Bouncing Zabaglione Bros.
Legend
I suppose you could say that the halo effect revolves around reputation. Like all reputation, it has it's own inertia.
The "halo effect" - is anyone really that dumb?
- Thread starter nicolasb
- Start date
I suppose you could say that the halo effect revolves around reputation. Like all reputation, it has it's own inertia.
Freak'n Big Panda
Regular
yeah exactly
You're right; I was thinking brand recognition by this point but this is not a category where the halo effect comes into play.
However, I do believe in the halo effect. People are definitely swayed by rumors from 4-6 months (or more) ago, and I can't explain it any other way. I believe the key is it has to be a field where a brand utterly dominated in the past. I.e., it really used to be that you couldn't buy anything but an Intel chip for an x86. And, when AMD first launched their budget x86 cpus, they sucked, comparatively. They were a nice cheap fallback but if you had the money there was no real reason to go AMD. After some time, they became price-competitive at many points but it wasn't until the a64 architecture was definitively superior that, well, my theory is that when doofs what write for print magazines and appear on TV news and such -- when those people finally acknowledge the up-and-comer due to obvious superiority, well, then people listen. Speaking of "the herd", here.
5-8 years ago, this was the case with graphics cards, and, honestly, a surprising number of people haven't forgotten. I believe a lot of people still think that because one company in the GPU market has a superior achitecture, that means that all of their products for that generation are going to be superior, as if ATi lacks the technology to build a competitive midrange part right now. This has occasionally been the case in recent years, but rarely so. My theory is that people are terrified of being caught in another 3dfx situation, where, if you were a relatively late-adopter, you got stuck in the cold when everyone sorta jumped ship and went HW T&L. Now I realize there are a lot of fallacies with this reasoning, but I think if you only pay attention to the graphics industry once every three years, and then only superficially, this is what happens to you.
To once again pick on my favorite print mags, do they even review budget cards? I've only ever seen highend reviews, but I'm sure some must from time to time. However, the big gaming mags often just review flagship cards (poorly), and I can bet you someone pays them well to do that. Lol, that may be untrue, but I think it's pretty certain that halo effect works in gfx cards. I myself usually go to the reviews and web stores every six months, intending to see what deals there are in my price range. When I do, I usually am surprised to find what I do -- which is to say that, while I'm open to buying either manufacturer's cards most of the time, I often find that new midrange parts are changing my paradigm. A good example of this is the nV 8600GT, which I assumed would be a far better part than it is. And it happens the other way all the time -- it's not really a very easy field to keep track of -- no, it's incredibly ridiculous, and probably 100% intentionally. The product numbering schemes are probably making both companies more money than any amount of marketing right now.
Last edited by a moderator:
Price?
If you look in the RV610/630 thread right now you will see that there are people even here who use the price point of the RV630 as an indication of where its performance will fall relative to its competition. And as long as price/performance is the single figure of merit that is used to communicate with consumers, that is actually a pretty reasonable assumption. This means that the consumer does not need to care about the brands involved, just decide on a price bracket, and the assumption that performance follows price will take care of the rest.
This is the core of the issue.
I'd say that the 7800GTX didn't enter into it - the other SKUs in nVidias line-up sold on their own merit.
Again, nobody is arguing that branding doesn't matter - it is the specific effect of having the top performing card on the sales of the rest of the line that is the issue.
As was pointed out above though, it's not necessarily the same person/people being intelligent/stupid. "Dude, buy a GeForce".
For a lot of folks market research involves asking their neighbours son what sort of graphics card he's got, because he's a l33t gamer who "knows a lot about them computers". He'll have the high-end NVIDIA card of course (for example), and chances are he's either one of a) a total NVIDIA fan-boy who wouldn't recommend ATI under any circumstances, or b) not particularly knowledgeable about the state of the mid-range and low-end market, because he's a l33t gamer. In either case his answer is likely to be "dude, buy a GeForce".
We can argue round the houses how likely such a scenario is going to be, but I'd propose that it's much more common to ask someone who you think knows than it is to do research on the Internet, and that when asked such people are more likely to recommend the brand they already own than they are the competitor.
For a lot of folks market research involves asking their neighbours son what sort of graphics card he's got, because he's a l33t gamer who "knows a lot about them computers". He'll have the high-end NVIDIA card of course (for example), and chances are he's either one of a) a total NVIDIA fan-boy who wouldn't recommend ATI under any circumstances, or b) not particularly knowledgeable about the state of the mid-range and low-end market, because he's a l33t gamer. In either case his answer is likely to be "dude, buy a GeForce".
We can argue round the houses how likely such a scenario is going to be, but I'd propose that it's much more common to ask someone who you think knows than it is to do research on the Internet, and that when asked such people are more likely to recommend the brand they already own than they are the competitor.
I don't see that this makes any difference. All that means is that the person who is being consulted has to simultaneously be educated enough to know about the high-end situation and dumb enough not to realise that the same doesn't apply to the mid-range.
Bouncing Zabaglione Bros.
Legend
You're looking at this backwards. You're saying it's unlikely that somone will be both dumb and clever at the same time, but the fact that the halo effect works proves that this is exactly the case.
People buy all sort of things on bizarre reasoning. Cars that look nice or have a "name". Designer clothes that don't flatter you, but have the "name". Houses that are dumps but have the "name" of being in the right area.
Companies invest vast amounts of money not just on promoting their products, but promoting their company and brand names. Why do you think that there are all these companies out there that you find are actually owned by a few conglomerates, so it seems that these many companies are actually competing against each other in the marketplace? It's because branding and company perception is important.
People buy Gillette razors, Bic biros, Apple Ipods, Dell PCs etc, and likewise they buy ATI or Nvidia because they know these names are big in the graphics field, but they don't know much more than that, but they think this is adequate to make a decision because that's what's on all the shelves and that's what is in every shop. They make what they think is an informed decision, even if it's not, and then simply backfill that they made the "right choice" because that's what they bought, and they wouldn't have bought the wrong thing would they?
The halo effect works because most people are like sheep. Most of them probably believe that Nvidia makes the best graphics cards in any segment because that's what the cover of the box tells them, and that alone is sufficient for them. You make the best technology at the top level, and most people assume you make the best technology at every level, because all they hear is you claiming to be "the best" all the time.
Last edited by a moderator:
You are acting as if "the halo effect" is some kind of natural law. It is absolutely not. Even people who are professionals in the field of marketing are very careful about when and where it may apply and to what degree. We are talking about a specific case here.
If this was the way it worked, we would see market dominance lead to more recommendations leading to even greater market dominance, leading to more... in short, a snowball effect. (Which is a ways removed from the alleged halo effect of top-end cards btw.) In the field of graphics however that is not what we have observed. 3dfx was overtaken by nVidia whose dominance was broken by ATI who in turn lost the market lead to nVidia again. If people recommending what they already had was a dominant effect, this wouldn't happen, so while your thesis sounds pretty reasonable (like the high-end halo effect) it just isn't significantly borne out by the overall market behavior. This in spite of recommending-the-brand-you've-got-and-are-happy seeming like a much more plausible market force than high-end halos.nutball said:
For the record, I find this a sign of health from a consumer perspective.
Last edited by a moderator:
Maybe they realise, but just don't care that the mid-range situation is different because it doesn't impact on them. Given the sheer variety of products in the mid-range with all those GTs and GTSs and GTXs flying around, it's a recipe for a bad headache. If you own an 8800GTX why would you waste your time reading up on the 8400? You know enough to know that the 8800GTX is better than ATIs offering, and that you're well aware that the situation is different in the mid-range, but can you really be bothered to do someone else's market research for them? I can't.
"What graphics card should I buy for £75?"
"I dunno, I don't know much about that bit of the market, do your own research"
"What sort of card have you got?"
"I've got a £300 NVIDIA card, but you might find ATI are better in your price bracket if you look around"
"OK, I'll buy an NVIDIA then".
So yeah, you're right, Mr £75 is smart enough to do his own research by asking someone, then stupid enough not to take their advice.
Well what I didn't suggest was that the people handing out the advice necessarily make the same vendor choice next time they come to buy their own graphics card. Halos can slip -- just go read the R600 thread here for a good example, a lot of very pissed off loyal ATI customers. Rampant fan-boys excluded, l33t gamers just want the fastest, so they'll flip-flop between vendors depending on whose got the performance crown. Next time round it's "dude, buy a Radeon".
For the record, blindly asserting that something is true does not constitute evidence.
Anyone doubting the importance of high-end cards should take a look at the current Steam survey. This survey usually shows a high tendency toward mid-range cards so it's a bit surprising to find 8800's near the top. Now how do you reconcile the irrelevance of the high-end with the fact that G80 is running far ahead of cheaper alternatives that have been around much longer? And then if you factor in ASP.....
A halo effect is when one product is sufficiently attractive to stimulate sales of other products.
Now, that said, during its 9 months of undisputed leadership in both performance and features, and being the only DX10 capable offering on the market until the last month, plus being available at prices down to roughly $300, the G80 variants has managed a market share of 3.8% in the Valve survey. And that is among online shooter players, it would be reasonable to assume that the volume WoW and SIMS markets are much less interested in these parts. (And the Valve percentage includes industry insider users.) And predictably, if we cross correlate with add-in board market data we can see that the Valve survey vastly overestimates the penetration of the high-end parts, even if we assume (justifiably) that just about everyone who bought a graphics card at $250+ the last quarter opted for a G80.
So it depends on what perspective you take. While you are right in that this percentage is higher than the norm for such parts, the message that I bring home from the survey data is just how low consumer interest is, even in this survey group, for such an outstanding offering as the G80 has been for the last 9 months.
But again, while this may help imply that the high-end has a negligeable halo effect, the actual market share of the high-end really is a different issue. There exists a market for $300+ cards with high power draw - we know that and we know the size of that market. It's tiny. The importance of that niche is another issue, and it might be useful to consider why it may be interesting to some parties to exaggerate that importance.
I didn't bring Steam up as evidence of the halo effect. Just like my post said, I brought it up as a counter-point for those people claiming that the high-end itself doesn't sell well like you did here.
Sigh. Why do you consistently ignore the fact that the universe of parts you are comparing G80 to has been around much longer than 9 months? So your "penetration" argument is invalid. If 8800's account for 7.61% of SM3.0 cards which includes cards from 2004 what percentage of NEW SALES since its release do you think it accounts for? Far more than 7.61% no?
Then when you consider that cheaper products sell in greater volume than expensive ones the penetration you speak of is actually quite high.
Er... that's another thread.
The existance, or not, of a high-end halo effect in graphics cards is theoretically independent of the market share of said high-end. In practise, since the price of a high-end (G80) card is quite low, it can be assumed that limited marketshare simply means limited interest in these products, hardly conductive to much of a halo.
Frankly, this is why I haven't responded until now. Your way with numbers shows that you aren't interested in information or facts, but in covering up reality as far as possible, for whatever reason. Why on earth do you use SM3 data at all? How can you assume that G80 cards will enjoy the same market share (or indeed remain high-end) for upcoming years? Does a factor of two make the slightest difference to the overall picture?
Why do you ignore the crossreferencing to market data which I referred to above, which shows that the total market share of $250+ cards is just over 2% for add-in boards. As I stated, (and as is obvious to anyone who actually checked the data,) this means that the Steam survey overestimates the impact of the G80 products. Furthermore, since high-end "add-in boards" do not apply at all to portables, (most PCs sold to private individuals these days), nor to systems sold and used with their built in graphics solution, the total market share among game players for the G80 products must reasonably be below 1%, probably much below. Will it grow a bit with time? Sure. Does that raise the high-end market share above an extremely small fraction of the whole? No.
The Beyond3D forums is a place that attracts 3D enthusiasts and industry insiders alike, from all over the world. This allows some fairly extreme opinions and priorities to prosper. Periodic reality checks may be in order.
I'm not sure why anyone even cares about the Steam numbers. The real numbers are out there: in Q406, which is the best quarter of the year, NVIDIA sold 400K G80s with ASPs of $125 for the chips, resulting in overall revenue of $50M and gross profits probably around $25M. That's with basically a 100% marketshare of the $399+ market, and IIRC without any contribution from the Quadro business.
If you want to compare that in terms of revenue, it corresponds to roughly 1/8th of NVIDIA's GPU revenue iirc, see: http://www.notforidiots.com/danvivoli/Slide7.JPG - that's hardly insignificant.
If you want to compare that to the overall worldwide PC industry's shipments, then that's 400K against 65.6M units in Q406, or 0.61% of the market. That is fairly meaningless however, since most of these PCs will never be used to play a 3D game. Also, how do you consider users upgrading to a higher-end GPU after buying a low-end one? etc. etc...
In terms of units, it's fairly obvious that G80-class products are insignificant. In terms of revenue, they're already more interesting. And it might be 1/7th or even 1/6th of NVIDIA's GPU gross profits. Now, if you consider that G80 is also used in the high-end Quadro models that bring in absolutely awesome ASPs and margins...
So overall, I agree that class of products is not very impressive from a unit POV. But it remains important from a financial perspective, which is also worth discussing on this subforum (but not only here!), and especially so when one of the two players manages to have a quasi-monopoly of that segment of the market (which is much easier than doing that in the mid-end or in the low-end).
However, it is true that the gross profits are not really high enough to dedicate a chip around that segment, because of the increasingly high per-chip R&D/design/tape-out costs. I think another key reason why G80-class single-chip solutions will die out is that IHVs must be realizing that so much power consumption & heat dissipation is bad from a marketing standpoint. Power efficiency isn't really what matters; it is perceived power efficiency that does, and 400mm²+ dies are out of the question there.
If you want to compare that in terms of revenue, it corresponds to roughly 1/8th of NVIDIA's GPU revenue iirc, see: http://www.notforidiots.com/danvivoli/Slide7.JPG - that's hardly insignificant.
If you want to compare that to the overall worldwide PC industry's shipments, then that's 400K against 65.6M units in Q406, or 0.61% of the market. That is fairly meaningless however, since most of these PCs will never be used to play a 3D game. Also, how do you consider users upgrading to a higher-end GPU after buying a low-end one? etc. etc...
In terms of units, it's fairly obvious that G80-class products are insignificant. In terms of revenue, they're already more interesting. And it might be 1/7th or even 1/6th of NVIDIA's GPU gross profits. Now, if you consider that G80 is also used in the high-end Quadro models that bring in absolutely awesome ASPs and margins...
So overall, I agree that class of products is not very impressive from a unit POV. But it remains important from a financial perspective, which is also worth discussing on this subforum (but not only here!), and especially so when one of the two players manages to have a quasi-monopoly of that segment of the market (which is much easier than doing that in the mid-end or in the low-end).
However, it is true that the gross profits are not really high enough to dedicate a chip around that segment, because of the increasingly high per-chip R&D/design/tape-out costs. I think another key reason why G80-class single-chip solutions will die out is that IHVs must be realizing that so much power consumption & heat dissipation is bad from a marketing standpoint. Power efficiency isn't really what matters; it is perceived power efficiency that does, and 400mm²+ dies are out of the question there.
ok someone get me £30 worth of the drugs Entropy is taking....
Well, a G80 can be had for just over $250. Most people who have any sort of income can scrunge up that kind of money in the industrialized world. If you're still too young to earn any kind of money, I guess you'd have to pester your parents.
If you work in a reasonably industrialized part of the world, and are interested in owning a top end card, you simply buy one.
I'm not interested in buying these cards though, even though I have the guilty pleasure of having 100 retail PC games on the shelf, and could easily afford them. Primarily due to power draw concerns, even for my stationary computers, but also value. Another notch up in resolution just doesn't make much of difference to my gaming enjoyment. (And now also since I don't plan on switching to Vista. Their added features are mostly useless to non-Vista users.)
But this thread was about the contention that I would be enticed to buy the 8600GTS due to the 8800GTX being the top-performance card. Erm... no.
Thanks for the data point.
It would be interesting to see the numbers for Q107 as well - as you say, the Christmas quarter is the best, and it would be useful to see just how special that quarter was, DX10 was made available around the same time to all early adopters and so on.
Do you have access to links or data for Q107?
Not only do R&D/design/tape-out costs enter into the equation but so do die yields vs. size and on-chip redundancy, assembly and testing costs, PCBs.... it is not a trivial equation by any means.
Just how the "monster-chip" approach has been justified up until now is probably a quite complex question as well, but that doesn't mean that the industry can/will pursue that path indefinitely.
It's particularly tricky to make market predictions right now. The dismal uptake of Vista (and therefore slow adoption rate of DX10) as well as the integration of CPU and GPU within the next couple of years are both significant. The market will have changed in two years, but how? Are high-end monster chips going to be phased out in favour of something more modular, or are they going to be the last stronghold for the gfx-ASIC as the low and middle end is eaten up by the new integrated approaches?
Similar threads
