What the work involves

You'll spend most of your time producing reference-quality artifacts and grading model attempts at the same tasks. A typical batch might ask you to build a bottoms-up TAM model for a vertical SaaS category with sourced assumptions and a sensitivity tab, then review three AI-generated versions of a competitor battlecard and explain — in writing, at the cell and slide level — where each one invents a funding round, misreads a pricing page, or buries the actual differentiator on slide 14. Other days it's a win/loss synthesis from raw interview notes, an industry benchmarking report, or a feature comparison matrix where the columns have to be genuinely comparable rather than cosmetically aligned.

The critique half matters as much as the creation half. Ethos is training a model on judgment, not just formatting, so your rationale needs to be legible to someone who wasn't in the room: which assumption is load-bearing, why a top-down market size sourced from a single analyst press release is unusable, what a CRO would actually ask when handed this deck.

What the platform screens for

  • Verifiable depth. Four-plus years of market research, competitive intelligence, or benchmarking work in an agency, consulting, or in-house strategy seat — and the ability to walk through specific models or battlecards you built and what decision they fed.
  • Craft across all three surfaces. Docs, spreadsheets, and slides. Strong Excel modeling with weak deck construction, or vice versa, tends to surface quickly in sample tasks.
  • Sourcing discipline. How you handle thin data, contradictory secondary sources, and the difference between an estimate you'd defend and one you'd flag.
  • Written precision. Your critiques are the training signal; vague feedback has no value.

Logistics

Fully remote and asynchronous, 5–20 hours per week with room to take more if you want it. No fixed meetings; work lands as batched tasks with deadlines rather than live hours. The interview loop includes an AI voice screen that probes your background conversationally and follows up on specifics, so have two or three concrete projects loaded before you start. The $70/hour figure is the rate observed on this listing and can vary by task type, batch, and reviewer tier — treat it as reported, not promised.