How to Build a Media Buying Team: Roles, Skills, and Structure

Build the team around work that must be completed and reviewed. Ad spend alone does not specify headcount: one market with reused assets differs from many languages, accounts and new shoots. Assign ownership for execution, creative and measurement before choosing job titles.

A workload-based staffing calculation

The following fictional week uses assumed hours, not a staffing benchmark or NovaStorm productivity result. Measure your own queue and include review, leave and rework.

If one full-time position supplies 30 productive queue hours after meetings and routine administration, 104 ÷ 30 = 3.47 full-time equivalents of capacity. Four positions provide 120 such hours, leaving 16 for variability in this illustrative plan. Skills can be combined or contracted; this does not prescribe four people for a particular spend tier. Additional leave or required coverage may change the staffing choice.

At a hypothetical fully loaded $50 per productive hour, 104 hours cost $5,200 per week. Compare a hire’s actual compensation, onboarding, supervision and output with a contract’s cost and review needs. Choose the next resource from the persistent bottleneck, not an unsupported claim that one role always has the highest return.

Responsibility and approval matrix

Role labels do not guarantee that a given platform offers exactly those permissions. Verify the available account roles and grant only the access needed for assigned work. Keep approval records, version IDs and an absence backup. Separate campaign permission from customer-data and payment access.

Training gates instead of a three-month promise

1. Foundation gate: explain outcome, currency, attribution window, contribution ceiling and new-customer denominator. Demonstrate a reconciled teaching dataset with duplicates and refunds.
2. Draft gate: prepare a paused campaign and creative brief; another reviewer verifies price, rights, placement crop, destination and intended outcome. No live spend authority yet.
3. Measurement gate: write a falsifiable hypothesis, assignment method, cap, maturation period and stop/incident rule. Identify an inconclusive result without inventing a winner.
4. Supervised delivery gate: a budget owner approves a bounded live task and access. The trainee records delivery, checks and escalation without exceeding the allowance.
5. Independent-task gate: a qualified reviewer signs off repeated correct behavior and incident response. Increase authority only after evidence, not after an arbitrary number of weeks.

This is a proposed curriculum; no claim is made that three juniors were trained in three months. Mentoring consumes senior capacity and belongs in the workload model. Track accepted drafts, missed defects, accurate reconciliation and adherence to allowance alongside commercial outcomes; do not reward rapid budget changes or creative volume alone.

Make preparation reviewable

A proposed NovaStorm workflow is for the buyer to prepare paused campaign and local ad-set drafts from the creative lead’s approved brief. The analyst supplies an external verified outcome summary and finance supplies the cap; a human owner decides release. Treat any routine-time saving as a hypothesis to measure with comparable tasks, including correction and review time.

Use the creative approval workflow for version ownership and the campaign naming guide for traceable handoffs. A weekly queue review can identify blocked work, but mature outcome reviews follow the actual conversion and refund delay.