Despite all the talk about adding value and services, some resellersservices - so say some. Others believe there's cash to be made if hardware is sold in the right way. still believe you can make plenty of money out of selling desktops - if you do it in the right way.
Philip Mitchell, managing director of IntraLan, says the reseller can successfully and consistently make money selling desktops to SMEs. 'On desktops, we make between £100 and £150 per seat, and then there are a number of services on top of that - usually ranging from between two and four hours' work pe box.'
Sometimes, the company sells Compaq PCs, but often it's a machine built by IntraLan itself. For SMEs, the price difference is significant, he says, especially on bigger boxes, but it is important to use quality components that have been properly tested and affirmed as fit for the purpose by Intel, Microsoft and other key vendors.
This is an important weapon against the direct vendors, he says. IntraLan is happy to work with machines that it has not supplied, but Mitchell insists they are quality systems and asks the user questions that they might not have considered - such as, is the machine network approved?
'The Dell Dimension range, for example, is not network approved and Dell won't support you if you attach a Dimension to a network - you have to buy the Optiplex range. That, of course, costs more and the same is true of all manufacturers' ranges, even Compaq's. Some are approved for network connection by Microsoft and Novell, others are not,' adds Mitchell.
Some smaller vendors even use motherboards that are not approved as Microsoft-compatible, says Mitchell. Knowing this and offering the user a lower-cost alternative with fully approved components, built and supported by a local dealer, can prove very effective in selling the idea of your own-built boxes. And, Mitchell believes, it is often the best option for the SME user as, under the box, the machines are more or less the same.
As for the cost of building them, Mitchell thinks it is a false economy to allow third parties to build. 'We have had people offering to sell us pre-built machines for £50 less than we can build them. We have been down that route a couple of times, but you are always dependent on their stock and their build quality.'
Ian Brooks, managing director of IB Business Developments in Cardiff, also supplies small businesses, but he takes a very different view. IB has just signed up as the first Gateway partner in the region. 'When we started out, we surprised some customers by saying that we didn't only sell computers. If they bought PCs from us, they'd have to buy the services too. We also said we'd allow them to buy from other places as long as it was not rubbish. But when push comes to shove, they tend to want to buy it from one place.'
To that end, Brooks struck a deal with a local Escom store, but that, of course, was ill-fated - the German manufacturer pulled out of the UK.
It then switched to recommending Gateway machines and did very well for a company not selling PCs, shipping £180,000 worth in 12 months. It is now an authorised Gateway distributor and makes a small margin on products.
But why Gateway and not a channel brand? 'The SME market is generally uneducated - to the buyers it is just a box. They can't perceive any difference and as long as it works, people will buy it. I have never sold a Compaq PC - I think they are overpriced and a bit proprietary. The company tends to try to differentiate its PCs by saying they will do things differently, and seems to be trying to drive the idea that if you buy a Compaq machine, all the add-ons also have to be Compaq.'
In the corporate market, the situation is different, but brand is not as important as it used to be, says a representative at Computacenter.
'It is still a healthy market but incredibly competitive, and brand recognition is decreasing. The differences between the products is decreasing due to the consolidation in component manufacturing.
'Customers are much more concerned about their desktop service provider.
They want to know what happens if it goes wrong - what if they want to refresh the model? That's what it is all about, not the product features.'
With these attitudes in evidence, it would seem that the key brand vendors have little choice but to look at their models and product lines. But clearly there are no easy answers. The approach and opinions of resellers vary enormously. Meeting the needs of them all might be impossible.
Resellers acknowledge that vendors have to make a change and respond to the direct threat, but they don't want to see a mixed-up policy that brings direct and indirect activities into conflict. Many resellers do not believe the hybrid model can work for vendors anyway, even in the consumer market - it has to be one thing or the other.
Imagine this scenario - it is Friday and a customer is calling Q Computers with the requirements for a home PC. Everything goes swimmingly - if the customer understands the stream of bits and bytes - and after the buyer receives details of the various options that are of interest, asks: 'How can I buy one of these?'
'I can arrange for one of our Q partners to call you on Monday,' says the telesales executive. 'What about tomorrow?' says the customer. 'I'm afraid our partners don't work at the weekend, but if you give me your work number I can arrange for them to call you on Monday.' The customer groans, but has to agree.
This quasi-direct model fails the first test of good business practice: 'Does it make it easier to buy?' It also fails on simple economics, says Jonathan Chapple, chairman of Equanet. 'If you look closely, it is adding an extra cost layer to an already expensive sales chain,' he says. 'This can't last. Such a model will only survive if it becomes fully automated - that is, it connects callers automatically and seamlessly with the appropriate reseller.'
In other words, human costs must be eliminated by technology and telephony.
All this is possible today, but vendors are just not doing it. And even if they do, they should not abandon the channel.
Matt Tedstone, business development manager at corporate reseller Tplc, says: 'The problem with a direct sales model such as Dell, is that it does not create any demand for product - it simply supplies into an existing demand. This is parasitic,' he says.
'If we are to support a successful and thriving IT industry into the future, we need to create other markets for applications and for the hardware which those applications leverage.
'It is simply not viable to cut out the reseller. It may deliver short-term financial benefits, but in the long run, manufacturers which look for short-term gain will suffer long-term pain,' adds Tedstone.
But there is no sign that Compaq, IBM, Hewlett Packard, Siemens or any other big-name vendors are going to abandon the channel completely. As one reseller points out, they are responding with channel-build initiative, but these are only just beginning to take effect.
'We are working with IBM, HP and Compaq on their schemes, but I don't think the speed to market is that much of a problem. They are having to respond more quickly these days and channel-build schemes will help, but it's early days. They will allow us to deliver a wider range of products to the customer at a higher quality and, of course, more quickly.'
Corporate customers don't necessarily always want the very latest processor in their machines.
Consistency and low running costs are important to them and that means as few upgrades as possible.
PC manufacturers need the channel because it is the best route to market for most of their products most of the time - especially if they plan to address the SME market. The channel has far better choice - no contest - delivery superiority and equally good, if not better, technical support potential.
'If volume growth dictates that key manufacturers must move downmarket to maintain business momentum, then they must create real direct business models for the SME market and distance them from their existing corporate brands,' says one reseller. 'If, for example, Compaq had bought Gateway instead of Digital, it might not have lost its chief executive.'
But the industry should not abandon the idea that money can be made on hardware sales. This could be a big mistake, as resellers might leave themselves exposed to the competition. Resellers which sneer at the idea of product supply may leave themselves exposed to attack from good product supply-focused companies.
Although Intel is under pressure from AMD's K6 and the eagerly anticipated K7, its low-end product, Celeron, is gaining in the credibility stakes.
Neal Grayston, managing director of Tulip Computers, believes Intel has responded well to the AMD and Cyrix challenges. 'While the first iteration of Celeron was a bit of a damp squib, it has now really come into its own. The market has very definitely polarised around the Pentium III at one end and the Celeron at the other, squeezing out much of the competition occupying the middle ground.'
But that could change if the rumours about the K7 are true. Reports from vendors say the initial tests for it are impressive, and with Compaq, IBM and others looking to use the alternative processors where they can, Intel will have to step up its programme yet again. If this is the case, it raises the possibility that the chip giant could try to flood the market with low-cost products, push life cycles too fast and leave the channel exposed to stock problems.
The potential involvement of IBM and Hewlett Packard in the development and support of the Compaq's Alpha processor must also be a concern for Intel as it strives to develop the IA-64 processor as quickly as possible. Although it makes Alpha chips under licence, Mitsubishi and Samsung do as well.
The arrival of IBM in the lab, with its knowledge of copper technology, would also be a significant and worrying development for Intel and its prospects of dominating the desktop. The disposal of Cyrix by National Semiconductor might also prove significant.
Even so, Grayston believes Intel will continue to dominate in the desktop sector. 'Although some serious competition would be a welcome fillip for the market and for margins, I don't think it is realistic. In the end, Intel is just too powerful. It goes back to the old moniker of not being fired for buying an A- brand and, as a result, there is always going to be this fear factor of buying into what may be seen as black sheep processor brands.'
Intel's low-end dominance in the processor market has been eroded recently - by AMD in particular, and by Cyrix. Due to the Celeron and the Pentium III, Intel appears to be enjoying a resurgence. AMD's K7, however, could change all that.
Intel has an aggressive launch plan for 1999 and will continue to launch its Coppermine and Xeon versions of the PIII for servers and notebooks as well as desktops. The Celeron will be enhanced further to address the growing consumer desktop market.
Intel is shipping 550 MHz PIIIs and has announced its intention to use its 0.18 micron technology to take performance beyond the 600MHz threshold and up to 733MHz later in the year.
Additionally, the chip manufacturer plans to maintain its 100MHz bus BX chipset, introducing a 133MHz bus version, and to launch a 433MHz product later in the year. Much of the company's focus is on the IA64 and mobile chips - Intel sees both as strategically important in 1999.
After the success of the K6 range, which fostered growth at the low end, and a successful alliance with Compaq, AMD is about to announce the release of the K7 processor, a direct competitor for Intel's PIII. Rumours abound that early testing of the K7 has produced some highly promising results - reports of unofficial benchmark testing in Germany suggest that the K7 platform is more than a match for Intel's high-end PIII Xeon.
While this should provide Intel with some food for thought, AMD still looks like David set against Goliath.
But, sheltered by some of the big PC vendors that want an active rival to Intel at the low end, AMD can continue to do well if it comes up with the right products. A great deal hangs on the K7's performance and on AMD's ability to produce it in large enough numbers at the right time.
Although it has been admired for its design technology and innovation, Cyrix has always struggled to make a real impact in the low-end processor market. As National SemiConductor has withdrawn from the semi-conductor market, Cyrix's future is far from certain.
According to several sources, there are already a number of parties interested in buying the x.86 manufacturer. National Semiconductor has confirmed that it has been approached by several prospective buyers and speculation suggests that ST Microelectronics, Taiwanese foundry TSMC, Samsung and even AMD are among them. This could mean one less rival for Intel to worry about.
MSP plans to use new acquisition to expand its security offerings
Reseller also saw its operating profit fall five per cent in its financial 2017
Wendy Bahr to bring 18-year spell at networking giant to an end
AdEPT says latest purchase will push revenue beyond £50m