Online dating site eHarmony gathers enormous amounts of data. It has over 20 million current registered users, each of which fill out 258 questions based on 29 “dimensions of compatibility.” That data is then run through predictive algorithms to assess the long-term compatibility between each of the members.
If you took only 1 million people, 500,000 men and 500,000 women, this would result in 124,999,750,000 billion potential matches. That’s a lot of dates.
In advertising, a single digital advertising campaign can generate 62 unidecillion (6.2×10^37) possible trafficking combinations. If you were doing one combination a second, it would take 3.18 years to develop a campaign.
Luckily, for advertisers and daters, there are systems that process all of this information, and there are extremely brilliant analysts to make sense of it all and distill the possibilities into the most actionable insights.
But, counterintuitively, finding and presenting people with a perfect match — whether it be a life partner or a new car — isn’t actually the best way to close a deal.
Too much of a good thing is a BAD thing. We often think we know what we want and only want to see examples of such things. But research shows that we are more satisfied, and make more and better decisions, when we are given options of somewhat less relevant choices. Other studies have shown people become overwhelmed and desensitized to many possibilities, pushing off the actual choice into the future (i.e. never).
The brain has evolved to be highly critical in processing information, and when presented with options that are similar, we will work to find a fault even if all the choices available are individually desirable. Spreading out the options lets people feel freedom to make decisions without making a “bad” one.
Consider an online dating scenario where a man has indicated he prefers women who run and are tall. Say he then receives 10 sequential profiles of females who are taller than 5 foot 6 inches and wrote in their bio they have participated in a marathon. Getting those profiles will likely decrease his interest in the matches and make him less likely to look at more matches in the future (confidence in future relevancy).
Why? He received too much of a good thing and didn’t have enough match diversity to instigate action. This is more fondly referred to today as “analysis paralysis.”
The same behavioral psychology manifests in advertising. If you know a particular user is in the market for a new car, showing that user only and too many auto ads on auto sites will likely make them quickly tune out the strikingly similar messages from different auto advertisers – and could go so far as to have the converse affect the advertiser wished their ad to instigate. Worst case scenario, too many of the same ad could delay the user from making a purchase at all because they feel overwhelmed by too many choices or — worse — annoyed.
Presenting consumers with the right balance and quantity of ad messages across a diverse media landscape is likely to influence positive interaction and reduce analysis paralysis. During the year 2000, Sheena Iyengar of Columbia University and Mark Lepper of Stanford University conducted a study that found when participants were faced with a smaller rather than larger array of chocolates, six versus 24 options, they were actually more satisfied with their tasting, proving quality opposed to quantity of diverse options is the key to understanding preferences, making better decisions, and understanding context effects.
Patrick Giordani is Adconion Direct’s head of Business Analytics and Optimization. Patrick spent six years as the Head of Matching for eHarmony (which has generated 600,000 marriages since 2005). Now Patrick matches people with the right ads, acknowledging the similarities between the science and psychology of analyzing big data for the world of online dating and digital advertising.
VentureBeat and marketing technology analyst David Raab are working on a new Marketing Automation usage and ROI study
. If you currently use a marketing automation system, help us out by answering the survey.
If you do, we'll share the resulting data with you.