Campaign Optimization Operating System | Manage Tests, Decisions & Learning
Jack RighteousStop Optimizing Campaigns in Isolation.
A useful test can improve one campaign. A useful operating system makes sure the learning survives long enough to improve the next campaign too.
This Build system sits above the free Campaign Control System, Campaign Review & Optimization and Campaign Test & Experiment Planner. Use those tools to run and diagnose individual campaigns. Use this system to manage the portfolio of campaigns, decisions, experiments and reusable learning over time.
Operating rule
ONE RECORD PER CAMPAIGN. ONE OWNER PER DECISION. ONE ACTIVE TEST PER QUESTION. ONE LEARNING CARRIED FORWARD.
The goal is not to create more dashboards. It is to stop repeating the same uncertain decisions because last month's reasoning disappeared into notes, chats or analytics tabs.
Who this Build system serves
You run your own marketing and need a lightweight way to know what is live, what is being tested, what was learned and what deserves attention next.
You need campaign ownership, approvals, shared test priorities and a record of decisions that survives handoffs.
You need the same discipline across multiple clients while keeping client truth, approvals, accounts, data and reusable learning properly separated.
Prerequisites
You need actual work and evidence. This is not a hypothetical planning exercise.
You should be able to state what happened, what appears weak and what decision came from the evidence.
When you decide to test, use one primary variable and record the important constants.
A spreadsheet, project system, database or document is enough. The operating logic matters more than the software.
The operating cycle
Add every campaign that deserves active management.
Give it a current status so the portfolio is readable.
Identify the most useful unanswered question or weak point.
Put proposed experiments into one backlog instead of launching every idea.
Run the smallest useful controlled test for the question.
Compare new evidence with the original objective and hypothesis.
Record what changes, what stays and who owns the decision.
Translate the result into a contextual learning that can inform future briefs.
Remove stale tests, obsolete assumptions and dead campaigns from the active queue.
Tool 1 · Campaign Portfolio Register
Do not use the portfolio register as another analytics dashboard. Its job is operational visibility: what exists, where it stands and what decision comes next.
Campaign name: ________
Business objective: ________
Audience: ________
Offer / desired action: ________
Primary channel(s): ________
Owner: ________
Approver / client approver if applicable: ________
Start / evidence window: ________
Current status: PLANNED / LIVE / REVIEW / TESTING / PAUSED / RETIRED
Strongest current evidence: ________
Current decision or unanswered question: ________
Linked review / test record: ________
Next action: ________
Next decision point: ________
Reusable learning so far: ________
Tool 2 · Experiment Backlog
Ideas enter the backlog before they enter the campaign. This prevents every suggestion, AI output or stakeholder opinion from becoming an immediate live change.
Question to answer: ________
Linked campaign: ________
Evidence behind the question: ________
Proposed hypothesis: ________
Primary variable: ________
Primary evidence needed: ________
Expected impact: HIGH / MEDIUM / LOW
Confidence in the diagnosis: HIGH / MEDIUM / LOW
Effort: HIGH / MEDIUM / LOW
Risk / reversibility: ________
Prerequisites: ________
Owner: ________
Backlog priority: NOW / NEXT / LATER / DROP
Do not turn the scores into fake mathematics. High/Medium/Low is enough if it forces an explicit comparison. A test with high potential impact but poor evidence, high risk and high effort may deserve less priority than a small reversible test that answers an important question quickly.
Tool 3 · Active Test Queue
Work in progress is a constraint, not a badge of ambition. Too many simultaneous tests can exhaust production capacity and make the evidence harder to interpret.
Default to one major active experiment at a time unless the tests are clearly independent.
Multiple tests can run when ownership, campaign boundaries and evidence are distinct. Set a deliberate work-in-progress limit.
Separate the queue by client/account. Never let one client's data, approvals or learning be treated as automatically transferable to another.
Active test: ________
Question: ________
Owner: ________
Start: ________
Decision point: ________
Primary variable: ________
Important constants: ________
Primary evidence: ________
Stop / hold conditions: ________
Current state: READY / RUNNING / WAITING FOR EVIDENCE / REVIEW DUE
Tool 4 · Decision & Learning Log
A result becomes useful institutional or creator knowledge only after you record the decision and its context. “That headline won” is not enough.
Date: ________
Campaign / test: ________
What we observed: ________
Decision: ________
Evidence supporting it: ________
Confidence / uncertainty: ________
What changes now: ________
What remains unchanged: ________
Context where this learning applies: ________
Context where it may not apply: ________
Reusable rule or briefing note: ________
Review / expiration point: ________
A winner is not automatically a rule.
If a result worked for one audience, offer, platform, timing window and creative context, record those conditions. Carry the learning forward as a stronger starting assumption—not as universal truth.
Portfolio review: look for patterns, not stories
The same type of evidence appears across more than one relevant campaign or test. It may deserve promotion into a stronger briefing rule or operating preference.
An outcome happened once or under unusual conditions. Keep it documented, but do not force it into every future campaign.
Different campaigns point in different directions. Check the audience, offer, channel, timing and creative context before averaging the contradiction away.
A platform, product, audience, price, policy or market condition has materially changed. Mark the old learning for retest rather than quietly relying on it forever.
Monthly or cycle-based Optimization Summary
Period reviewed: ________
Campaigns live: ________
Campaigns reviewed: ________
Tests completed: ________
Most important evidence: ________
Strongest learning to carry forward: ________
Assumption that became weaker: ________
Pattern worth retesting: ________
Campaign / test to retire: ________
Highest-priority unanswered question: ________
Next active experiment(s): ________
Briefing rule or process change for the next cycle: ________
Use the system differently by operating mode
You may be strategist, producer, approver and analyst. Keep those decisions separate in sequence so a production impulse does not overwrite the evidence.
Every live campaign needs a responsible owner. Every material decision needs an approver. Every test needs a clearly assigned review point.
Add client/account, authoritative source material, client approver, data/tool restrictions and explicit reuse boundaries. A useful learning from one client is not permission to reuse that client's confidential information elsewhere.
How this connects to Campaign Repurposing
The Campaign Repurposing System remains useful when the problem is asset expansion: you have a strong project and need more purposeful content, promo and CTA-backed pieces from it.
This operating system solves a different problem: deciding which campaign question deserves attention, which test enters the queue, what the evidence changes and what learning should be reused later. Use repurposing as an execution resource when the portfolio decision calls for more or different assets—not as a substitute for diagnosis.
Completion task
Load three recent or active campaigns into the system. If you have only one real campaign, start with one rather than inventing two more.
- Create one Campaign Portfolio Register row for each real campaign.
- Add the unanswered questions and proposed experiments to one backlog.
- Select no more than the next one to three genuinely active experiments.
- Complete one Decision & Learning Log entry from existing evidence.
- Mark one stale, duplicate or low-value item PAUSED, RETIRED or DROP.
- Define the next portfolio review point.
Success criteria
- Every active campaign has a current status, owner and next decision.
- Tests answer explicit questions rather than random ideas.
- Duplicate experiments are visible before time is spent repeating them.
- Active work is limited enough to manage and interpret.
- Decisions record evidence and uncertainty.
- Learnings include the context in which they were observed.
- Old assumptions can be retired or retested.
- The next campaign can start from documented learning instead of memory.
Navigation
Supporting Build resource: Campaign Repurposing System when you specifically need to expand one core asset into additional campaign-ready content.
What this system does not promise
It improves the quality, continuity and auditability of campaign decisions. It does not create causal certainty, remove platform volatility, guarantee performance or make every lesson transferable to every audience, offer or channel.
Primary road: Find Your Brand · Build + Tool · Intermediate. Access depends on the applicable member, resource or implementation arrangement.
