A interesting new startup, Affective Interfaces, gave a presentation today at the TechCrunch50 conference. They’re not exactly the slickest of salesmen but what they’re pitching has enormous potential. Basically, they’ve created a piece of software that can read emotion on a human face. I’d imagine it’s not the most nuanced algorithm yet, but it seems to be a good start.
Advertising and market testing appear to be their primary focus, and with good reason. Market research is a huge business, but the technology that drives it is still pretty basic. Focus groups and surveys seems to be the standard. It’s the “take a look and tell me what you think” approach. However, surveys require a level of self-consciousness on the part of the person being tested. People can be embarrassed by their responses and lie with ease.
The last big technological jump in market research came from aggregating the demographics of customer loyalty cards. You probably have a few in your wallet or on your keychain right now. Who can deny those savings?!
But why would a company give you a discount unless you were giving them something of value in return? That would be bad business. In reality you’re giving them something of immense value: information. Rather than asking you what you think you would buy, loyalty programs watch what you actually bought. Walmart knows how much you hurricane survivors love pop-tarts.
Along the same lines, this software takes that more immediate and unconscious reaction, the one you can’t lie about, and applies it to the media we’re experiencing at the time. (i.e. Do BMW ads make middle-age males the most happy when the see the car or the logo?)
Charting the emotional response to advertisements is a good first step towards understanding the cognitive link between the media we consume and our emotions, but in terms of creating effective ads there’s still the pesky little problem of delivery. Ask yourself, how many television ads have you watched in the past week? How many do you think you saw five years ago, ten years ago? We’ve gotten awfully good at avoiding advertising, even has it has become increasingly prevalent.
Taking this a step beyond advertising, think about what this software means towards entertainment. Whatever you’re watching could potentially react to how you’re feeling at the moment. Bored with a show? The TV will anticipate you want to change the channel. Storyline uninteresting? A game might change on the fly to create more excitement.
Film test screenings would know moment by moment where a film was working, where it lagged, where it made you laugh or cry. This is a hugely important tool to studios trying to understand how much money to pump into reshoots and marketing. However, often they’re terribly inaccurate.
Even more interesting will be linking this type of technology to evolutionary algorithms, but that deserves its own post. Needless to say, finding a way to not only monitor but engage with our emotions in an immediate way could create some of the most effective media yet to be experienced, and also the most seductively terrifying.