In the age of internet dating, companies have made millions off users who rely on the 'secret sauce' algorithms that promise ultimate matchmaking abilities. With the rise of big data, dating websites quickly increased the amount of data they collected on its users. Questionnaire responses on likes/dislikes, personality traits and relationship goals formed part of the site’s growing data set. At the same time, the amount of users across the sites has increased significantly in the last 10 years, creating a large collection of data points. Through the development of machine learning algorithms, the online dating platforms should be able to predict with more certainty, matches that lead to ‘true love’.
This year, Match.com launched a survey on U.S. adults’ opinions on love and relationships – “Match’s Singles in America Survey”. (A way to get even more valuable data to use) With over 35,000 respondents, some of the results were quite surprising. If you were picking a restaurant food genre for a first date, would sushi seem like a good choice? Personally, I would have said no – raw fish, use of chopsticks, complicated menus – too many unknowns to think about when you’re already worried about making a good first impression. But yet, the survey data suggests that the sushi first date increases the chances of a second date by 170%! Worried about how to handle the check? According to the survey, splitting the bill is more permissible after a few dates (48%) versus 29% who approve of this on an official first date. All of that new data can now be used to help refine what the system uses to determine someone’s compatibility with another.
So what does it all mean? Users turn to those sites because they trust that the data used to power the site’s secret mix of algorithms and data science, will connect them with someone they wouldn’t have otherwise found. That trust is what feeds millions of dollars of revenue into the online platforms. But what if the data used to match individuals was bad? Romantic digital matches are dependent on the accuracy of the data being collected. Things like inconsistent fields, missing information or out of date data can be the difference between a future “I do” and a “Let’s just be friends”.
Here’s an example… Amy filled out a profile back in 2010 before the dating website had a field for whether or not she was looking for a partner who likes animals. She has 2 dogs and a cat, but the missing data field was not pushed to her as a mandatory update to her profile, so she is matched with someone who is highly allergic to pet dander and sneezes the entire night. That missing data led to wasted time, and ultimately a bad experience for Amy and her date.
Data quality, and a process to govern that data long-term, is imperative to trust the output or insights from any system. This could be a dating site, or it could be your organization who is making big decisions based on data being pulled in from multiple systems, enterprise-wide. Regardless of the purpose of the data, accuracy and reliability leads to trust. And we can all agree that a stronger level of trust leads to better business outcomes.