When you make the rounds of the venture capitalists, words become a vital part of your asset base. So important are the meanings attached to particular words that old Mr. Webster (the lexicographer, not the legislator) would probably have gone into investment banking had he been making a career choice nowadays.

The situation goes something like this. One of the most important issues for the VC is salary leverage. Salaries make up a good part of the appetite that eats away venture funding, and the investors want to know that the salary expense can turn into not just good revenue growth, but explosive revenue growth. They need the occasional explosive earner to pay for the 95 others that don't survive to go public.

But salary leverage is all about how the money you pay your engineers turns into the money customers pay you. In a design services model, for instance, you hire geniuses, but one genius can support only so many clients. So the ratio of salary expense to revenue, give or take, is a fixed factor. Bad news.

Compare that to the intellectual-property model. You hire geniuses, who write some RTL one weekend. The code is widely recognized as the next Pentium and licensed by huge numbers of clients, each of which sells huge numbers of royalty-bearing products. The ratio of salary to revenue is limited only by available market size for your customers. Good news.

The problem is that most VC firms lack either the time, the inclination or, shall we say, the intellectual bent to figure out just what the connection is between salary and revenue for your particular company. So they try to extract this data by visual pattern recognition: They look for particular words in your business plan.

That's why nobody puts "design services" in their business plans these days, but you see lots of "fabless IC vendor."

But dig deeper. If you talk to those fabless IC startups in these lean times, you soon find out that, to attract corporate funding or even to attract evaluations, they've reworked their "standard-product" design to include lots of special, customer-specific features. Naturally these features pull the product in the direction of being a customer-specific chip. Nay, say the chip designers. Yes, says the silicon.

So in what way do these startups, which are now doing customer-specific COT designs for systems OEMs, differ from design services companies? Why, just try to find those words anywhere in their business plans.

http://www.eetimes.com/