Fog Creek Software
Discussion Board

Strategy Letter V vs. Spring in Cambridge

I think my attempt at an eyecatching subject has failed miserably, but here's my point anway:

Joel points out the big win IBM is making by becoming an IT Consulting Company.  However, he already said that, "A consultant working on a project makes ten or twenty percent profit. If you license software, each marginal unit you sell is 100% profit. That's why software companies have valuations that are something like 25 times richer than consulting companies." (

Is this the only option IBM has to capitalize on Open Source?  Why else would they make such a move?

Joel - Thanks for a great site.  If you're looking to open an office in Boston drop me an e-mail! ;-)

Sunday, June 23, 2002

Not everything IBM does is open source (WebSphere for example).  By becoming a premier consulting company IBM essentially gets paid to put salespeople in your organization to suggest that you buy WebSphere, IBM Servers, IBM Mainframes, etc.

Not that they won't use SUN, HP, Microsoft, etc. if you want them to, but all things being equal, they will use IBM products.

Not every move is explained by complements.  There are many reasons to do something, and  being a consulting company is a great way for IBM to build a channel for its other products, even if the margins are not great.

Monday, June 24, 2002

"and being a consulting company is a great way for IBM to build a channel for its other products, even if the margins are not great. "

IBM consulting strategy walks a fine line.  They want to screw up enough to give the hiring company a feeling that they would be lost without IBM in all of the complexity of what they are trying to achieve, but not enough to show this strategy of incompetence.

I have worked with them on many projects.  There is a limit, but they have a wide range where the more they screw up the more money they can milk.  And they do.

Joe AA.
Monday, June 24, 2002

The services business isn't always an easy business to be in - this post from Larry Augustin lays it out quite nicely.

Stuart Fox
Monday, June 24, 2002

Interesting, but IBM is neither a services pure play nor a product pure play, but somewhere in between.

Any comapany with products needs a channel.  Microsoft's channel is Windows for all products except Windows (and OEM's for Windows).

IBM uses it's services arm as a channel.  It doesn't really need to make that much money as long as it doesn't lose too much.  Remember, you pay IBM to put its salespeople in your company.  This results in more sales of IBM products (the point of the channel).  IBM sells enterprise software that costs $100'sK.  The gross margins are huge, but the potential customer base is small (compared to Word, Excel), so IBM's channel needs to:
(a) have access to enterprises that might use the software
(b) touch C level executives (CEO, CIO, etc) who have the buying power

That's why IBM benefits from a services business more than EDS does even if the profitability of those parts of the business is the same.

Monday, June 24, 2002


Uhm hello!!!!
This article is makes sense only if you assume EDS' consulting is done by elves at night instead of paid engineers. Think about it: Couln't it also be that EDS' margins are so low because they DON't do research & development, i.e. every consulting project starts with a blank slate, and no proprietary code is ever reused!!!!

Think about DELL, what is more likely- that DELL wastes money on R&D because they are stupid, or does the 1-2 percent R&D that they do create the automation they need to kick compaq's and HP's ass, by REDUCING OPERATING COSTS, which is one of the potential benefits of R&D (The other being competative priducts, and innovative products with no competitors)

In both EDS/DELLs case R&D should reduc operating costs by more than the R&D itslef costs (otherwise what i sthe point of technology). There are two catches though
1. Over automation, some operations (assmbling pc boxes) are much harder and more expensive to automate than others (creating a image database for installing software on your pc boxes) so for dell more than 2 percent  R&D would waste money. Like any other business expense, dont pay more for automation than it is worth

2. The second is accounting (this is not at all my are) but maybe EDS does a ton of R&D and simply does not want to show it on the books

Daniel Shchyokin
Monday, June 24, 2002

Actual R&D is usually capitalised and so appears on the Balance Sheet and not on the P&L which is where Operating Expenses live.

You can capitalise the entirety of R&D including the staffing and establishment costs.  There's usually significant tax advantages and grants available under most regimes.

There used to be an equation about the proportion of R&D Spend to Net Worth, for a strictly technology company (ie one that doesn't really involve itself in service revenue or significant product trees), that proportion can be quite high.

Twenty percent I think would be a healthy norm for a company such as Dell or HP.  A company like Microsoft, however, which is fundamentally not a technology company but a marketing company, I'd expect the ratio to be much smaller.

The reason I said 'used to be' is that the sane model of defining a healthy technology company went out of the window about 8 or so years ago and the five year silicon valley cycle replaced it (three years at a humungous loss, 1 year at approaching profit, sell in the 5th year, or IPO which is the same thing, on the basis of future profits).

After all the shakeouts, we might approach a more sane view in the future.

Simon Lucy
Tuesday, June 25, 2002

"Twenty percent I think would be a healthy norm for a company such as Dell or HP. A company like Microsoft, however, which is fundamentally not a technology company but a marketing company ..."

Interesting. Could you elaborate on this? Personally I thought the opposite would be true. Dell, mainly a pure distribution play would seem to me to be fundamentaly operations & marketing, wheras Microsoft seem to depend on perpetual innovation, and thus need a substantial R&D effort. Since R&D is rather scale invariant, I am all the more surprised by what to me seems an extraordinary R&D percentage for such a large company.

Just me (Sir to you)
Tuesday, June 25, 2002

Microsoft's R&D largely comes out to be product development.  They purchase or make use of R&D happening elsewhere, but again its almost always product related.

You could say that the NT development was R&D but that core R&D was done elsewhere originally, DEC.  Perhaps the one original development Microsoft have made was in using the double exception, shut down the processor to move from protected mode to real, but that now is all superceded.

Compaq/HP might be a better model of a technology company than Dell.  HP spends considerable amounts in medical equipment research areas, Compaq should be spending considerable amounts in R&D in motherboard designs, ASICs, novel bus management, etc, etc.  Even Dell should now be spending around 20% of their net worth just to ensure that they have further to go, cheaper to produce.

This not to say that a software company can't be a technology company, of course they can.  I happen to think mine is in a very small way.  When, though, the software company depends on churning their existing products in dot releases or changing their revenue model from licences sold to a subscriptions or fixed term leasing I think you can say that they certainly aren't a technology company.

Simon Lucy
Wednesday, June 26, 2002

Just as a further clarification, the average proportion of Business R&D to National Net Worth is around 2% for the G8.

Individual businesses tend to show between 1-3% R&D spend on their balance sheet.  I'd still maintain that a technology company should have a much higher proportion spent showing on the balance sheet. 

The vast majority of businesses though show no R&D at all they are middle tier or retail businesses.

Simon Lucy
Wednesday, June 26, 2002

*  Recent Topics

*  Fog Creek Home