Intel Inside: how a part nobody sees became a reason to buy

In early 1991 Intel had a naming problem. Its processors were sold by number, the 286, the 386, and rivals such as AMD were making compatible chips. In March 1991 a federal judge ruled that “386” was a generic term Intel could not protect as a trademark, so a competitor could put the same number on its own chip. Meanwhile the people paying for computers saw a brand on the case, IBM or Compaq or Dell, and nothing of what made the machine fast.

Intel’s answer, led by marketing executive Dennis Carter, was a cooperative advertising programme called Intel Inside, launched in 1991. A share of what PC makers spent on Intel chips went into a marketing fund, and Intel used it to subsidize their advertising on one condition: the ads, and later the machines, had to carry the Intel Inside logo. Many manufacturers joined. In 1993 the next processor line arrived with a name instead of a number, Pentium, and in 1994, as Intel moved into television, it added a short five-note sound by the composer Walter Werzowa that played with the logo.

Proving what you cannot see

A processor is the clearest case of a benefit the customer cannot perceive. Nobody opens the case, and few buyers could compare clock speeds with any confidence. Our books are blunt about this: a benefit that is not perceived, strategically, does not exist. Proof is the mechanism that makes quality perceivable, and when the product is invisible the proof has to be designed from scratch.

Intel designed it in the places buyers already looked. A sticker on the case turned a component inside a sealed box into something a shopper could point at. The logo in the PC maker’s own print ads repeated it, and the sound did the same job on television. The signature carried almost no technical information, and repeated often enough it began to read as a guarantee.

Renting the customer’s shelf

The cleverness was in the channel. Intel did not need its own stores or its own shelf space. It paid to appear inside the communication of the companies that sold to the end user, which put its name on the ads of many brands at once. The PC makers got cheaper advertising, and Intel got the most valuable placement available: inside the decision itself, on the product being bought.

The relationship had a cost on the other side. A buyer who asks for a computer with Intel Inside has already made part of the choice before the PC brand speaks, and the processor brand takes some of the meaning that used to belong to the box maker. Over time the makers became, in part, the channel for someone else’s promise. For a supplier the trade is excellent. For a manufacturer it is worth reading the terms before signing them.

Visibility works in both directions

In 1994 a mathematics professor, Thomas Nicely, found that early Pentium chips could return wrong results in some floating-point divisions. Intel at first described the flaw as rare and offered replacements only to users who could show they needed them. Public pressure grew, and in December 1994 Intel agreed to replace the chip for anyone who asked. In January 1995 it took a pre-tax charge of 475 million dollars.

A flaw in an anonymous part would have been a problem for engineers. A flaw in a branded part became a story for the evening news, because Intel had spent years teaching consumers that this chip was the reason to trust the machine. The books say that trust takes the minimum: ten proofs and one public refutation leave a wound. The lesson Intel drew, replacing on request, matches what the promise had implied all along. Once you put your name on a hidden component, the customer holds you to it as if they had bought it from you directly.

Ingredient branding is open to any company that makes something inside someone else’s product: a fibre, a material, an algorithm, a sensor. The work is to choose a visible signature, a channel that carries it into the moment of purchase, and a policy for the day it fails. If you want to build that kind of proof with us, see what we do and where we would start.