This article explains Upwork’s use of placebo auctions to test the effectiveness of Boosted Proposals. In these randomly selected job posts, boosts are visible to freelancers but not actually applied—no extra Connects are taken, and proposals are ranked organically. The goal is to gather unbiased data to improve the feature’s accuracy and usefulness without affecting real job opportunities. - AI helped create this summary
Boosted Proposal placebo auctions are a step we take to measure the success of the Boosted Proposal feature.
Now, a very small percentage of job posts may have placebo auctions for Boosted Proposals. Here’s what that means:
- Freelancers can still see the placebo auction and submit boosts, but no Connects will be taken for the boost. (Only the Connects needed to apply to the job will be taken.)
- The job posts that have placebo auctions are still real and clients will still receive all proposals submitted. The proposals will just all be ranked organically instead of appearing at the top of the client’s stack, and none of them will have the “Boosted” label.
- Freelancers won’t know the difference between placebo actions and normal Boosted Proposal auctions until after they submit their proposal. After they submit their proposal, they will be notified if the auction was a placebo.
- Only a very small percentage of auctions are placebo. Whether a job’s auction is standard or placebo is entirely random and not impacted by a client’s spend, job post characteristics, etc.
- The placebo auction testing is only being used to allow us to accurately measure data and improve this feature.
Our testing of Boosted Proposals revealed that the feature can be helpful. Boosting your proposal can increase your chance of being hired up to 24%. This placebo feature allows us to monitor the average benefits users can expect from boosting on an ongoing basis, helping us build and improve its usefulness.