On a spreadsheet, hiring your own annotators looks cheaper than a platform. The spreadsheet is usually wrong, because it only counts the hourly rate.
What the spreadsheet misses
Recruiting and verifying experts, getting worker classification right across jurisdictions, running compliant payments, and building quality and provenance tooling. Each is a project, and each pulls senior engineers and ops people away from the model.
The senior-time tax
The most expensive line item is invisible: the attention of your best people spent on data operations instead of research. That is the cost that decides most buy-vs-build calls once teams are honest about it.
The costs that never make the spreadsheet
Beyond recruiting and pay, an in-house team carries: verification and vetting, worker classification across jurisdictions, payment rails and tax handling, guideline and QA tooling, provenance infrastructure, and coverage for churn and holidays. Individually small, together they are a standing team and a maintenance burden.
There is also flexibility cost. An in-house team is a fixed capacity: painful to scale up for a push and awkward to scale down when a project ends.
How to run the comparison honestly
Do not compare a platform's price to your annotators' hourly rate. Compare it to your fully-loaded cost: pay, plus the ops headcount, tooling, compliance, and the fraction of senior time consumed. When you cost it that way, the gap usually narrows or reverses, especially across multiple domains.
Skip the tax
Pathwize gives you verified experts, clean classification and built-in provenance so your team stays on the model. Book a demo to compare against your estimate.