The seven inputs
- Daily volume by content type
- The share AI resolves automatically for each type
- Throughput: items a moderator can decide per hour for each type
- Occupancy: the share of paid time spent actually reviewing
- Coverage: business hours, 16 hours a day or 24/7
- Languages: each one needs its own people on every shift
- SLA: tight report SLAs need buffer staff
Step 1. Items that need a person
Multiply each content type's daily volume by the share AI cannot decide. Text comments might need people for a minority of items, profiles for a majority, appeals for all of them.
Step 2. Moderator hours per day
Divide the human items for each type by throughput per hour, then divide by occupancy. Occupancy is never 100 percent: training, QA feedback, breaks and calibration take real time.
Step 3. People per shift
Spread the hours across the hours you cover. Round up per language, because a moderator who reads only English cannot work the Spanish queue, and every covered language needs at least one person on every shift.
Step 4. People on the payroll
Positions per shift are not people. Covering 7 days a week, holidays and sick leave means more people than seats. Then add team leads and QA reviewers on top.
A worked calculation
Take a community with 50,000 comments a day in English only, 24/7 coverage, and a one hour SLA. These are assumptions to replace with your own measurements.
| Input | Value |
|---|---|
| Daily comments | 50,000 |
| Share AI resolves | 70 percent |
| Comments that need a person | 15,000 |
| Decisions per moderator hour | 300 |
| Occupancy | 75 percent |
| Moderator hours per day | about 67 |
| Average moderators on duty, 24 hours | about 3 |
Three seats on every shift around the clock, plus a multiplier for weekends and leave, gives a payroll of roughly a dozen moderators, plus at least one team lead and one QA reviewer. Add a second language and the minimum grows by at least one seat per shift, even if that language has little volume.
What changes the result most
Languages and coverage usually move the number more than volume does. A small platform in five languages with 24/7 coverage needs a bigger team than a large one in one language during business hours.
Run it for your platform
The plan builder applies the same method to your content mix, languages, SLA and coverage, and shows the team per shift and a monthly estimate. See also how to hire content moderators and the outsourced content moderation comparison.