Phonetic friction: why customers shorten your name

Stress placement, consonant clusters and the three-syllable ceiling.

Customers are ruthless editors. If a name takes effort to say, they cut it down to something that does not, often within a month of launch, and they do not ask permission. The shortened version then becomes what people search for, what they type into an address bar and what they call you when you are not in the room. Phonetic friction is the effort a name costs to say, and we screen for it because it predicts what customers will do to the name.

Stress placement

English speakers bring strong expectations about where the weight of a word falls. When a name puts its stress where listeners expect it, they repeat it without thinking. When the stress is ambiguous, they choose, and they do not all choose the same way. Take an invented example, Corvida. Some listeners will say COR-vi-da, others cor-VI-da. Neither is wrong, and that is the problem: the name now exists in two versions, and each one gets repeated by the people who heard it that way.

Consonant clusters

Clusters are runs of consonants with no vowel between them. English handles them reasonably well at the start of a word and far less well where two parts of a compound meet. A name like Brandstrom, used here only as an example, asks the mouth to move through five consonants in a row. Almost nobody does. They say Branstrom, and before long the easier spelling starts turning up in emails and searches. Clusters are also the first thing to collapse in a noisy room, which is exactly where our phonetic screen tests a name.

The three-syllable ceiling

In our experience, three syllables is about as much as a name can carry before customers start trimming it. Four-syllable names tend to become two, usually the first pair or the stressed pair. That is not always a disaster. A shortening you planned for can be owned: you can hold the dot and the handles for it before launch, and use it in your own copy so the clipped form is still yours.

So when a candidate runs long, we ask two questions before anyone falls in love with it. What will customers cut it to, and is that cut a name we could also own? If the answer to the second question is no, the long version rarely survives the shortlist. It does not matter how well it reads on the page if the version people actually say points somewhere else.