It's Easy to Design for "a" Customer
by Paul McGoldrick
In previous lives I have been on the receiving end of proposed product specifications from my sales managers to consider for the engineering department to design and bring to manufacture. Every company has limited design capacity and the more complex the task the more resources need to be thrown at the work - and there is less design effort going into other projects - so great care obviously needs to be taken at those early decision points.
I've never believed that making final decisions on the use of resources should be limited by a single company department. However sure a marketing group is that a particular product can be sold at a price that meets company profit goals, it cannot tell engineering that it can be done within a required time frame, nor even that it can be done at all with the facilities and people available. Nor can engineering double-guess marketing decisions, although we have a couple of vendors in the analog world that very much rely on FAE inputs directly to engineering to determine product directions - and that mostly turns out OK.
I've worked in companies where engineering said no to virtually every project that came along (I later took over the group after Mr. No was finally canned) and I've worked at places where marketing got nearly everything it wanted - or at least some, sometimes twisted, version of everything it wanted - because engineering couldn't say no. But by far the major danger is when you push through a product because it apparently satisfies the needs of "a" customer.
Having a single customer identified at the start of a design project is really not a very good, or very smart, place to put yourself and my experience has been that a proposal that comes from sales is nearly always dubious. I still keep my nose tuned for such projects, from my different position as a journalist, and I see a few a year get into production as finalized ICs.
There was one a couple of weeks ago. A major vendor wanted to have a telephone briefing on a product that has now been released. I acquiesced (which I do on maybe a quarter of the requests I receive - not because the products don't interest me but usually because I don't need somebody else to explain the obvious to me.) But the day before the briefing I received the embargoed materials and data sheet. It was a one-horse product…
Apart from the fact that the customer this product was designed for is obviously very well aware of its existence - and doesn't need my input on top of his own - the use is so limited that I could not think of anyone else who would want the too-precise specifications that were allowed for. I cancelled the briefing but the company came back asking for it to go ahead so they could hear my inputs. My response was it was not my job to do their marketing for them. Maybe a little curt, but fact.
There are so many things that manufacturers could be spending their resources achieving. There will always be a small percentage of single-customer projects that squeak past the check stages. But in this case of this particular vendor I have seen complete groups collapse after management decisions withdrew key resources from projects that would have been very profitable, even after having started the distribution of first-generation silicon to an eager market.
I don't miss those previous lives and the hardest sell of all: Within
your own company.