STRATEGY - Small change, big plan

Intel wants its Celeron processor to create more competition at the low end. How will a market with an upgrade mentality and concerns about margins react to the idea?

For those who take an idle interest in the computer industry, thehe low end. How will a market with an upgrade mentality and concerns about margins react to the idea? past couple of years have been rather interesting.

Until recently, the industry's strategy was to throw things into the marketplace, watch what worked or didn't, and then launch a similar product. Software companies which had one good idea were usually unable to come up with another - so there would be one good, successful product followed by a series of daft ones, swiftly followed in turn by the demise of the company and its product.

Similarly, players in the hardware industry were obviously making it up as they went along. New standards fixed problems that had arisen a few years before and would then introduce two more.

But in the past couple of years, all this has started to change and these companies seem to be marching to a better tune.

However much it might be disliked, Intel has been looking more and more like a company that is working to a plan. In fact, since it introduced the 386 (does anyone even remember the posters showing the 286 chip with a big red cross through it?), it has had to be much more savvy about the future.

Every 18 months or so the company had a different iteration of processor to sell, which would promptly alienate those who had just bought the earlier, out-of-date processor. It needed a way of handling that and staying in business.

Add the likes of AMD and Cyrix snapping at its ankles and some sort of strategy was needed.

Since Intel realised that it could no longer rely on the 80X86 label, it has been forced to apply its talents to producing names for its chips.

Thus its latest offering, the Celeron, highlights the fact that if Intel dared to try and survive on its product-naming ability alone, it would have been dead well before Ashton Tate and MicroPro.

As it is, Intel, unlike Microsoft, has shown a market canniness that has not lead to envy and vilification. The Celeron launch looks like part of something that could be called a plan - although that could easily be an optical illusion. It seems to be a real attempt to create a low-cost PC with specific low-cost PC hardware, as opposed to just one with discounted parts like hard disks or monitors. It also seems specifically designed to introduce the likes of AMD and Cyrix to the Titanic principle - and experience suggests it would be wise to start looking in their direction for lifeboats.

Based on the P6 architecture first seen on the Pentium Pro, the Celeron is the processor previously known as Covington. As such, it was demoted by Intel chairman Andy Grove in February. Although Intel seems to be coy about its speed, it is reported to be clocked at 266MHz, which is important because Intel believes the clock speed number is a fundamental part of the sales pitch. Celeron will have a 32K cache, but not the other nice go-faster bits that drove the Pentium Pro and made it so expensive to produce.

The Celeron comes with a cheaper Slot One and there are unconfirmed reports of a new chipset around it called the 440EX - a cut-down version of the current 440LX chip set.

The Celeron PC is to be priced between $800 and $1,200 and has limited expandability. It will support only one processor, two DIMM sockets and three PCI slots. It has become apparent that the P6 core technology is going to be the basis of this and the more powerful systems which Intel actively separate as being for the enthusiast or professional market.

The word is that there will be a new brand name for the low-end processor range.

The launch also marks a change in the way these things are done. Usually, Intel comes up with a new processor and then - while desperately trying to get the yields up - launches it. Then the processors using that technology get faster and faster - and the people who moved to it in the beginning grow more and more irritated as they are left behind.

Now it looks as if Intel is - for the time being at least - sticking with a processor technology and is prepared to develop it horizontally rather than vertically, launching specifically targeted low-end processors and associated chipsets as well as high-end processors for the low-powered users to aspire to.

According to the Celeron press release, the market will be broken down into three categories. The first is the Professional PC, which 'has the power to handle visually rich content applications while providing superior multitasking capability with manageability innovations'. (How nice to see that the engaging management speak of the 80s is still with us.) Intel will aim this at the $2500 price range.

Then there are Performance PCs which are 'targeted at home and business users alike who are looking for a full-featured PC that offers solid performance and flexibility to run the latest education, creativity, and entertainment applications'. These are priced at between $1,300 and $2,400.

Last and least is the Basic PC pitching in at between $800 and $1,200.

These 'provide a base level of functionality with limited expandability to meet the core needs and affordability requirements common to many new home and business users.' Which decoded, is likely to mean 'cheap PC'.

Despite the toe-curling hyperbole, there is at least an attempt to split the market into specific segments that buyers will understand - at the moment 'segments' is an important word at Intel.

Obviously, there will be problems if the market doesn't want to play the 'three niche game'. Chris Hogg, market development manager of Europe at Intel, claims the evidence for the need of this stratification is there.

'We see in the US an emerging market segment for sub-$1,000 PCs, probably being bought by first-time users but undoubtedly by business users as well,' he says. 'We think there is an ability to grow this segment of the marketplace because there are people going in to stores buying PCs who hadn't bought PCs before, and we think there is an additional marketplace we haven't been addressing.'

But what kind of businesses will be looking towards a brain-dead Pentium Pro for the office? Hogg replies: 'They are the people who know exactly what they want to do in terms of fixed applications; they are not that worried about running next year's operating system or next year's applications.

Affordability is a key driving factor in their choice.' And it is for these people that Intel will be launching the Celeron processor.

But doesn't the sub-$1,000 PC suggest a whole PC, not just a processor and chip set, and will Intel be building them? 'We aren't going into PC manufacturing,' says Hogg. 'All we say is people are likely to use these chips in systems that are sub-$1,000. Obviously, we don't have any control over the entire system cost, although in order to get to those price points, there are a number of things you can do apart from produce a low-cost processor.'

Such as? 'Well, we have defined a new motherboard form factor called MicroATX, a really dinky motherboard. And along with that, we have been working with power supply manufacturers to produce some cost-focused power supplies as well. So it isn't just the processor, it's other aspects as well.'

An obvious question is, will Intel seek to influence the price of the PC that comes from the manufacturer? 'No,' says Hogg firmly. 'It is up to the marketplace.'

Does this mean the price for all low-cost PCs will be #799 then? 'People buying high volume for business will probably be paying less than #799, but realistically in the retail market model that's going to be #699 or #799,' replies Hogg.

So it is sub-$1,000 - sub-#799 in the UK - only if the manufacturers make it so, which is really no guarantee of anything. This looks like just another Intel marketing scam. Hogg, not too surprisingly, disagrees.

'What we are going to see in the market in 1998 is segmentation,' he says. 'We are no longer doing a one-size-fits-all, where the user has to have a full-spec Pentium for every market segment. The focus will still be on Pentium II at the high end on the servers and workstations.'

Apparently, it is for the good of everyone. 'We are trying to put a gap between the basic PC and the performance PC,' says Hogg. 'At some point the basic PC may adopt some of the performance PC technology, but the performance PC will be markedly more powerful.'

The other side of the equation has to be the manufacturers - if they don't buy into the low-price segmentation then the battle is lost right from the start.

Dave Cheetham, product manager of desktops and servers at Acer, believes in the new strategy and also thinks it is good for the market. 'The cost-down chipset, together with the cost-down processor, is essential to hit some of the entry-level prices we need to hit this year,' he says.

'Intel is obviously looking to push prices of the Pentium II down to prices approaching those of the current Pentiums, and you can't do it with the current technology. The cost of the motherboards, chipsets and processors don't allow that, so it is very important that we get these cost-down components.'

This thing isn't going to go off half-cocked, either. Acer is ready to go with the new technology.

'We were displaying the new motherboards with the new chipsets and processors at Cebit,' says Cheetham. 'They were engineering samples but we hope to have the finished systems to market by about June or July.'

Cheetham doesn't think this will immediately exclude the likes of AMD and Citrix. 'There are a couple of issues there,' he says. 'For a start, it is cheaper to produce a Pentium motherboard than a Pentium II motherboard.'

He points out that Acer runs AMD K6 processors on Pentium motherboards.

'So if you are asking if the cost-down processor is going to be the last nail in the coffin of AMD and Cyrix, I'd have to say I don't know, but we will continue to support all the technologies.' He adds: 'For example, in the US and Germany, we ship vast quantities of the K6 systems.'

Then there is the corporate market. How are the corporates going to view the machines? Ray Badminton, desktop product marketing manager at Dell, thinks there might be a little confusion.

'There is an element of education that needs to be done to the market,' he says. But he believes the move is entirely logical and that the corporate market will understand it. 'Intel realises it has to be more competitive in the low end of the market, particularly with the likes of AMD and Cyrix around. The corporates are viewing it as Intel differentiating between different elements in the market.'

All the same, is the differentiation really necessary? Does the industry really need a new brand name - or is this just another marketing exercise?

'I guess there is an element of that,' admits Badminton. 'But any company in the world with a reasonably strong marketing operation will try to position its products into different market segments. Look at BMW, for example. The BMW 3 Series offers exactly the same build quality as the 5 or the 7 Series, but they are all in completely different markets and price ranges. I think Intel has probably looked at this and said "we can do this as well".'

So it seems there is an Intel plan, at last. After a mere 17 years, we are seeing a technologically defined stratification being applied to the industry. The only question is, will the price drop actually get through to the punter?

It is all very well for Intel to produce cheaper chipsets and tell the whole world that the sub-$1,000 (or somewhere around #699 or #799 - no one is quite sure) PC is here, but what if the channel, which has seen its margins all but disappear over the past five or six years, decides to pay the rent?

And what will Intel do if it finds people selling Celeron PCs at more than $1,000? Can it control the market that tightly, and if it does, will the subsequent bad feeling take the heat off Microsoft? The answers will be with us in the next six months.