What did Shannon actually say?
He defined information as the amount of surprise in a message.
In 1948 Claude Shannon showed that the information in a message depends on how unlikely it is. A message you could have predicted word for word carries nothing new, however well it is written. The less expected a message, the more it tells you.
Neuroscience arrived at the mirror image. Karl Friston's predictive-processing work describes the brain as a prediction machine that spends most of its effort reducing surprise. Attention goes to what breaks a prediction; comfort comes from what confirms one.
What does it mean for marketing?
Attention needs surprise; preference needs predictability.
Category-standard copy is perfectly predictable, so it carries no information and earns no attention. That is why the tenth agency promising to be data-driven is invisible. One sharp, specific, checkable claim carries more than a page of adjectives.
Once a buyer has noticed you, surprise stops helping. From then on they are asking a different question: can I predict what this company will be like to work with? Every answer that matches the last one makes you easier to choose.
If your competitor could run your ad with their logo on it, it carries no information.
What do we do differently?
We plan one surprise, then a long run of confirmations.
- One idea per campaign, specific enough to be wrong.
- The same promise, in the same words, on the ad, the landing page and the proposal.
- A publishing rhythm buyers can predict, so presence becomes a signal of its own.
- Novelty saved for the first contact, not sprinkled over every touch.
How we measure it
Surprise shows up as a step change in click-through rate on cold audiences. Confirmation shows up later, in branded search and in the share of deals that arrive already convinced.
The Brand Memory Planner sets the cadence: it fits a forgetting curve to your audience so the confirmations land before the message decays.