Upwork is a web based platform that provides remote work. It connects freelancers with clients. They have about 300 employees and 1,000 freelancers. Upwork provides the service for individuals and enterprise. Upwork uses algorithm and machine learning to connect lancers with clients. At Upwork they have freelancers actually working with them in the company so they are using their own product.




At Upwork they value the ideas that come from both internal and external sources. From the outside both their clients and the freelancers provide feedback from their perspectives. In addition to the feedback from the user’s, individual contributions come from within the company as well as their stakeholders in order to answer the question of if the idea satisfies their goals or not. In addition to this, Upwork must make sure that they have enough time and resources to devote to the execution of the new idea because they can’t afford to waste resources like some of the bigger companies such as LinkedIn or Microsoft.



Upwork is very big on dogfooding (using your own platform as a service). By being a large customer of itself they are well aware of the pain points and teams are designed to make the platform stronger through an iterative process. The idea being that ‘new engineers are trying to put maintainers out of a job eventually’.  Upwork does the most by the book version of Agile, by practicing Scrum in 2 week sprints. Every new feature is built around testing and metrics.

For Upwork, as a data-driven company, the data will show and prove the success or failure for each project and the company as a whole. Furthermore, these data is divided into certain metrics for which some are seen as more important for the measure of success. The main metrics used when measuring for success at Upwork are:

  • Freelancer earnings (Measuring how successful people are around the world, quality of the work through the platform provided by Upwork)
  • Number of freelancers/hirings/sign ups – the number of users all around the platform
  • User feedbacks



Since the platform connects freelancers with projects/companies, the main focus on metrics are around the earnings, number of people on the platform and feedback from all users. Each project also includes several A/B testing to evaluate each project.

The decision to exit a project if successful vs failed follows the same strategy as for the metric of success – it is all data-driven based decisions. The decision to exit a project is then evaluated against a success-matrix from which a score will tell whether the project should continue or not. Also, the evaluation of a project is most of the part done on team level, where each team responsible for a particular project has to make a decision, based on data, whether it will be a successful implementation or not.