Prolific selling managed data pipelines

This shift means Prolific is selling a managed data pipeline, not just access to workers. In the old marketplace model, a researcher picked a cohort, set pay, and Prolific took a visible fee on top. In frontier lab training, the lab buys an ongoing program where Prolific recruits the right people, pays them, checks the work, and runs the daily workflow, which makes revenue larger, less transactional, and closer to an outsourced training team.1 2 3

• The product changed with the customer. Prolific still offers self serve studies, but frontier labs are using API connections and managed services for red teaming, safety evals, cultural nuance testing, and post training work that need bespoke pools, oversight, and fast turnaround.1 2
• Bundling also changes what revenue means. In this category, headline revenue often includes contractor payouts and service delivery, so retainer style deals can make top line jump even when the actual software and service margin is much smaller. That is why net revenue and margins matter more than gross volume.1 4
• This puts Prolific between Mechanical Turk and BPO style vendors. MTurk offers cheap raw labor with little provenance, while firms like Invisible and Handshake package labor, QA, and workflow management for labs. Prolific is moving up that stack while trying to keep its advantage in verified, deeply profiled participants.1 5 6

The next step is more wallet share from fewer, bigger AI accounts. As model builders spend less on generic labeling and more on trust, safety, cultural fluency, and personality grounded evaluation, vendors that can run always on human data programs with reliable supply and QA should keep taking revenue away from simple marketplace models.1 3 5