This article discusses new research in how the brain makes buying decisions and other choices — what is now called “neuromarketing”. Neuromarketing researchers seek to discover, and influence, the neurological forces at work inside the mind of potential customers. According to the article, most decisions are made subconsciously and are not necessarily rational at all – in fact they may be primarily governed by emotions and other more subtle cognitive factors such as identity and sense of self. For example, when studied under a functional MRI, the reward centers of brains of subjects who were given “The Pepsi Challenge” lit up when they tasted Pepsi, but Coke actually lit up the parts of the brain responsible for “sense of self” — a much deeper response. In other words, the Coke brand is somehow connected to deeper neurological structures than Pepsi.
Neuromarketing is interesting — it’s actually something I’ve been thinking about on my own in an entirely different context. What I am interested in is the question of “What makes people decide that a given meme is ‘hot’?” Each of us is immersed in a sea of memes — we are literally bombarded with thousands or even millions of ideas, brands, products and other news every day — But how do we decide which ones are “important,” “cool,” and “hot?” What causes the human brain to pick out certain of these memes at the expense of the others? In other words, how do we differentiate signal from noise, and how do we rank memetic signals in terms of their relative “importance?” Below I discuss some new ideas about how memes are perceived and ranked by the human brain.
Let’s call an individual instance of a meme at a particular time and place an “occurance.” An occurance might be a mention of the meme in the media, or in a direct or overheard conversation, or an advertisement, etc. What’s interesting is that the occurance patterns of memes do not have the same space-time dynamics. For example, some memes — such as mentions of common nouns — behave pretty much like “noise”. They occur in our experience with random distributions. Other memes, such as a steadily growing or declining trend, have linear dynamics — the number of occurances per unit of time is either constant or gains or declines in frequency with fairly predictable dynamics. Then there are nonlinear memes that behave chaotically — they exhibit erratic growth, sudden inflections, and are hard to predict.
Memes with random dynamics are quickly tuned out by the human perceptual system and by the brain. Memes with linear dynamics are treated differently by the human brain depending on whether they are unchanging, gaining or declining in frequency. Unchanging frequency is quickly tuned out as “background” noise. Memes that linearly gain or decline in frequency are attenuated as signal in proportion to their slope: The steeper the slope of frequency change, the more they are regarded as “signal.” Memes with nonlinear dynamics however are perhaps the most potentially interesting to the human nervous system — this is because it appears that humans are tuned to attend to “novelty” above all else. Nonlinear dynamics tend to be more novel than random or linear dynamics because they are hard to predict — the trajectories of nonlinear memes are “full of surprises” — they can suddenly change frequency and inflect. But not all nonlinear dynamics are equally interesting to humans — at the extreme of nonlinearity they are just random noise.
So the “ideal” dynamics is somewhere between chaos and order. What trajectories are optimally balanced between chaos and order? If we can figure this out, this is the key to a whole new science of marketing and advertising. Why is that? Because for example, it would enable us to mathematically optimize the frequency of an ad campaign in order to spend less money while actually generating better results in the minds of the target audience.
What’s interesting about this is that the way most advertising and marketing campaigns are conducted may in fact be almost maximally non-optimal. For example, showing an ad with regular, random or linear frequency over space and time in a market will quickly desensitize subjects — their brains will start to tune the ad out as “noise.” In other words, the brain attenuates to ad campaigns over time. So to get more attention, advertisers and marketers should vary their ad campaigns to make them more novel, more interesting, and less like “noise” or “background.” But how should they vary them, on what schedule, and in what manner, in order to generate the maximum effect? This is the interesting question to explore. What advertising distributions are perceived as most “interesting,” “memorable,” “positive,” and “important” to the human brain?
I believe it is possible to empirically test for the memetic dynamics that are MOST effective at getting attention in noisy perceptual environments by humans. By doing this we can, in a double-blind manner, figure out just what patterns work best. An experiment of this nature could be conducted in the following manner:
1. Create a mosiac of 100 frames, each capable of displaying an image
2. At each step in time (where a step is an intervale of time of fixed length, for example .1 seconds), show an image in every frame of the mosaic.
3. The particular image that is shown in any given frame at any step in time is determined by a computer.
4. At the start of the experiment the computer assigns a different “trajectory” to each of a set of 100 images. A trajectory can include frequency distribution over time as well as x/y coordinate trajectory over time.
5. Randomly select a set of test subjects to watch the slideshow for 1 hour.
6. After the slideshow, ask the subjects to rank the images they saw in order of “importance”
7. Two weeks later, without showing the subjects the slideshow again, ask the subjects to rank the images again in order of importance to see which images are retained as important for longer.
8. Test to see whether images that were shown with particular trajectories are consistently ranked as more or less important by subjects.
9. Run the experiment many times with different sets of subjects and many different candidate trajectories.
10. For each experimental run, the computer can use the trajectories that scored best in previous runs in order to narrow in on the best trajectories.
Over time this experiment might yield a set of trajectories that are most effective in getting and retaining attention in noisy perceptual environments. This would be a very useful discovery — it could be applied in many fields to overcome “information overload.” It would not only be of use to advertisers and marketers, but also to lobbyists, analysts in homeland security and intelligence, pilots in noisy cockpit environments, and even to help knowledge workers manage their email, prioritize search results, and monitor news more effectively.