Can Big Data + Big Dating = True Love?

Guest author Mat Young is senior director of products at Fusion-io,

In recent years, big data and technology have infiltrated nearly every part of our lives – and dating is no different. Today, approximately 20% of relationships begin through some sort of online or mobile dating service. It’s estimated that there are 90 million single people over the age of 18 in the United States. According to Match.com, 40 million of them online dating services – more than 40% of the country’s single population. If those numbers don’t seem big enough, China is the world leader in online dating, with more than 140 million users looking for a match online.

(See also Top Dating Apps For iPhone, iPad and Android)

The unsung hero of online love is the infrastructure that supports these dating applications. As more people turn to online and mobile dating, databases containing profile info, photos and messages between users continue to grow as well. The scale means that innovative tools are needed – along with more powerful infrastructure – to keep matching up potential pairs.

Cupid’s Changing Industry

It’s estimated that in 2010, users spent more than twice the amount of time connecting to dating websites than they did using mobile dating apps. Just one year later, in 2011, these numbers flipped, as users began spending slightly more time on dating apps than websites. Whatever platform users prefer, online dating is booming as services use technology to match up people with shared interests and values.

Today, the number of people using dating apps is growing faster than all other apps combined. Of note is that the average age of the online dater is 48 – proving that Baby Boomers are more connected than ever. This coincides with general trends showing more adults than ever are connecting via mobile devices – the same devices they are using for mobile dating.

These well-connected users have helped propel the online dating industry to top $1 billion in annual revenues, providing online dating websites with even more reason to continue searching for innovative solutions.

Of course, this kind of growth adds more stress to mobile and wireless networks, not to mention the servers hosting the applications – a welcome challenge for online dating providers.

Match.com found out how important scalability was when it began struggling to keep up with a growing user base and its 70 terabytes of data. With an upgraded SQL Server solution from Microsoft, Match.com was able to move user data across more than 100 servers in two seconds, while also providing sufficient infrastructure for future growth.

Big Data Meet Online Dating

While many users may not see, or care, how data is crunched to help them find the perfect mate, database admins, Big Data scientists and mathematicians all play an important role in formulating the perfect match.

When a user signs up on eHarmony, for example, they are required to fill out a 400-question profile outlining personal preferences, physical traits, hobbies and many other telling details. Using that data, an Oracle 10G database makes a few suggestions for possible matches.

Then the real magic happens.

The system compares the users extensive list of answers to the site’s 20 million other users – a process that requires a billion calculations for each bachelor or bachelorette. After comparing information using a sophisticated algorithm, the site provides the user with matches. eHarmony’s dataset alone exceeds 4 terabytes of total data, not including photos and other information.

These fine-tuned formulas and powerful datacenters are the trade secret of these online dating services. As users change preferences, provide feedback on proposed matches and refine their profiles, eHarmony uses that information to refine search results for proposed matches.

Optimizing For True Love

Just like with any database, the quality of output is based on data input. Users who accurately answer questions and provide truthful information are likely to receive the best love matches.

Considering the human element, though, perfect matches venture far beyond math and science. Even if every user tried to be completeley honest in their responses, personality matches might not be completely accurate. People often perceive themselves differently than how the rest of the world sees them.

Some observers look for an objective system of understanding potential compatibility, while others believe a completely automated matchmaking service that removes human input could never work.

Most likely, the next innovations in compatibility science will be come through a technological love triangle, where advanced hardware and software solutions in enterprise and hyperscale datacenters will enable development of unique new applications that can help real people find true love more efficiently.

(See also Taking My Dating Mobile: A Social Experiment)