What the work involves

You write the problems, then you judge the answers. A typical batch might ask you to build a scenario — a regional wholesaler carrying 4,000 SKUs facing a margin squeeze from a new key-account rebate structure, say — and specify what a correct analysis must contain: the reorder-point math, the true cost of the rebate after freight allowances, the merchandising trade-off, and the reasoning that separates a competent answer from a plausible-sounding one. Then you review model responses against that rubric and write feedback explaining precisely where the reasoning failed, not just that it did.

Subject matter spans inventory management and turn, pricing and markdown strategy, planogram and category decisions, buyer/vendor negotiation, sales forecasting, loss prevention and shrink, and customer service escalation. Channel breadth matters: brick-and-mortar, e-commerce, and B2B distribution behave differently, and questions that only hold true in one of them get flagged in review.

What the platform screens for

  • Operator specificity. Screens push for numbers you actually worked with — turn rates, GMROI, shrink percentages, minimum order quantities, payment terms. Vocabulary without figures reads as coursework.
  • Defensible answer keys. Retail questions often have several acceptable answers. You need to explain what makes a response right and where legitimate judgment ends.
  • Written precision. Almost everything you produce is prose read by strangers. Sloppy or ambiguous scenario wording is the most common reason submissions get rejected.
  • Calibration under disagreement. Reviewers are asked to hold a grade when a model argues back convincingly — and to change it when the model is genuinely right.

Logistics

Fully remote and asynchronous, contractor engagement, hourly billing. Contributors typically pick up work in batches and set their own hours; there is no fixed shift, though projects run in waves and availability during an active wave matters more than total hours pledged. Onboarding usually includes a paid or unpaid calibration task before volume work opens. The $75–115/hr band reflects observed rates and varies with depth of category experience and review responsibility — it is not a guarantee.