Algorithms – helping you find love, furry friends and star employees

Algorithms – helping you find love, furry friends and star employees

 

Algorithms seem to be popping up everywhere today. And while the term probably elicits thoughts of professors and computers whizzes engaged in highly complex calculations, the fact is that algorithms are commonplace among many of the activities we do every day.

 

Put simply, an algorithm is a process or set of rules to help solve a problem or perform a task, traditionally carried out by a computer. The beauty of algorithms is that they can analyse extremely large amounts of data to help us make sense of information, draw conclusions, and make effective decisions (that would otherwise be nearly impossible to do).

 

Activities such as creating music playlists, searching for love, choosing a pet, conducting a Google search and navigating the quickest route from A to B are just a few examples of how algorithms are changing the way we experience everyday life. And with the rise of big data in recruitment and selection, it’s not surprising that these types of systems and processes are becoming increasingly popular to help make better hiring decisions.

 

 

Finding love via eHarmony

 

Algorithms may not be the first thing that spring to mind when you think of love. But for dating website, eHarmony, the two are closely related. eHarmony is one of the largest online dating sites globally. Their matchmaking success is said to come from a sophisticated algorithm used to pair potential soul mates.

 

Researchers and former physicists working for eHarmony are said to use methods inspired by quantum physics to create the matchmaking algorithm. Individuals complete an online relationship questionnaire and are paired with others in eHarmony’s expansive database, based on factors such as levels of agreeableness and extraversion, spirituality, levels of optimism, and preference for emotional intimacy.

 

And apparently the brain does know what the heart wants, because according to a 2012 study the divorce rate of married couples who met on eHarmony was approximately 50% less than those who met via other avenues.

 

However, even if the outcome of online matchmaking isn’t entirely successful for you, there is also an algorithm that can help with predicting the success of couple’s therapy by analysing a variety of voice qualities!

 

 

Finding the perfect pet via PawsLikeMe

 

Not in the market for a hook-up via eHarmony? Perhaps a furry friend is more up your alley. Lucky for you, there is a company that will help select the most suitable companion for you. PawsLikeMe, a US based company, created the first of its kind human-to-pet matching algorithm that is said to have a success rate of 90% in terms of predicting compatibility.

 

Originally developed by animal behaviour expert Coleen Johnston and social worker Marianna Benko, the algorithm uses the results of a personality assessment based on four traits that are said to influence the human-animal bond. In addition to personality, the matching algorithm is also said to take into account lifestyle factors and home environment when generating a score. An added bonus is that all dogs listed on the website are from shelters or rescue groups.

 

 

Finding the perfect music via Spotify

 

How algorithms can be applied to a field as subjective as music is pretty amazing. But Spotify’s Discover Weekly does just that, with seemingly great success.

 

Created by software engineer Edward Newett, Spotify’s Discover Weekly algorithm connects data from over 2 billion user playlists with your own profile to suggest new music likely to get your feet tapping.

 

More specifically, the algorithm analyses other users’ playlists based on songs you have listened to and recommends other songs in that playlist (on the assumption you have a similar taste in music to the creator of the playlist). The algorithm also uses your taste profile (very specific genres of music you have a preference for) to make targeted suggestions for new tunes to listen to. Discovering your new favourite songs made easy!

 

 

Finding faces via Facebook

 

Facebook uses numerous algorithms to help with functions such as targeted advertising and what to display in the news feed that many of us see multiple times a day (if you ask me, the algorithm could do with some tweaking!).

 

But face recognition is another area where algorithms are ‘enhancing’ the user experience. Can’t be bothered searching for your friends to tag them in a photo? Not to worry, Facebook’s algorithm will automatically identify people in your photos approximately 83% of the time (even if their face is obscured). The algorithm uses cues such as facial features (distance between eyes, nose, mouth), clothing, hair, and body type to determine who’s who in the photo. The set of rules they have to identify individuals is based on profile pictures and photos they have been tagged in previously.

 

 

Finding the best people – recruitment and selection

 

Steve Levy, the director of global sourcing job site Indeed, sums up nicely why algorithms are useful in recruitment and selection. He says algorithms are used “to take massive amounts of data being generated before, during and after the recruiting process and turn it into actionable information—with one goal being to predict whether a person will be right for the job, the team and the company”.

 

LinkedIn uses algorithms to help recruiters sort through millions of online profiles to identify people they might want to hire. They do this by analysing information such as who recruiters have engaged with via e-mail, what candidate suggestions they have ignored, and even by looking at geographical trends of which industries and roles tend to see candidates relocate.

 

Gradit, a new graduate recruitment jobsite started by Joe Afif, assists with hiring graduate talent in the tech sector. Students create a profile, and organisations post jobs on the site. Algorithms match the two, in turn reducing the amount of unsuitable jobs/profiles to trawl through. Joe explains: “If a job description looks for X Y and Z, and the candidate has X and Y then the candidate will receive 2/3 and therefore a 67% fit for the job. Of course, this is oversimplified. There are many more variables and they’re scored a lot more intelligently, but you get the picture.”

 

Psychometric testing is another area in which algorithms help turn data into user-friendly outputs.  And more recently, game-based assessments have gained interest as an innovative way to reliably assess candidates’ suitability for a role. Game-based testing differs from traditional testing in the amount of data captured during a test sitting. For example, Revelian’s Cognify tracks players’ actions (e.g., hesitation, mouse clicks, quality of decisions) every few milliseconds, in turn capturing thousands of data points per session.  These data points are then combined into algorithms that help make sense of the information in order to measure job-relevant traits.

 

While there are many ways we can use algorithms during recruitment and selection, the most important factor to consider is whether both the information gathered, and the output generated, is relevant, valid and reliable for the decisions you’ll make. As with any recruitment and selection tool, it’s always a good idea to use it as one piece of information to consider alongside all of the other information you have about a candidate.

 

Then again, perhaps one of the most attractive features of such algorithms is their removal of bias from the assessment process: they take a purely objective, scientific and equitable approach to recruitment, minus the inescapable biases we all have that influence our thinking despite our best efforts. And according to one study, algorithms were much more effective at determining the best candidates for low-skill service sector jobs than humans.

 

I think we have a way to go until algorithms could take over all aspects of recruitment across all roles and besides, as humans we do have to interact and work with each other. Nonetheless, there are some interesting movements happening in this space that we’ll be watching with interest.

 

 

Kate CervettoAbout the Author

 

Kate Cervetto joined our Client Services team way back in 2006, while completing her Masters in Organisational Psychology. Shortly afterwards, she moved to the Psychology team and started helping our clients to select the right person for the job and build more satisfied, productive and committed teams.

 

Kate has worked with a broad range of clients and businesses, from large-scale graduate programs, to consulting with individual managers on how to develop future leaders. Willing to take on any HR challenge – no matter how big or small – with compassion, enthusiasm and humour, Kate is highly respected and valued by both our clients and the rest of the Revelian team.