Campaign Workflow Automation
Key Insights — Weekly creative optimization that no one had time to do
I led the design of an internal workflow that turned campaign optimization from a manual task shared by every game title owner into a repeatable process run by one operator.
The Problem
21+ Hours of Manual Work
Optimization Kept
Getting Delayed
Unwritten Decision Rules
What I did
Workflow Analysis
Data-Driven Rules
System Design & Rollout
Business Outcomes
21+ hrs → 83 mins
Reduced weekly operating time by approximately 93%.
7 Owners → 1 Operator
Made weekly optimization a repeatable process.
Extended to Moloco
Applied the same approach to a second advertising network.
Background
A manual routine on the company's largest ad channel
Google Ads accounted for the largest share of Bitmango's ad spend. Each title owner managed campaigns for their games, but reviewing existing creatives was often postponed as new campaigns took priority.
The Problem
Everyone agreed the work mattered, but nobody did it
A full optimization cycle required over 21 hours across seven title owners. Each had to review campaign performance and decide which creatives to keep or replace, making an important task difficult to sustain alongside daily responsibilities.
WORKFLOW ANALYSIS
The rules existed, but not in a form we could use
I mapped the workflow with the UA and Creative teams and found that campaign data was scattered across tools, while optimization decisions depended on each owner's experience. The team knew what to do, but those decisions had never been documented.
Budgets also ranged from $500 to $100,000 per day. Different campaign objectives and levels of risk meant one set of thresholds could not work for every title.
PROJECT GOAL
Turn tacit judgment into rules a system can execute
The goal was to turn recurring optimization decisions into a process one operator could run consistently. We started with lower-spend titles to validate the approach before expanding to larger campaigns.
Workflow roadmap · Analysis, rule definition, system design, and rollout.
DEFINING THE RULES
From experience-based judgment to measurable thresholds
I worked with the team to identify which performance signals mattered, when action was needed, and what should happen next. I then analyzed campaign data in Python across three periods—one year, three months, and one month—to define thresholds by campaign type and budget.
These conditions became the system's decision logic, allowing the same rules to be applied consistently across campaigns.
Workflow DESIGN
The spreadsheet was the interface
Since marketers would operate the system, I designed it around a familiar spreadsheet rather than introducing another tool. Python scripts connected the inputs and execution trigger to the Google Ads API.
I defined the workflow, decision logic, and requirements while engineers implemented the system. Rollout began with lower-spend campaigns to limit risk.
Before and after · Seven title owners' weekly work consolidated into one operator's workflow.
OUTCOME
A 21-hour team task became an 83-minute job
One operator could review five ad groups and around 200 creatives in 83 minutes, reducing operating time by approximately 93%. More importantly, optimization became a consistent weekly process rather than a task repeatedly postponed.
The approach was later extended to Moloco, a second advertising network.
REFLECTION
The automation was the smaller half
The hardest part was not building the system, but making the team's existing judgment explicit and getting everyone to agree on the rules. Once those decisions were clear, we could build a workflow that one operator could run without having to make every decision from scratch.
Thanks for coming by.
© 2026 Hailey Moon