VELOS offers two data quality options: single or multi-contributor
The integrated workforce in Hivemind VELOS is made up of a large number of contributors that belong to a comprehensively managed workforce. This means each contributor has been vetted for skill and character, is offered opportunities for development, and is paid by the hour—not the number of tasks they perform—so they’re incentivised to spend more time on the trickier cases.
Within this framework, VELOS offers you two options for data quality, as explained below.
Option one: each piece of your data is handled by an individual contributor
With one contributor, each row is completed once by a fully vetted, trained and motivated managed workforce.
For simpler text annotation projects where the classification categories are intuitive and unambiguous, a single contributor’s responses may be ideal given the scale, cost, and accuracy requirements of your project.
If your classification categories contain bespoke business logic, or the content is potentially ambiguous, or accuracy is paramount for your project, then it may be more suitable to take a multi-contributor approach.
Option two: each piece of your data is handled by three independent contributors
With the multiple-contributor option, each row of your data is distributed to three independent contributors. VELOS’s text annotation consensus algorithm compares every annotation made by each contributor, returning clean annotations where at least two contributors agree.
This approach maximises the quality of your data by minimising inevitable human error and cutting through ambiguity by aggregating multiple independent decisions.
The difference in cost between option one and option two
There is a cost difference between single and multiple-contributor options because of the additional work involved. The cost of each option will be displayed openly on the screen. Simply toggle between the one contributor and the three contributor options on the screen to see the cost differentials appear.