|
||
|
|
|
              |
|
 |
 |
||||
 |
|
|||
By William Van Winkle |
||||
LIFE IS ABOUT BALANCE, YES? Any good action movie needs occasional “quiet” scenes so the viewer can catch his breath; too much action for too long gets boring. I’ve freelanced for tiny local rags as well as some of the biggest magazine names in computing, and I’m telling you that it’s easiest and most profitable to make a living in the middle where the work load and pay achieve equilibrium. In a successful marriage, partners take turns being “right” or sharing the decision-making power. Maybe my viewpoint is tinged by the climate of my native Oregon, where we break through 100 or dip under 20 degrees only once or twice a year. Success is about balance, about finding your groove in the middle ground. One of the things that I think has contributed to Intel’s success, even in down times, is the company’s ability to target the middle ground and strike a balance between the high and low ends of the market. Look at the company’s desktop boards. You’ve usually got a couple of Extreme models (currently the DX38BT and D975XBX2) and a smattering of Essential SKUs (currently based on the aged 945G/P and the SiS662 chipsets). In the middle, we have the bulk of Intel’s board business, nearly all of which is based on the new 3 Series and last year’s 965 chipsets. Perhaps it’s telling that the board I hear about most from Intel isn’t its $270 Extreme Series superstar or the $69 SiS board with a soldered-on Celeron 220 but its $112 DG33BU Viiv board for consumers and $135 DQ35JO workhorse for business desktops...right smack in the middle. The odd thing is that you’d think more companies would take the clue. Hey, targeting balance seems to have paid off for Intel, after all. But this isn’t the case. Instead, we have massive anticipation for and media coverage of the latest and greatest GPU. The press frenzy for NVIDIA’s GeForce 8800 and AMD’s HD 2900 parts when they arrived was huge. No doubt, much of the attention was deserved. This GPU generation introduced a ton of new functionality and significant architectural overhauls. Unified shaders? That’s a big deal. DirectX 10? Also a big deal. But what about the ability to smash a prior benchmark record by 50%? Not such a big deal. What does that mean, after all? So this new thing was 50% faster than the last new thing or the competitor’s newest old thing in some new or recently new benchmark app. What does this miraculous feat have to do with reality? Nothing. When the benchmark app comes out with a patch update, the results change. When one GPU performs 4% faster than another GPU taken from the exact same bin at the exact same time from the exact same yield batch, the numbers change. When the air conditioning kicks on by the test bench, the numbers change. So what exactly do these numbers mean? Will anyone care about this GPU’s benchmark scores in three months when the price drops and new SKUs with a faster memory type or more stream processors or an incremental fab shrink take their place? I never have. And yet we’ve heard over and over about the “halo effect” or (for all you ‘80s kids) “trickle-down” inherent in owning that top benchmark slot. The 5% of consumers who will actually shell out a car payment or two for a graphics card are the “influencers” who dictate the rest of the market. They’re the “experts” who advise their friends on what brand to buy, and these people in turn go out and spend their money in the mainstream and value segments. I’ve listened to industry analysts and execs for years rehash this idea. I even had one tell me about it again today—the High-End Influencer Effect. And that got me wondering... QUESTIONING REALITY How do we know it’s true? Do we have measurements that support the High-End Influencer Effect in the market? Is there some shadowy metric from Google or comScore or whomever that can measure the number of semi-literate exclamation points and their proximity to model names in hardware forum posts and thereby ascertain the impact these Influencers are having on the world? I’ve never seen such measurements. When I went to buy my Honda Civic in 1993, I didn’t care that Honda made the supposedly ritzier, better benchmarked Acura brand. I couldn’t afford an Acura, and I’d never even been inside of one. I knew people who owned BMWs and Audis and raved about them—until they faced the repair bills. In 14 years and 180,000 miles, I’ve had exactly one major repair on the Civic, and it was strictly from usual wear and tear. I still drive it every day because my criteria were safety, fuel efficiency, and dependability. Not once did I seek brand-buying advice based on someone’s experience with an ultra-high-end vehicle. In fact, I don’t even know anyone with a car that costs more than $50,000. So much for the High-End Influencer Effect there. Let’s bring it closer to home: hard drives. The top-performing hard drive on the market for quite some time has been Western Digital’s Raptor. I’ve tested it myself. I know first-hand what an awesome drive it is. I am qualified to be a High-End Influencer. But would I ever recommend one to a mainstream user? Of course not! Does your mom need a Raptor drive? Or your boss? Or your neighbor? My three best friends are all gamers, and I wouldn’t recommend a Raptor to any of them because none of them is enough of a split-second, benchmarking control freak to notice the tiny performance improvement in their everyday usage. I’d recommend something like a Seagate Barracuda, which delivers a far better mix of high capacity, satisfying speed, and durability for a much lower price point per gig. Here’s another example: flat-panel displays. If you read this month’s cover story, you know that I bought a 50” Fujitsu PlasmaVision a few years ago because it was supposed to be the best. Shortly after the purchase, I ended up interviewing an executive geek at Samsung who was familiar with my Fujitsu model. “That unit looked the best in the showroom,” he told me, “because the shop set it on a wall by itself and calibrated the display. All the other plasmas were shown next to other on different walls, right?” They had been. “And most of them probably weren’t precisely calibrated. Both tactics make the Fujitsu look better in comparison and justify its price tag, which of course makes the store more money. But I’m telling you that if you would’ve seen my plasma next to that Fujitsu and had them both properly calibrated, you would not have been able to tell the difference between them.” Had I known this a month before, I would have saved $3,000. So when people asked me which plasma I would recommend, what do you think I said? The Samsung, which offered nearly all of the high-end performance regular people would want in a mid-level price band. Mind you, I’m not advocating cheap. Like everybody else, I’ve been burned too many times to count by cheap products—drive motors that melt, cars that break every month, TVs with terrible images. I’m saying sensible balance is where we want to be. So for now, I don’t believe in the High-End Influencer Effect. Maybe it applies in 3D graphics, maybe not. But I’m not willing to say it extends far in the world of computing beyond GPUs. Some people say that Athlon chips sold better because of the halo effect from Opteron’s success. I would counter that Athlon took market share from the Pentium 4 because of more aggressive pricing and better performance. Likewise, I wouldn’t make any correlation that the success of Core 2 Duo and the Xeon 5000 had anything to do with each other. I’ve seen no evidence to indicate that server admins were going home and saying, “Oh, man, this new Xeon chip rocks! You guys need to go buy Core 2s!” No, the Core 2 Duo turned the tables on Athlon, period. The Core 2 Duo and now the Athlon 64 X2 after a refresh or two are both excellent chips with a lot of value and functionality to offer mainstream buyers. No trickle-down required. And I’ll go one step further. I’ll say that there can be times when there is an inverse relationship between the benchmark leader and the graphics card that regular users should buy. When the GeForce 8800 parts arrived, they swept the field for benchmark scoring. When AMD’s Radeon HD 2900 arrived, it still couldn’t best NVIDIA. So would I as a knowledgeable High-End Influencer tell my buddy who wants to build a little living room media center box to go buy an NVIDIA card because “NVIDIA is the best, man!”? No, I’d actually tell him to go buy a low-profile, passively cooled Radeon HD 2400 card, because dollar per feature for his specific needs, that’s the most sensible choice on the market today. And it just so happens to be in the product segment’s middle range. (OK, maybe a bit below the middle.) DON’T IGNORE THE CENTER You didn’t see AMD spend a lot of time talking about its mid-range products in the last quarter. Neither did NVIDIA. Seagate and Hitachi were preoccupied with talking about hitting 1 terabyte while value-adds such as full-disk encryption on mid-range business drives barely made a dent in the headlines. When’s the last time you heard about a mainstream cordless desktop? Or an affordable DVD drive that had the key selling point of best compromise between high speed and low noise? My point is that, yes, most people buy mainstream products. That’s where most of the money resides in our business. But almost everyone, vendors most importantly of all, ignore this middle ground. This is where you make all of your upsell profit, for cryin’ out loud! There’s no margin at the low end and no volume in the high end. The OEM-matching budget price gets people in the door, but it’s the value-adds you get by picking special mid-level parts that target the user’s specific needs that can make your system more special and desirable than one from anybody else. I’m saying you should give some attention to the high-end news just to keep abreast of where the market is going, but focus most of your component efforts and in-depth understanding at the middle. This is where you’ll find and can pass on value-rich, high-end features for only a few bucks more. Mid-range is often perceived as being “ordinary” or “average.” If your competitors want to think that, fine. But don’t make the mistake of following suit. Master the middle ground and rise above the rest of the pack. |
||||
 Back to top |
||||
Copyright © 2007 RAM Magazine. All rights reserved.
Do not duplicate or redistribute in any form. |
||||
