%
Finished Goods $8.5 $11.4 35% $20.1 75.6%
Revenues $185.9 $186.2 0.2% $172.3 -7.4%
Receivables $98.8 $102.9 4.1% $102.1 -0.7%
You often see an inventory build-up around the time of a new product introduction, so the high percentage of finished goods at the end of the first quarter wasn't disturbing. On the other hand, the build-up at the end of the second quarter was. Although Diamond's second quarter 10-Q wasn't filed until August 14, the inventory breakdown was probably available July 9. An added reason for worry was that accounts receivable were declining more slowly than sales. Add all that to Diamond's disclosure of disappointing Monster II sales and the fact that 3Dfx barely beat second quarter estimates, and it makes more sense that the market was factoring in a weak third quarter even before the pre-announcement.
For more on 3Dfx -- CEO Greg Ballard comments on the outlook for the company's Q3, Q4, and its new Voodoo Banshee product in a FoolWire article.
FOOL ON THE HILL Archives
|Are you a Foolish investor?|
|The Motley Fool Recommends...|
|The Weekly Fool
The best thing since sliced cyberspace -- it's the Weekly Fool! Get more info or order.
New format! A stock idea, industry overview, top players, and financials -- every two weeks! Get more info or order.
You give it the data, and it does the rest: PEG & Foolish 8 valuations, ratio valuations, and lots more. Get more info or order.