Historically, the memory industry's business model has been production-based and one-size-fits-all. Companies adopt a factory-driven strategy of a limited product portfolio, aggressively striving to drive down manufacturing costs . The theory is that the lowest-cost producer will win in the long run, and all that success means churning out the narrow product offering relentlessly.

But that "recipe for success" is not customer-oriented. In today's memory marketplace, a real customer-oriented strategy requires companies to develop and produce the multiple technologies that customers actually need to support their varied products and business lines. Success comes through offering multiple solutions, top-quality product and cutting-edge technology.

This approach requires unparalleled engineering strength and immense manufacturing capability, as well as capacity. That combination allows a company to switch production quickly among the products in its portfolio to meet customers' actual demands and avoid depending on any single product. For example, now that the SDRAM market has cratered, companies need to offer products in the pockets of opportunity that remain strong, such as Rambus and double-data-rate DRAM, as well as asynchronous and other SRAM.

This approach requires significant investment in research and development, capacity and human resources. It also means investing continuously through the market cycles inherent to the memory industry. A company must be positioned to mass-produce next-generation products, such as advanced DRAM technology and DDR II. It must aggressively move to finer process geometries.

With the correct strategy, semiconductor memory remains an attractive long-term business. Companies can be well-positioned to capitalize on the coming rebound in the memory sector by maintaining favorable cash flow, managing capital expenditures, and having technological and manufacturing advantages.

When it comes to semiconductors, one size does not fit all. Successful players need to focus on the big picture, and on being a total memory provider. They must be committed for the long run. They need to recognize market cycles as a part of business, think beyond the downturn and continue investing in chip R&D. They cannot afford to be distracted by fleeting tangential business opportunities.

The key to prospering in the memory business is to offer all memory products and to offer them better than anybody else. Any other strategy and you're toast.

Tom Quinn is Vice President of Marketing at Samsung Semiconductor Inc. (San Jose, Calif.).