What a cruel title. But if it drew your attention to continue to read that is good; if you are involved in the marketing of analog parts that is even better. If you come with an open mind about your work, welcome. My bandwagon today is one that a number of manufacturers have heard in the privacy of their plants. And, please, I am not anti-engineering. Engineering is the word on my academic qualifications, it's the word on my professional memberships; it has even been the word on my business card from time to time.

But I have also been in marketing and I have had a golden rule that even the engineering departments have been happy to go along with. If you can't show how many you can sell, at what price, don't make them. Each potential product should have a discrete business plan and forecast P+L.

This is a fundamental in a manufacturing business and on the digital side of electronics it seems to be followed. In our analog world we seem to have a problem, probably not singular.

Let's put volume and cost into perspective. The majority of analog semiconductor vendors, with their own fabrication facilities, have little problem meeting their cost targets for the majority of their products. The exceptions tend to be in areas where they are fighting another technology -- using semiconductor devices for isolation, for example -- and, obviously, where the arrangements with an outside vendor are insecure or poorly arranged or managed.

To be a Home Run product a couple of years ago you had to produce a product that sold 1M pieces per year. Most manufacturers would accept that because of shortening product lives in general, the number is probably now 5M pieces per year. You might think that this would not apply to "exotic" devices, but then you have to question what exactly exotic means? Is that just expensive? Well, if it is and it is not a consumer part then it probably should be made under individual contracts. If it is consumer then the price is immaterial, the quantity should be there. So, does that mean that a product selling at the 1M piece level is a failure? No; it's not a home run, but it wasn't a mistake either.

Because, Because

If a product isn't a winner it must be a loser? No, that generally is not the case: It nearly always turns out to be a "because" product, labeled with one of any number of excuses: "Because we needed to make them to fill out our product line." "Because we needed to make them so we had a complete solution." "We had to make those variations because we didn't know which were going to sell." "We had to make them because company B is making them." "We had to make them because the [choose one or more] CEO/VP of Engineering/Chuck (our sales guy in LA) said so." I doubt I've heard all the excuses because there is probably no limit to the varieties. If you examine them all, however, they have a dual-thread that seems to be constant. The first thread is that there is a genuine fear driving the decision to allow the product to be produced, sometimes a fear that a customer will go elsewhere, sometimes pressure from a single -- or higher -- source, sometimes a fear from the inability to read the data. It is a very difficult message to accept that you shouldn't do all that is possible to get all the potential business out there. But you have to accept it: There are thousands of businesses who go bankrupt every year after increasing their revenues. The fact that it is possible to make something is not a justification to do so.

The second thread is a simple inability to follow the rules of analyzing the marketplace and then defining the products; producing all the parts you can in an ý la carte manner to make the catalog look like a complete matrix works only on Chinese restaurant menus. It doesn't cost the owner anything but ink to list the variations, or to make them as ordered. With semiconductors it is a different story and I have found that you can talk yourself to death trying to get across the point that there is no excuse for making a part that is not a winner. It is, indeed, grossly irresponsible if you know in advance.

Win-Ratios

What sort of numbers are we talking about in the majority of analog manufacturers? The product win-ratio is between 10 and 25% with the spread depending on the type of products, the nature of the company, and the nature of the managers looking after the product groups. I have known lower numbers and I have known a few higher (the latter being from focused operations in very limited fields.)

Let's say that the average win-ratio is 20%. What does it mean to the company? What is the cost of the 80% of non-winners? A simplistic statement would be that 80% of your engineering, manufacturing and marketing/sales costs have been wasted, but that isn't the case when so many of the losers are caused by unnecessarily-manufactured derivatives. In the excuses game it is usually claimed that these extra costs are negligible and "go" decisions are, I think, often made in a real belief that it is so.

So, it's not 80% of costs that are wasted, and it's not 0%. And in my conversations with many companies I would guess that few, if any, know what the number is, although it is probably measurable. But whatever it is there is not only the actual loss but also the lost potential: What could the human efforts have been doing in working on other, profitable, projects? Totally immeasurable.

I don't think that I have been in any analog semiconductor company which says it has enough designers. Would they have enough if the product hit ratio was higher? Probably not, because every company has an enormous library of things they would like to get into, but at least they would have started a lot more of those!

How To Decide?

I feel that we not doing the requisite marketing research on the analog side of our business. If you do the research and the person authorizing sign-off for proceeding is not involved in the discussions you will have the best basis for going ahead. Those projects that are not recommended should not be offered for a second chance to the sign-off authority; that person should be presented only with those projects that are recommended. Handing over the file that contains the product specification, the price, business plan and forecast should not be an opportunity to open the discussion again; if the file doesn't speak for itself the process is faulty. Nor should the company demand a certain number of product introductions per annum, otherwise it will be guaranteeing a proportion of second-class products.

The analog manufacturing business is quite a profitable area. Some of our vendors have done extraordinarily well and it is my belief that they could do even better. Whenever I see one of the companies change their project selection process it seems to always be with the best intentions. But those intentions either quickly disappear or get folded out as other politics take over again and things get back to "normal." I think, in particular, that many senior managers do not realize the power they have when throwing out casual comments or suggestions, which come back as fully-fledged projects.

Our best analog companies started because the founders had the vision to see what could be done. After a few years of nurturing the operation to a reasonable size it is unlikely that one, or a couple, of people can see the whole picture. There is absolutely a time when you use the people you hire: And there should be no penalty for disagreement, ever.

If I have even partly converted you let me know.

By: Paul McGoldrick
Sr. Technology Editor, EDTN

Analog Main | Product of the Week | Columns | Editorial | Tech Notes

		Click here to get your listing up.