Build a Weekly Optimization Loop for Your Live Campaign
For most sellers, campaign optimization isn't a routine, it's a one-time panic move. The numbers look bad, you panic, and you change five things at once: the image, the copy, the budget, the audience. Whether the result improves or gets worse, you'll never know WHICH change actually did it, because you moved too many variables at the same time. The real issue isn't improving the campaign, it's turning that improvement into a MEASURABLE routine. This prompt builds you a checklist to open on the same day every week, the discipline to change only one variable at a time, and a change log that tracks what you changed and when. Three weeks later, you'll have a real answer to 'which decisions actually improved this campaign,' backed by evidence instead of guesswork.
You are a performance marketing consultant who optimizes live ad campaigns systematically. Your job isn't handing me a one-time improvement list, it's building a weekly checklist I'll repeat, a single-variable discipline, and a change log system that lets me see over time what actually works. This isn't an audit checklist. It's about building a habit I repeat week after week. Use the following information: - Platform: [Meta, Google Ads, TikTok Ads, etc.] - Campaign goal: [sales, leads, traffic, etc.] - Weeks live: [number] - Current performance summary: [brief, e.g. ROAS is inconsistent, CPA is climbing] - Levers I have access to: [budget, creative, audience, bid strategy; which ones I can actually touch] - Time I can give this campaign per week: [minutes/hours, realistically] RETURN THE OUTPUT IN EXACTLY THIS ORDER, WITH THESE EXACT HEADERS: ═══ WEEKLY CHECKLIST ═══ A concrete 5-7 item checklist I'll run through in order on the same day every week (example: has frequency crossed the threshold, which ad is the worst performer, can the best performer's budget be raised). Each item states in one sentence what to look for, and the whole list has to actually fit inside the time budget I gave you. ═══ THE ONE-VARIABLE RULE ═══ Explain why I shouldn't change more than one thing at a time in this weekly cycle. Give a simple decision rule that allows only one change per cycle (example: find the single biggest problem metric first, then take exactly one action tied to it). If there's more than one problem metric, tell me which one to prioritize. ═══ CHANGE LOG TEMPLATE ═══ Give me a simple log format to fill out every week (week number, the one change made, that week's key number, observed result). Make it copy-paste ready, a table or a short list format works. ═══ THE EARLY-DECISION TRAP ═══ Tell me how long I need to wait to see the result of a change, and put a number on how many days of data is misleading to decide on. Explain why making a second change before that window passes contaminates the read on the first one. ═══ WHAT TO DO AFTER FOUR WEEKS ═══ Once four to six weeks of log entries have piled up, explain how I should read that data looking back, what patterns to look for (for example, seeing which type of change keeps working over and over). RULES: - No made-up industry benchmarks or fake "average improvement rate" numbers. - Make the routine genuinely usable: not a ten-item marathon list, a short list that can actually get done every week. - Do not loosen the single-variable discipline. Never suggest changing more than one thing at once, under any circumstance. - Do not build a routine that exceeds the weekly time budget I gave you. A routine that doesn't get followed is worthless. - Write in clear, systematic, repeatable language.
How to use it
- Give the basics of your live campaign (platform, goal, how many weeks it's been running, a quick performance summary). On first setup, this becomes the template for your weekly routine.
- Run the prompt and get your personalized weekly checklist plus the one-variable rule. Open that checklist on the same day every week (Monday, for example) and work through it in order, without skipping steps.
- Log the ONE change you made each week, plus that week's number, in the change log. After four or five weeks pile up, feed the log back into this same prompt and read together which changes actually worked.
Example / tip
Input: Meta campaign, sells handmade jewelry, live for 5 weeks, ROAS is inconsistent. AI output gives a weekly checklist in this order: check frequency, pause the worst-performing ad, gradually raise budget on the best performer. It tells you THIS week to pause only the one ad with high frequency and touch nothing else, then log it as 'week 6: ad X paused, reason frequency 6.2' in the change log.
Usage example
Copy the prompt as is, fill in the bracketed fields for your own business, and paste it into ChatGPT or Claude. The more concrete your input, the more useful the output. Take the first draft, ask for a one-sentence fix on whatever you don't like, then give it a final pass in your own voice.
When not to use it
This prompt is for campaigns that are already set up and have been collecting data for at least a few weeks. It shouldn't read as advice for early intervention on a campaign that's brand new and still in the learning phase. Let the data build up first.
Output quality checklist
- Is the weekly checklist concrete and sequential, or generic advice.
- Does every cycle really recommend changing only ONE variable.
- Is the change log format simple enough to fill out every single week.
