Andreas Weigend | Social Data Revolution | Fall 2014
School of Information | University of California at Berkeley | INFO 290A-03


Audio: weigend_ischool2014_4.mp3 (dropbox)
Transcript: weigend_ischool2014_4.docx (not completed yet)

What is the future of work? Can social data transform work? How to make data driven decisions to optimize and decompose work?
To answer all these questions,Odysseas Tsatalos and Panagiotis Papadimitriou will join us in class and
and discuss how Elance-oDesk revolutionized the online workplace.

Timeline Oct 28, 2014 (Part 1)

3:30 Getting ready
3:40 BEGIN
4:30 END
Odysseas Tsatalos, CTO and Founder of oDesk
Panagiotis Papadimitriou, Head of Data Science at oDesk

What is Elance-oDesk?

It is the largest online workplace. Their goal is to put data at the heart of the company and build data products rather than just using data to guide products; and their mission "to create economic and social value on a global scale by providing a trusted workplace to connect, collaborate, and succeed.”

Who uses Elance-oDesk?

It serves businesses of all sizes and allows them to hire freelancers globally.
For start-ups, small and mid-size companies, they are the fastest and cheapest staffing solution available. Companies like Nextdoor hire and manage programmers, designers, customer service agents, content creators and moderators, market researchers all through Elance-oDesk. A growing number of clients use the platform as a freelancer management system to manage their existing contractors. Mozilla has moved over 30 of their engineers to Elance-oDesk. Bigger companies like J&J are building private talent clouds to access specialized talent that can be shared across teams and departments on demand.

What are the problems they trying to solve?

  • Freelancers
    • Identify Great Talent
    • Predict Quality and Expertise
    • Infer Availability
  • Clients
    • Understand Hiring Needs
    • Predict life-time value
  • Match
    • Clients and Freelancers for Great outcomes
  • Fight Fraud
  • Data-driven Policy Making

How did they solve those problems using data?

  • Matching Freelancers with clients: oDesk have all the data they need about the freelancers. They have their emails, resumes, applications, preferences. They also have all their interactions with employers and feedback the employers leave. All this data can be used to match talents with jobs and predict quality and expertise
  • Understand hiring needs: There is a difficulty when it comes to writing concise job descriptions. The employers sometimes describe the job in terms of outcome especially if they are posting in different language than their native language. oDesk tries to find a match who speak the language of employers which makes the communication easier.

Did they really revolutionize the hiring process? Show us data!

When Elance-oDesk started, most jobs posted were IT jobs so most candidates where in IT business as well. However, with time they witnessed a change in types of jobs posted. The change was slow but eventually the IT jobs became only about half of the posted jobs not all of them.
external image rKAS08v4f5FQdpgNuNDQJOjtiKx_I5o-jTdd6JmUiOPnlPqy3kMoU7L7GQm-RgG5vSp_ZlgBCkVXVdjYah-ZKtMItyWGiAKtmhp2NJoKnKO8toDgeNUpi5VGKwsxbC1MUA
From graph of current job postings, we can see many categories and how they are growing. What is fascinating is that this grow was done without active push from oDesk, as a result, it reflects the growing interest in many areas of doing work online. Because of new growing categories, they are starting active category management to nurture early categories & make them grow, e.g., data science (collaboration with IBM).
Now they have 2.4M job posting annually, 22 candidate per job, and in 24 hours about 25% of jobs are filled.

Questions for thought?

  1. What is the equations for business?
  2. Can we achieve transparency in workplace?
    • -Because there are so many different standards of what "transparency" includes according to that respective business and their goals, I don't think that having a standardized idea of what "transparency" means is really appropriate. For example, micromanaging may not be the style of one business and may not place any value on that particular idea of transparency. They may not want to know every detail accomplished throughout the day, and rather would prefer to focus on the bigger picture.
      • --- Noelle Reyes
  3. Can we use social data collected from Facebook and LinkedIn to find perfect employee?
    • Probably not the absolutely perfect employee, but the information from those sites (as well as possibly other social media sites) can definitely help you make informed decisions when it comes to hiring.
    • Social media is just one facet of how people choose to represent themselves online. It is by no means an indicator of their professional personality. For example, someone's Instagram could have no pictures relating to their field of expertise, and another person could have all their posts related to their profession. This difference in how people choose to present themselves in a casual social setting designed to be shared with friends, not recruiters, does not reflect how they will perform as an employee. I believe social media can be used as a precautionary filter, but it does not showcase enough (or any) of that person's skill level.
  4. Do you see oDesk as a platform to change the real world jobs not only online jobs?
    • Yes - the methodology that oDesk provides in order to choose freelancers is very similar to the shift that the world is undergoing today. Many job applications are now through LinkedIn, which provides testimonials and proficiencies very similar to that of an oDesk profile.

This page initially created by (3 students total):
Hala Diab ( ),
Farin Namdari ( ),
Kamel Dabwan ( )