Corporate manipulation may be classified into two, the consumer manipulations and manipulating other corporations. Manipulating consumers is relatively easy if you’re a large corporation. You have dedicated staff, a lot of resources, and you know a lot more about what you’re doing than the typical consumer. Manipulating other corporations, however, would seem to be a lot tougher – after all, they’ve got the same assets you do.
It turns out it’s not particularly hard to manipulate other companies – as long as you have a dominant market share.
Let’s take some examples. Take Intel, for example. Intel makes the chips that go into most PCs, and as such it is a colossus – worth over $100B today. It dominates its industry – more than 80% of the chips that go into PCs come from Intel. You would think that a company with that sort of position would have little need for manipulation – and you’d be wrong.
In 1995, for example, Intel was behind its main competitor, AMD, who was making better, cheaper chips. Intel doesn’t sell to consumers – it sells to computer makers, who then sell to consumers. As more and more of these manufacturers defected to AMD, Intel found itself under pressure. The key manufacturer, at the time, was Dell. It became urgent for Intel to make sure that Dell would buy its chips, and, more importantly, that Dell would not buy AMDs. So Intel basically cut Dell a $1B cheque. That was sufficient for Dell to ignore AMD’s better chips, and continue to buy Intel products. As the CEO of Intel, Mr. Otellini, wrote in an email, Dell was “the best friend money can buy.”
The cheque itself came in the form of ‘rebates’. Basically, Intel would give out special rebates to Dell on Dell’s purchases of Intel products, timed to coincide with Dell’s quarterly result announcements. As a result, Dell consistently could show strong quarterly results, and Intel kept AMD out of the largest manufacturer for a long time.If Dell bought AMD chips, Intel would withhold those payments.
The tactic was apparently repeated – and often. In 2008, the European Commission fined Intel $1.44B for making payments to Dell, HP, Lenovo, and Acer meant to delay or refrain from using AMD chips. S. Korea and the US have launched their own investigation.
Despite the ‘rebates’, AMD continued to make headway in the manufacturer ranks. So Intel upped the ante. As per this FTC formal complaint (warning: PDF) it started a ‘bad cop, good cop’ strategy: for manufacturers that bought its chips and excluded AMD, it promised guaranteed supply of chips; it offered to pay to help the ‘good guys’ win bids against AMD products; it offered indemnification in terms of patent lawsuits, and more.
As the ‘bad cop’, Intel dealt harshly with those who did dare to buy AMD chips. It raised the prices on its chips (and other products) for those manufacturers who also bought AMD chips. It shut down technology sharing and partnerships with those players (a big blow in such a deep technical field). It reduced supply to those same players, essentially putting them at a disadvantage (since AMD didn’t sell a full product line, most manufacturers wanted some AMD chips but also needed some Intel chips).
These were all moderately successful tactics – they slowed AMD significantly. Intel also did some other, less obvious maneuvers to slow them down further. It changed its compiler and software library to make it run not more effectively on Intel hardware, but less efficiently on AMD hardware. This was particularly devious, because users who saw a performance gap between the two platforms would naturally assume that the AMD hardware was to blame – who would think that the software itself would penalize one platform over another?
We could keep going with Intel, since they seem to be doing all of this all over again in the graphics arena, But let’s pick on someone else for variety.
How about Rambus? Rambus was a relatively low-profile company until 1999 – it made connections for memory chips, which was not usually considered a cutthroat industry until then. In 1997, the various companies responsible for making memory chips got together to agree on the next standard for making memory modules (called SDRAM). The companies worked together and eventually published the new standard, and manufacturers started to incorporate the design into their machines. You should know, at this point, that coming up with a new memory standard is tough – hundreds of companies have to collaborate so that chips made in a factory in Singapore will talk to memory modules made in Japan and work with the software written in California. This is not the kind of effort you do every year, or every every ten years. So after hundreds of hours of work, the committee responsible for the standard disbanded, with the satisfaction of a job well done.
As products started to actually roll off the plants, a couple of years later, Rambus sent out letters to every memory manufacturer, mentioning that they owned a patent on the critical module that the memory standard specified, and that all PCs were likely to incorporate for the next several years.
It was a nice coup. Rambus had advocated their technology to the committee, but when told that they would need to essentially sell their patent “under reasonable terms” if the technology was adopted, they withdrew from the committee. But they also decided to tailor their patents towards the standard. Once the standard was adopted, Rambus was in a great position: it threatened to sue anyone who didn’t pay their licensing fee. Several manufacturers paid the license fee, not having the stomach for a prolonged legal battle. Others, like Micron and Hynix, fought in courts, and, in 2009, a federal judge ruled that Rambus had a clear show of bad faith (partly due some destruction of evidence), and could not sue Micron.
Nevertheless, Rambus made hundreds of million of dollars by manipulating the timing of its patent applications (don’t feel too bad for Infinions and Hynix and others ‘victims’, by the way – they were found guilty of price-fixing memory chips in 2004. Ironically enough, it was the FTC who discovered the price-fixing while they were looking for evidence to convict Rambus). The EU launched an investigation into Rambus, calling what they did “patent ambush”.
These are great manipulations: not overtly illegal (when the EU has to come up with a new name for what you did to sue you, you are creative at the very least), very effective, and very profitable. They are different from corporate manipulation of consumers, mind you – they are not gradual, they involve a lot of stealth, and they rely more on understanding and surfing the ‘grey’ areas of the law.
This is par for corporate – to – corporate manipulations. It turns out that just having similar resources does not mean that companies can’t manipulate each other – they just have to leverage their position in interesting ways, and hope that what they do doesn’t attract the attention of the regulators (although in most cases, the fines the regulators impose are less significant than the benefits of the manipulation, if it works).