Goal: Retention of a diverse, on-demand workforce for crowd-sourced data projects that cannot be automated, such as image moderation, sentiment analysis, and categorization.
Challenge: CrowdFlower specializes in crowdsourcing for business data collection and other discrete tasks that require human input on a massive, micro scale. To achieve quality and speed, they need to motivate millions of on-demand workers worldwide to produce quality work.
Solution: Add game mechanics to the crowdsourcing application to reward accuracy and experience with higher paying opportunities.
Step 1: Gamify the CrowdFlower dashboard used by contributors to complete a task.
Step 2: Create badges and rewards for task completion and accuracy (determined by the crowd).
Step 3: Use completion data to create a “reputation” for each worker.
Step 4: Segment opportunities based on reputation levels.
Results: Average user accuracy increased 8.2% in the first week and continues to improve, tracking toward a goal of 90% accuracy.