How the Sevendays recommendation system works


At Sevendays we want to make sure that you can find the freelancers that fit your needs without the hassle of interviewing and reviewing applications.


The way we're solving this problem is by doing the work up front for you, by keeping track of the freelancers have the skills you need, but also by trying to determine who will the best chance at making your gig a success.


We use machine learning to help us learn which freelancers will best suit you. This is a complex process but it can basically be divided into two activities:


  • Subjective matching: We look at the ratings that you've given to freelancers in the past, and compare your rating profile with that of other clients. This helps us find the freelancers who you'll likely have a successful collaboration with. Of course, the more we know about you, the better this works. This is also the reason why you should always be honest when you give feedback to freelancers. When you give a good rating to a freelancer, we assume that this means you want to work with similar people more often.

  • Proficiency scores: We generate a rating called a proficiency level related to each skill a freelancer has advertised. These proficiency levels are based on the ratings they received from clients in past gigs, and takes into account the clients' rating habits, when the gigs took place, how consistently high the ratings were, etc. These proficiency levels enable us to estimate how good freelancers actually are in the skills they advertise. Based on these levels, we can assign the very best freelancers to the gigs that pay the most. This is in turn keeps the best freelancers motivated to remain active on the site.

We are constantly improving our algorithms as we hope to continually offer a better service to our clients and freelancers.