I’ve been in the board industry for about 30 years. That probably makes me smarter than the average bear, but this thesis might prove otherwise, depending on how you see it as an engineer or a manager. This industry is undergoing some of the most difficult changes in its history. We are moving from parallel buses to serial fabrics. We have the tin/lead transition to Restriction of the use of certain Hazardous Substances (RoHS). We have a depressed telecom equipment industry. We have experienced a slowdown in MIL/COTS spending for electronics. We have Waste Electrical and Electronic Equipment (WEEE) on the horizon. Some market segments have become commoditized. We are experiencing a larger number of chips going End of Life (EOL), creating obsolescence problems across all market segments. So, let’s take a look and see if my “Logic Chain” theory is valid for you and your company.
Here is how I see it. As commoditization goes up, innovation goes down; as innovation goes down, differentiation goes down; as differentiation goes down, variants of standards go up (an ineffective form of differentiation); as variants of standards go up, margins go down (for example, the volumes are too low to capture any economies of scale, and engineering costs rise to do the custom engineering); and when margins go down, the number of patents filed on open standards goes up (an ineffective way to raise margins). Somewhere along this chain, we see some spur lines branching off. Those are mergers and acquisitions. Various mergers and acquisitions began in the industry at several junctures, mostly when differentiation went down, and later when margins went down. Let us discuss them one by one.
We have all seen the desktop chips and architectures commoditize many segments of the board business. That is a given, particularly in the telecom and industrial board segments.
Innovation goes down as an inverse function of commoditization going up. Back during the Great Telecom Revolution, every Pentium board announced that complied to certain board standards looked exactly alike. That is empirical evidence that commoditization occurred in certain segments. Therefore, in an effort to differentiate, all the product manufacturers got wrapped around the axle with Linux or Microsoft code and drivers. Those differentiators lasted about six months, and we were back to a purely commoditized market with undifferentiated products. Innovation continued to decline as companies integrated undifferentiated products with undifferentiated software.
As product differentiation went down, companies began to create variants of certain specifications. These variants allowed some companies to lock out competitors, albeit for a short period of time, and to lock up some customers. Different connectors, custom-routed backplanes, unique software (for example, managed switches), and special chassis metalwork were all examples of this trend.
These variants helped companies to differentiate their products through custom engineering for their few customers. However, those customers were still buying low volumes of these unique products. When you consider the additional engineering costs and the low-volume manufacturing of the products, costs tend to go up and margins tend to fall. Companies were willing to assume those higher costs and lower margins in anticipation of large orders from customers sometime in the future, as predicted by the market researchers. But those volumes never materialized, thereby exacerbating the margin problems. At this stage, companies making these low-margin/low-volume products in specific market segments start eating the bark off the trees in their parking lots in an effort to survive.
When margins are depressed on such products and the market researchers continue to report unrealistic forecasts for demand, companies must find another way to maintain their customers and thwart low-cost competitors. So, they patent their implementations of those standards-based products. Thus far, I have seen about 70 patents on serial fabric architectures covering mechanicals, backplane routing techniques, and the architectures themselves.
As I stated in this column, mergers and acquisitions occur within an industry at several points along the Logic Chain. At the beginning of market development, when market researchers proffer huge forecasts, companies will buy the smaller innovative companies who have a head start in the market with products. For example, Intel bought Ziatech back in 2000. As the market develops, companies will buy other companies to gain market share. In 2004, Motorola bought Force Computers.
As a market develops and defines itself as a low-margin/low-volume/undifferentiated-product market, some companies choose to diversify to protect their overall margins. They buy low-volume/high-margin specialty companies in other market segments. For example, in 2005, Mercury Computer Systems, Inc. bought Echotech. Schroff bought Electronic Solutions in 2005. GE Fanuc bought Condor Engineering, SBS Technologies, Inc. and Radstone in 2006. Diversification is just another method of protecting overall margins for a company that may be too exposed in a developing or declining low-margin/low-volume market segment. Such acquisitions promise some protection for the bark on the trees, depending on which way those low-volume/low-margin markets go.
Every company makes decisions based on its own needs and goals, but those needs and goals are primarily influenced by market conditions. The embedded board markets are very fragmented by application segment and product requirements for each segment. Some segments offer high growth potential but low volumes and margins at this stage (telecom). Other segments offer low volumes with high margins and moderate growth (MIL/COTS). Some segments offer very low growth rates, accompanied by severe competition and price/margin erosion (industrial). The Logic Chain I have proposed is causing many of the changes we see in our industry. Also, I believe this Logic Chain will continue to affect product decisions in our industry for many years to come, both on the user side and the supplier side.
For more information, contact Ray at [email protected].