Lesson 11 of 11
Benchmarks, Prompts & 30-Day Plan
The reference kit and your launch sequence. The executive summary distills the whole system; the benchmarks give you targets; the prompt library gives you a running start; and the 30-day plan turns a course you read into a machine you built.
Executive Summary — The Whole System in One Page
Paid growth is a machine of four numbers: CPA, conversion rate, landing-page performance, and LTV. Your job is to build a qualifying funnel — cold traffic to a paid entry offer (the $5 quiz) that filters buyers and feeds the platforms a purchase signal — then capture the email, upsell at peak trust across three tracked touchpoints, and follow up non-buyers within the 90-day cookie window. Underneath it all sits conversion tracking: without verified events and real conversion values, every decision is a guess.
On Google (intent), you start with Search campaigns and walk the three-phase progression — Maximize Clicks to buy data, Maximize Conversion Value once you pass ~10–15 conversions, then Target ROAS to scale profitably — never skipping steps and never disturbing the learning phase. On Meta (audience), you feed the algorithm with pixel + Conversions API, seed it with a Custom Audience and 1% lookalike, and let retargeting convert the warm traffic; high-ticket offers warm to a call or event rather than closing cold. Creative is the last great lever — win in the first three seconds, test angles not adjectives, and keep a refresh pipeline because every winner fatigues. The landing page is never finished; you improve conversion rate one disciplined test at a time. You scale only on the machine's green light ("limited by budget" + stable CPA), gradually (vertical) then wide (horizontal), matching aggression to your payback speed. And you run the whole operation with an AI layer — Claude as analyst, copywriter, creative engine, and CRO strategist — on a daily-look, deliberate-change rhythm. It is all numbers, and numbers do not lie.
KPI Benchmark Cheat Sheet
Directional ranges to orient you — your real targets depend on your margins and market. Use these to spot "clearly broken" vs. "healthy."
| Metric | Healthy-ish range | Improve by |
|---|---|---|
| CTR (Meta cold) | ~1–2%+ | Better hook / angle / audience |
| CTR (Search) | ~3–8%+ (intent) | Tighter keyword-to-ad relevance |
| Landing conv. rate | ~2–5% (cold), higher for quiz | Message match, friction cuts, proof |
| CPA | Below conversion value / confirmed LTV | Raise CVR, lower CPC, better data |
| ROAS | Depends on margin (aim well above breakeven) | Raise AOV, lower CPA |
| Frequency | Watch above ~2–3 on cold | Refresh creative, widen audience |
| Retention (sub, 90d) | Know it before scaling | Onboarding, fast value, engagement |
The Rules — What You Never Do
- Don't change the landing page while ads run — unless it's one isolated change you're watching.
- Don't dramatically change the daily budget — every big change forces a relearn. Step ~20% and let it settle.
- Don't touch the campaign in the first 3–4 days — early erratic results are normal.
- Don't run Performance Max or AI Max before you have conversion data to verify it works.
- Don't guarantee specific results in month one — after a month of optimization, you can project.
- Don't send paid traffic to a free product when you need a real conversion signal.
- Don't scale a subscription on projected LTV — wait for confirmed 60–90 day retention.
- Don't launch on unverified tracking — fire test conversions and confirm they land first.
The 15 Most Expensive Founder Mistakes
- Judging ads on feel instead of the four numbers.
- Sending cold ad traffic to a free offer, teaching the algorithm to find freebie-seekers.
- Launching before conversion tracking is verified.
- Starting with Performance Max to "keep it simple."
- Switching bidding phases before hitting the conversion threshold.
- Changing five things at once, then not knowing what worked.
- Scaling on two good days and triggering a relearn.
- Scaling a subscription on projected LTV without retention data.
- Skipping the upsell at the moment of peak trust.
- Never sending the 2-day follow-up email.
- Outsourcing the checkout page and losing the conversion signal.
- Running the Meta pixel without the Conversions API.
- Testing adjectives instead of angles in creative.
- Riding one winning ad until it collapses with no pipeline behind it.
- The heroic landing-page redesign that teaches you nothing.
40 AI Prompts for Running Ads
Load your Ads Command Center context first. Fill the [brackets].
Strategy & economics
- Calculate my max allowable CPA for
[product/price/margin]and the conversion rate I'd need at 3 CPA levels. - Model my funnel:
[traffic, CVR, upsell take-rate, AOV]— show revenue and CPA at each stage. - Is my business better suited to direct-sale or subscription economics given
[details]? Explain the scaling implications. - Model 3 retention scenarios (30/50/70% at 90 days) for my subscription and when each makes scaling safe.
Google Ads
- Generate a keyword list for
[offer]grouped by intent, plus 30 negative keywords. - Write 12 RSA headlines (30 char) and 4 descriptions (90 char) for
[offer]emphasizing[benefit]. - Here are yesterday's metrics
[paste]and my phase[X]. What should I adjust — or should I wait? - Have I hit the threshold to switch from Maximize Clicks to Conversion Value?
[paste conversions] - Review my search-terms report
[paste]— which to add as keywords, which as negatives? - Explain what my Quality Score components suggest I should fix:
[paste].
Meta Ads
- Propose a Meta audience testing plan for
[offer/customer]: customs, lookalikes, when to go broad. - Write 5 primary texts and 5 headlines for
[offer], each a different angle, in this voice[paste]. - Write 3 retargeting ads for visitors who didn't buy
[offer], each addressing an objection. - My ad-set metrics:
[paste]. Which audience is winning, and what do I scale or cut? - Diagnose likely creative fatigue from these trends
[frequency/CTR/CPA over time].
Funnel, email & CRO
- Write my results-page inline upsell tied to result
[X]for premium offer[Y]. - Write my results email (deliver value + upsell CTA) and 2-day follow-up for non-buyers.
- Write 5 hero headline variations matched to this ad
[paste ad](message match). - Here's my page copy and drop-off data
[paste]— biggest leak and the one test to run first? - Give me a prioritized list of 10 A/B tests for this page, ranked by likely impact.
Creative & offers
- Give me 15 distinct advertising angles for
[offer]targeting[customer]. - Write 20 scroll-stopping hooks for the '
[angle]' angle, ranked by stopping power. - Write three 30-second UGC scripts for
[offer], each opening with a different hook. - Here are 50 reviews
[paste]— extract the exact phrases, pains, and desires; turn the best into hooks. - Suggest 5 offer variations (price, framing, bonus, guarantee) to test for
[product].
Analysis, reporting & automation
- Turn this raw export
[paste]into a clean weekly report: what moved, why, and next actions. - Walk down the metric ladder on these numbers
[paste]and find the weakest rung. - Find the bottleneck in my funnel from this data
[paste]and the highest-leverage fix. - Design the Zapier/Make automation to
[connect tools / send follow-up / log conversions]. - Draft a month-end scaling review: am I ready to scale, and by how much, given
[data]?
Setup & QA
- Walk me through conversion tracking for a 3-event funnel; which primary vs secondary and what values?
- Give me a pre-launch tracking QA checklist for
[funnel]. - Generate consistent UTM parameters for these campaigns
[list]. - Explain enhanced conversions and offline conversions for my
[funnel type]and whether I need them. - Write a Lovable prompt to build my
[landing/checkout]page with sections[list].
Positioning (for ad-service sellers)
- Analyze this prospect's current ads/feed
[paste/describe]and find 3 specific problems to raise on a call. - Write a discovery-call script that builds credibility by diagnosing their account live.
- Draft honest expectation-setting language for month one (no guarantees) that builds trust.
- Turn this case study
[paste]into proof of process, not a promise of results. - Write outreach to a prospect who's "been burned before" by a previous agency.
Your 30-Day Action Plan
| Week | Focus | Outcome |
|---|---|---|
| Week 1 | Foundation: nail unit economics + max CPA (M1–2). Build the qualifying funnel and upsell touchpoints (M3). Set up & test conversion tracking (M5). | A funnel and verified tracking — ready to spend |
| Week 2 | Launch: Google Search in Phase 1 (Maximize Clicks) and/or Meta with pixel+CAPI, custom + 1% lookalike (M4, M6). Ship 3 creative angles (M7). Don't touch it 3–4 days. | Live campaigns buying data |
| Week 3 | Optimize: hit 10–15 conversions → switch to Conversion Value. Run search-terms + negatives. Ship one CRO test. Refresh creative (M4, M7, M8). | Stabilizing CPA, first optimizations |
| Week 4 | Scale prep: set Target ROAS once CPA is stable. Watch for "limited by budget." Plan first vertical step + a horizontal path. Build your Ads Command Center + daily Claude rhythm (M9–10). | A machine ready to scale — and the habit to run it |
Recommended Tools
Run the ads: Google Ads, Meta Ads Manager.
Track & measure: Google Tag Manager, Google Analytics 4, Meta Events Manager (Pixel + Conversions API).
Build pages: Lovable (fast AI page builds), Webflow, or your site platform — plus a standalone checkout (Stripe).
Email & follow-up: your ESP of choice for the results + 2-day sequences.
Automate: Zapier or Make to connect the stack.
The AI brain: Claude (analysis, copy, creative, CRO, reporting) and ChatGPT as a second generator.
Bonus — Positioning for Anyone Selling Ads Management
Plain-English Glossary
CPA / CAC — cost to acquire one customer (CAC includes all sales/marketing costs). Conversion value — immediate revenue per conversion. ROAS — revenue ÷ ad spend × 100.
CTR / CPC / CPM — click rate / cost per click / cost per 1,000 impressions. CVR — conversion rate.
LTV — total revenue from a customer over time. AOV — average order value. Payback period — time to recoup acquisition cost.
Learning phase — the 3–4 days a platform re-optimizes after a change. GCLID — Google Click ID cookie (90-day attribution window).
Pixel / CAPI — Meta's browser tag / server-side Conversions API. Custom / Lookalike audience — your data / people who resemble it.
Match types — broad / phrase / exact keyword interpretation. Negative keywords — terms you exclude. Quality Score — Google's relevance rating.
tCPA / tROAS — target CPA / ROAS bidding. Frequency — average times a person saw your ad. Fatigue — decay from overexposure. Enhanced / offline conversions — first-party and off-site conversion imports.
End of Course
Final Reflection
- Do I know my four numbers — CPA, conversion rate, landing-page performance, LTV — well enough to make decisions today?
- Is my funnel qualifying buyers with a paid entry and capturing them with tracked upsells?
- Have I earned the right to scale — verified tracking, stable CPA, and (if subscription) confirmed retention?
- Am I running an operation with a rhythm, or touching my ads emotionally?
- Am I keeping the judgment and automating the labor with my AI team?