Case Study / Creative Operations Turnaround

The Situation

A creative engine that had outgrown its own processes.

The brand was scaling fast on paid social, but its creative production had grown organically and hit a ceiling. Briefs were piling up faster than they could be edited, turnaround was slow, per-asset costs were climbing, and there was no reliable picture of how much the team was actually producing.

A distributed international freelance editing team was running without shared processes, clear incentives, or a real onboarding path, and the founders had no data to steer by. I was brought in on a fractional basis to take ownership of the editing function and turn production around.

The Work

Five parallel workstreams, run at once.

Compensation

Rebuilt the pay model around output.

Moved the editing team from flat retainers to a transparent per-batch model: tiered rates by format complexity, a bonus for clean first-pass work, and performance-linked bonuses for ads that won in-market. Pay now tracked directly to volume and quality, while per-asset cost dropped by roughly 44 percent.

Why a retainer costs more per asset

cost per asset

complexsimplefewer finished, every asset costs more

finished assets per month, more to the right

Flat retainer: a fixed amount divided by finished assets. When work jams, every asset gets more expensive.Rate per batch: the price follows the format tier, two tiers shown as an example, not the volume.Bonuses are added only for a clean first pass or an in-market win.Shaded: where the operation sat before the switch.

MeasuredCost per asset about 44% lowerafter the switch to per-batch rates

Schematic without values; the figure in the box is measured.

Visibility

Built a data backbone from scratch.

Stood up the team's first end-to-end production tracking, including an automated daily recap covering output versus intake, per-editor throughput, backlog and cycle time, plus a forecasting method built on real intake rate, run rate and editor capacity. For the first time, leadership could decide from data instead of guesswork.

How the daily numbers fit together

Intakenew briefs per day

Backlogopen briefs; cycle time roughly equals backlog divided by output

Outputfinished briefs per day, equal to the number of editors times per-editor throughput

this is where the curve turns

IntakeOutputBacklog

weeks

If throughput drops for everyone, it is the process; if it drops for one person, it is that person.

Forecast: intake rate against run rate times editor capacity, that is, against the output of the coming weeks. Will the queue grow?

Schematic, no values: the logic behind the daily recap. Below: while intake runs above output, the backlog grows; after that it shrinks. That is how the backlog turned within the first weeks.

Diagnosis

Fixed the system, not the people.

Separated structural bottlenecks, jammed review stages, voiceover routing errors, incomplete briefs and tracking gaps, from genuine performance issues, and resolved them at the source. A 16-point quality checklist cut revision rounds and standardized output before delivery.

Four bottlenecks, solved at the root

BriefIncomplete briefs
AssignmentVoiceover routing errors
Edit16-point checklist, before the cut is delivered
ReviewJammed review stagesRework goes back to the edit; the checklist shortens this loop
Ready to launch
Rework goes back to the edit; the checklist shortens this loop

Tracking gaps, across the whole chain

bottleneckadded check, not a board status

The five statuses of the board at the time, each bottleneck in its place; the diamond is the check that was added. None of the bottlenecks sat in the edit itself.

Team

Led a distributed international team.

Managed around nine editors plus voiceover specialists across multiple time zones: onboarding, capacity planning, routing each format to the right strengths, and daily hands-on support, while holding firm but humane standards through real-world disruptions without losing throughput.

Stakeholders

Translated between speed and cost.

Sat between founders pulling for raw speed and others pushing cost discipline, anchored every decision in data, and replaced ad-hoc pressure with clear forecasts and joint decision-making.

How the forecast turns speed vs. cost into a decision

Two demands: more speed (founders) and holding costs (others)

Forecast: the weekly gap between intake and outputCapacity: around nine editors plus voiceover. Same arithmetic as the daily numbers above.

Add capacityfaster, more spend per week, same rate per asset

Prioritizesame cost, some work waits

Change nothingqueue and cycle time grow

A joint decision instead of ad-hoc pressure

Schematic: the options the forecast makes visible and what each one costs.

The Outcome

A chaotic operation became a measurable one.

Within the first weeks, the department had a new compensation model, a working data and reporting system, a standardized quality process, and a backlog that had turned the corner from growing to shrinking, with weekly output climbing toward target. An organically grown operation had become a measurable, process-driven function with the structures in place to keep improving.

What was in place after the first weeks

Built (3)new pay model: rate per batch plus two bonusesdata and reporting system: daily recap and forecaststandardized quality process: 16-point checklist

Turned (2)backlog: from growing to shrinkingoutput: rising week by week toward target

Measured (1)cost per asset about 44% lower

Three structures built, two curves turned, one figure measured.

I rebuilt a stalled ad operation into a measured, lower-cost production engine.