The problem at Dell wasn't the CEO. It's the commodity R&D

Michael Dell's reassumption of the CEO's post at Dell is a shell game that can do little more than temporarily assuage Wall Street. Let's be honest.
Written by David Berlind, Inactive

Michael Dell's reassumption of the CEO's post at Dell is a shell game that can do little more than temporarily assuage Wall Street. Let's be honest. As chairman of the company, Michael Dell had whatever opportunity and access he needed (to the now-ex CEO Kevin Rollins as well as other Dell execs) to implement a course correction. When Michael Dell spoke, everyone listened and listened closely. A more honest assessment of Dell's current situation is that Michael Dell knew that the business model on which his namesake company was built and made fortunes would eventually run out of gas. Now, that time has finally come and the big question is, what to do next.  

A lot of people think Dell is a computer company. They're mistaken. It's a bank. One that's been trying to diversify, but a bank nonetheless.

A while back, I did a whiteboard video that explains why Dell is a bank (see The Bank of Dell). It works like this: Dell assembles PCs. It doesn't "make" them. It takes off the shelf parts from household name suppliers for processors, hard drives, memory, etc. and assembles them into the PC that someone is going to order tomorrow. The suppliers of these parts have inventory depots across the street from where all this assembling takes place and, since Dell has been in business so long, it has a very good idea of how many of each part it will need at 3pm tomorrow. And that's when the race starts (or perhaps I should call it the financial clock).

Just because Dell takes delivery of a hard drive at 3pm tomorrow doesn't mean it has to pay for it at 3pm tomorrow. Using common deferred payment terms (eg: "Net 15", "Net 30", etc.), Dell usually doesn't pay for that hard drive until well after it's not just assembled into a computer, but also after that computer has been shipped to its customer who has already paid for it. In other words, there's a window of time where Dell has collected money for a hard drive that it hasn't yet paid for. Now picture the volume of business that Dell does on these terms and you can begin to see why I see Dell more as a bank than a computer company. Making massive amounts of money on this sort of financial manipulation, often called "the float," is usually the domain of financial institutions. 

This financial efficiency didn't happen overnight. Inventory-laden companies are often measured on a metric called Day Sales of Inventory (DSI). DSI essentially measures how many days it takes a company to turn the raw materials in its possession into sales. The key? Don't possess the raw materials for very long. Or the final product for very long. Back in the day when the company was called PCs Limited and Michael Dell first pioneered the direct model (selling directly to customers instead of going through the retail channel), his DSI was not even remotely close to a day or two (as the aforementioned "3pm scenario" implies). It was probably more like a month.

But even so, a one month DSI was significantly better than the DSIs of the rest of the PC makers who operated through the retail channel. For them, the time it took for a raw material like a hard drive to finally end up in a computer on someone's desk was probably three to six months, if ever. Computers had to be manufactured, shipped in bulk to a wholesaler, and then from there, to the stores, and stocked on shelves (where inventory burns holes in financial pockets), and then someone had to walk into the store and buy it. And, unlike with other stocked merchandise (eg: a blender), Moore's Law was constantly conspiring against this model. Before too long, the inventory that was on the shelves was obsolete, costing the entire channel even more money.

Out of the gate, Dell's direct model already put the company at a significant supply-chain advantage since the entire channel inventory problem was a non-issue (or at least less of an issue). Then, for a while, Dell's competitors, with their legacy channel businesses, watched like deer in headlights as Dell whittled its DSI down by squeezing more and more efficiency and cost out of the supply-chain side of the business; everything from doing a better job predicting inventory needs to having suppliers across the street simply converted into a lower DSI which in turn meant more float time. 

But now, two decades later, Dell is no longer alone when it comes to supply-chain efficiency. Sure, Dell's competitors went through some painful conversions as they straddled the traditional sales channel and direct-to-customer businesses. But today, Dell's supply-chain story is no longer the competitive advantage that it once was. But that's not really what haunts Dell. Stripped of its huge supply-chain advantage, what now haunts Dell is that it has never been able to break away from the Fabrication, Assembly and Test (FAT) model. Dell is and always has been primarily an integrator of off-the-shelf parts. Dell claims it does Research and Development (R&D). But it's not the sort of R&D that Dell's suppliers do; the kind that yields really profitable intellectual property. Like other system integrators, Dell's real R&D departments are at Intel, AMD, ATI, NVIDIA and its other suppliers. Unfortunately for Dell, it's the same "commodity R&D" that the company's competitors have access to.

So, if the R&D playing field is level and the supply-chain field is level (or the gap is significantly closed), is there anywhere for a FAT-specialist like Dell to hide? 

In meeting with executives over the years, Dell has clearly tried to diversify. It has gone the TV and printer route (and maybe the consumables market will eventually pay-off for Dell the way it has for HP). But in terms of hardware, the area of servers and networking infrastructure is where the gold is if only you can close the business. Dell may have made some good inroads there. But, if desktops and notebooks are difficult markets to be in if all you have in terms of competitive advantage is an efficient supply chain, then servers and network infrastructure are an even worse place to be. For starters, while it's always an issue, supply chain efficiency doesn't matter nearly as much with servers and network infrastructure since the sales lead times tend to be longer, the deliverables are far more customized to the end-users' requirements, and buyers generally don't need the products within a day of ordering them.

But to make matters worse for Dell, relying on commodity R&D really puts a server and network infrastructure player at a significant disadvantage since it's in this segment that R&D can make or break a sale. 

I've always been suspect of Dell's server strategy, calling it for what I think it has always been: a strategy that's tied to Intel's R&D. Dell has shown us what it believes to be its own innovations on top of the R&D of others. But the phrase "blue blinking lights on the front" comes to mind. I remember a server briefing where some Dell server product manager was explaining how the company had, based on customer feedback, put the blinking lights on its rack mounted networking infrastructure on the front rather than on the back. Who am I to poo-poo what the customer is saying? But is this the R&D on which fortunes are made? Not quite.

Dell has forever said the way to scale your computing infrastructure is to scale-out. Coincidentally (not), Dell's ability to scale up was completely tied to whatever Intel had to offer it in the way of off the shelf parts, none of which were scale-up parts. To do a two, four, or eight way system, Dell had to wait for Intel (or some other supplier) to offer two, four, or eight-way parts. Dell was incapable of doing what IBM and HP were doing: R&D that yielded multiprocessor server chip-sets (EXA and ZX1) that could scale Intel's processors to 16- and 32-way systems where some real margins live. Even so, there was a period of time where, for customers, the penalties for scaling out versus scaling-up were few. The hardware was cheap. Energy was cheap. And the Internet wasn't yet a viable medium over which to distribute computing off premises. So, a commodities player in a scale-out market (validated by Google's approach) with an efficient supply chain like Dell's has a good shot as long as market conditions remained unchanged.

But then, everything changed when it came to that one potential sanctuary of margin (servers and infrastructure). Between managed hosting services or something like Amazon's EC2 and S3, buying your own servers and storage makes far less sense than it ever did. Not only that, if you must stay on premises, then with technologies like VMWare and XenSource on the loose, why buy five servers when you can get away with better utilization of one (one that looks like five). And then, with both Software as a Service (SaaS) and services oriented architectures (SOAs) finally proving themselves to be viable, components of a business process (or the whole enchilada) are far more easily outsourced than they have ever been. Think about it. For every customer that Marc Benioff signs up to use salesforce.com, that's probably one to twenty less servers that get purchased by some business. 

But it gets worse (for Dell). The energy situation is a mess. Talking scale-out smack (even though that's all Dell could do because of how it relied on Intel's R&D) may have worked when there wasn't a war in Iraq and there wasn't a giant question mark hanging over the energy sector. But now, scale out isn't looking quite as smart as it once did. Generally speaking, scale out means more machines: Machines that need their own electricity; Machines that need to be cooled off. Suddenly, having lots of machines driving up your energy costs isn't looking so good. Even worse, with no real R&D of its own to correct the problem, what's a commodities player like Dell to do? Answer? Hold a press conference to talk about how energy can be saved by scaling out. I laughed. I cried "Say what!!!???". I recorded the press conference for podcast and wrote about it

Now, energy is on everyone's mind and Dell's reliance on its suppliers to give the company a compelling energy story is out in the open. The emperor's clothes are off. Whatever Dell's suppliers come up with to address the energy situation will be the same, at the very least, as what Dell's competitors (those suppliers' other customers) will come up with. Meanwhile, companies with real R&D -- companies like Sun -- are showing up in the market with scale up solutions (Niagara) that save enough energy to motivate utility companies like PG&E to issue rebates for buying them. I personally can't validate that Sun's current or forthcoming scale-up offerings perform better at a reduced energy cost. Others including fellow ZDNet blogger George Ou have questioned the claims and no standard benchmark involving energy consumed, overall cost, and units of compute power currently exists.

But I do believe two things. First, scaled-up systems (especially when virtualized) will eventually prove to be so much more energy efficient than scaling out that even Google will have no choice but to reconsider the approach given its potential impact on the bottom line (Google is clearly keeping an eye on the situation). Second, whether it's energy or some other important end-user reflection of market conditions, Dell, with its heavy reliance on commodity R&D, is incapable of coming up with any potential big game changers on its own. Not for its customers. Not for itself. Now, with Dell's supply-chain advantage having been whittled away, it's going to take a lot more than a shell game in the management ranks to regain its competitive edge.

Finally, if I had to point to one other problem at Dell which may be a source of its woes, that would be agility. Yesterday, as I was putting the finishing touches on ZDNet's coverage and 42-slide image gallery of four Windows Vista-based notebooks from Toshiba, I came across a notebook from Asus (shown here in the image gallery) that had actually done a better job than Toshiba at implementing a Microsoft technology found in Windows Vista called SideShow. Both Asus and Toshiba were at CES showing off their Vista "innovations." Other than recommending Windows Vista for its existing systems, Dell, which in the past has been known for some innovation in the area of mobile computing, has been relatively silent.

Michael Dell needs to be asking someone at the company why a Dell-branded PC wasn't on stage at CES with Bill Gates instead of the Toshiba Portege. He also needs be asking why Asus (a company that hardly anyone has ever heard of) upstaged every household name in mobile computing, especially Dell, with an innovative Vista-based notebook. Finally, he needs to ask himself why his company's own home page for Windows Vista buries the path to buying a Dell computer with Vista installed on it (on the lower left-hand side). Not that I have any vested interest in fixing Dell's problem. But having once run a computer shopping site (computershopper.com), I know that if the objective is to sell something (and what other objective should most pages on Dell's site have), then the ability to buy something should be top, dead, center.

Editorial standards