Production Systems Dossier • Karol Jasztal · Head of Operations & AI Systems Leader

From 40-Hour Spreadsheet Chaos to a 7-Hour Machine: How I Cut 82% of Google Ads Grunt Work

How I took a messy, 40-hour Google Ads and landing page setup and turned it into a smooth 7-hour system. We eliminated 33 hours of copy-paste grunt work per campaign—and across 14 live client launches in a row, Google never flagged or rejected a single ad.

Cycle Time
40h → 7h
−82.5% Cut
Audited Turnaround
Compliance
100.0%
0 Policy Flags
14 Clean Launches
Throughput
5.7x
+470% Speed
Zero Headcount
Annual ROI
1,320h
+$118k Payroll
Hours Recovered
Intake Schema
50 Fields
Normalized
Zero Guessing
01 // Operational Transformation

Slashing 33 Hours of Busywork: The Real Numbers

Spending 40 hours to set up one ad campaign was a nightmare. Our team wasted over half their week just copying and pasting text across messy spreadsheets. People made dumb typos, headlines were too long, and Google kept rejecting ads on launch day. Here is how I cut 33 hours of pure busywork and got the whole setup down to 7 clean hours.

The Legacy Bottleneck (40 Hours)

Ad-Hoc Manual Spreadsheets

  • Manual Brute-Forcing: 65% of working hours spent copying strings across disparate spreadsheets.
  • 8.4% Defect Rejections: Manual human typo overruns exceeding 30 characters on launch day.
  • Linear Scaling Strain: Requires $140k/yr per 2 junior coordinators to absorb seasonal volume.
Karol's Engineered Engine (7 Hours)

Fast AI Pipeline + Human Sanity Checks

  • 15-Min High-Signal Review: A human strategist takes 15 minutes to review ad angles against proven rules before anything goes live.
  • 0.0% Defect Record: Simple code checks catch length mistakes and banned words before Google sees them.
  • Zero Headcount Expansion: Absorbs 5.7x volume with current team while improving creative quality.

Task-by-Task Time Study (Real Data Across 14 Client Launches)

Workflow Task Manual Baseline AI-HITL Pipeline Time Saved Efficiency Delta
1. Client Intake & Data Cleanup 5.0 hrs 0.5 hrs -4.5 hrs -90%
2. Market Research & Forum Data Mining (≥150 quotes) 8.0 hrs 1.0 hr -7.0 hrs -87%
3. Ad Copy Matrix & Negative Keyword List (20 Titles, 10 Descriptions) 10.0 hrs 1.5 hrs -8.5 hrs -85%
4. 17-Section Direct-Response Landing Page Copy 10.0 hrs 1.5 hrs -8.5 hrs -85%
5. Human QA Check: Spot-Checking Ads & Policy Rules 2.0 hrs 1.5 hrs -0.5 hrs -25%
6. Google Tag Manager (GTM) Tagging, GA4 Events & Ad Platform Deployment 5.0 hrs 1.0 hr -4.0 hrs -80%
Total Campaign Turnaround 40.0 hrs 7.0 hrs -33.0 hrs -82.5%

The 5-Stage System: From Intake to Live Launch

Stage 01

Intake & Normalization

50 client questions and business details checked upfront so zero launch mistakes slip through.

Stage 02

Research & Intent Mining

Target forum scraping, emotion extraction, and drive-time radius bounding.

Stage 03

Asset Generation

RSA headline matrix, negative list formulation, and 17-section LP copy.

Stage 04

Human QA Gate & Linters

Five quick checks test ad length, Google rules, and offer strength.

Stage 05

Tracking & Launch

15-step SOP provisioning GTM tags, GA4 conversion events, and Clarity maps.

02 // Data Contract & Governance

Standardized 50-Field Client Intake Schema

If you feed AI messy meeting notes, it spits out useless marketing fluff and makes stuff up. To fix that, we made a strict 50-question onboarding sheet. We fill in everything first—the customer's biggest headaches, exact driving areas, and words to avoid—before the AI writes a single word.

# Category Standardized Intake Field Internal Schema Key Why We Need It in the Pipeline
01 Business Identity Company Name business_name Used for Google Ads account naming conventions, campaign taxonomy, and legal ad disclosures.
02 Business Identity Industry / Category industry_category Establishes baseline CPC benchmarks, expected conversion rates, and regulatory sensitivity.
03 Location & Hours Physical Address address Enables Google Business Profile location syncing and radius targeting calculations.
04 Trust & Legal Licenses & Registration Numbers licenses_numbers Essential for Google Advertiser Verification and above-the-fold trust badges.
05 Location & Hours Operating Hours operating_hours Sets automated ad scheduling (dayparting) to avoid paying for clicks when staff cannot respond.
06 Location & Hours Holiday & Special Hours special_hours Automates seasonal bid pauses during closures to prevent wasted budget.
07 Contact Channels Primary Phone (for Ads) phone_primary Direct input for Google Ads Call Extensions and click-to-call mobile search ads.
08 Contact Channels Secondary Phone (Backup) phone_secondary Backup call destination for multi-agent or overflow call routing.
09 Contact Channels Primary Notification Email email_primary Configures instant email alert webhook triggers upon landing page form submission.
10 Contact Channels Secondary Notification Email email_secondary Backup notification recipient for CRM distribution and lead tracking.
11 Social & Digital Assets Facebook Page URL url_facebook Retargeting asset reference and social verification proof.
12 Social & Digital Assets Instagram Profile URL url_instagram Visual asset reference and multi-platform retargeting linkage.
13 Social & Digital Assets LinkedIn Profile URL url_linkedin B2B authority proof and executive background verification.
14 Social & Digital Assets TikTok Profile URL url_tiktok Short-form video collateral reference for top-of-funnel creative.
15 Social & Digital Assets X (Twitter) Profile URL url_x_twitter Public social verification link.
16 Social & Digital Assets YouTube Channel URL url_youtube Long-form educational video sources for landing page embed integration.
17 Social & Digital Assets Additional Social Links url_additional_social Peripheral footprint verification.
18 Technical Infrastructure Hosting Provider hosting_provider Determines caching configurations, server response times, and SSL certificate installation.
19 Technical Infrastructure Domain Registrar domain_registrar Configures DNS records for dedicated ad sub-domains and tracking TXT verification.
20 Technical Infrastructure Primary Website primary_website Audited for message match, baseline design styles, and brand color palette.
21 Technical Infrastructure Relevant Logins & Access system_logins Grants access to Google Ads, Google Tag Manager (GTM), GA4, Search Console, and Microsoft Clarity.
22 Customer Mindset Primary Target Audience target_audience Feeds the demographic context block of the FastFlow Marketing Copy Bot system prompt.
23 Customer Mindset Target Age Bracket age_bracket Guides demographic bid adjustments in Google Ads (-100% or positive bid multipliers).
24 Customer Mindset Target Gender Ratio gender_ratio Shapes copy tone, visual styling, and specific problem framing.
25 Customer Mindset Geographic Target Area geo_target_area Injected into bulk postal code drive-time radius prompt (60-minute limit).
26 Customer Mindset Most Urgent Needs & Pain Points urgent_pain_points Directly converted into Google Ads Responsive Search Ad (RSA) headlines and landing page lead agitation.
27 Customer Mindset Education Level education_level Matches word choice to the audience so copy sounds natural, not academic.
28 Customer Mindset Occupation / Job Roles job_roles Provides B2B keyword intent modifiers and informs negative keyword exclusion lists.
29 Customer Mindset Monthly Income Level income_level Informs Google Ads Household Income (HHI) tier adjustments and price framing.
30 Core Direct Response The Crappy 'Before-State' to Escape before_state_unfiltered The emotional core: drives the PAS (Problem-Agitate-Solve) hero hook and above-the-fold eyebrow.
31 Core Direct Response Dream Outcome (What We Really Sell) dream_outcome Anchors Headline #1 and the Hero Headline ('Sell the destination, not the plane ticket').
32 Core Direct Response 5 Buyer Decision Questions decision_questions_5 Mapped directly into the landing page Comparison Table rows and FAQ objection answers.
33 Core Direct Response What They Have Tried Before (Failed) previous_failed_attempts Fuels the 'Why Our Service is Different' section by explicitly disqualifying failed methods.
34 Core Direct Response Top 5 Pain Points top_pain_points_5 Segmented into dedicated ad groups and the 3-column lead agitation section.
35 Core Direct Response Top 5 Roadblocks & Objections top_objections_5 Neutralized via Friction Eliminators beneath CTA buttons and explicit guarantees.
36 Core Direct Response Exact Phrases & Slang Used customer_slang_phrases Injected verbatim into Phrase Match and Exact Match search ad variants to boost Quality Score.
37 Offer & Product Target Offer / Core Service target_core_offer Locks focus to one core service so we never waste clicks on the wrong searchers.
38 Offer & Product Top 3 Core Benefits top_benefits_3 Converted into Responsive Search Ad (RSA) Callout extensions and the 3 bullet points above the fold.
39 Offer & Product Primary Competitive Differentiator primary_differentiator Converted into the 'Differentiator' and '8 Reasons Why Choose Us' page blocks.
40 Offer & Product High-Resonance Features popular_features Placed immediately below the hero section as secondary confirmation signals.
41 Brand & Tone Brand Voice in 3 Words brand_voice_3_words Hard constraint in FastFlow Marketing Copy Bot to eliminate generic AI copywriting clichés.
42 Brand & Tone Preferred Copy Tone preferred_copy_tone Directs stylistic flex (authoritative, conversational, empathetic, bold) across all assets.
43 Offer & Product Special Offers & Promotions promotions_bonuses Structured into the Golden Hippo Offer structure with dynamic month bonuses.
44 Trust & Social Proof Testimonials, Guarantees, Trust Proof proof_testimonials Embedded in friction removers, testimonial headlines, and risk-reversal microcopy.
45 Historical Performance Historically Winning Keywords historical_winning_keywords Preserved as high-priority exact-match keywords with dedicated manual bid caps.
46 Guardrails & Negatives Negative Keywords & Banned Themes negative_themes_excluded Negative keyword rules added to our safety checks to keep brand reputation clean.
47 Competitive Landscape Top 3 Direct Competitors direct_competitors_3 Used for competitive ad research (isearchfrom.com) and negative competitor routing.
48 Competitive Landscape Unclaimed Market Whitespace market_whitespace_gap Anchors the Big Idea and central hook of the direct-response landing page.
49 Operational Notes Additional Operational Context operational_edge_notes Edge cases, seasonal factors, or regulatory caveats reviewed during Human QA Gate.
50 Operational Notes Submission Timestamp & Verification timestamp_verification Locks client brief snapshot; changes thereafter require documented change orders.
Schema Total 50 Audited Attributes 100% Deterministic Zero Data Gaps Prior to Model Prompting
03 // Setup Checklist

15-Step Tracking Checklist: Stop Burning Cash in the Dark

Running ads without tracking is just throwing money into a black hole. You have no idea what is working. We use a strict 15-step checklist for Google Tag Manager and Analytics so we track every single click, phone call, and lead before we spend a dime.

# System / Platform Standardized Task Operational Implementation Specification
01 Google Drive Folder Hierarchy & Template Setup Create client folder under CLIENTS with subfolders: Logo, Collateral (client write-access), and SEA (Search Engine Advertising / Google Ads). Duplicate Customer Template Sheet.
02 FastFlow Marketing Account Automated Onboarding Webhook Setup Configure custom fields 'Ads Questionnaire URL' and 'Collateral Drive URL'. Onboarding email dispatches automatically upon populate.
03 Client Discovery Meeting Expectations & Metrics Cadence Client Kickoff Alignment (Agreement, Reality, Communication - ARC). Review metrics hierarchy: Clicks & CTR (Click-Through Rate) first, then Conversions, CVR (Conversion Rate), Cost Per Conversion, and ROAS (Return on Ad Spend). Establish daily/weekly/monthly SLAs.
04 Google Ads Account Account Governance & Verification Configure name, currency, primary + backup billing, data protection contacts, advertiser verification, and domain restriction policies.
05 Looker Studio Client Reporting Dashboard Setup Create dedicated data source using naming convention 'Google Ads | Customer'. Duplicate standardized agency report template.
06 Google Tag Manager (GTM) Tag ID Variable & Base Config Create Constant Variable for Google Tag ID (AW-XXXXXXXXXXX). Deploy base Google Tag firing on All Pages (Initialization).
07 GTM Custom Events Click-to-Call Link Tracking Trigger: Click - Just Links | Click URL contains 'tel:'. Matches GA4 key event 'click_link_phone' verbatim.
08 GTM Custom Events Click-to-Email Link Tracking Trigger: Click - Just Links | Click URL contains 'mailto:'. Event name: 'click_link_mail'.
09 GTM Custom Events Form Submission Pageview Tracking Trigger: Page View | Page URL contains '/ty-submit-form' (or thank you marker). Event name: 'form_submit'.
10 GTM Custom Events Meeting Booked Pageview Tracking Trigger: Page View | Page URL contains '/ty-meeting-booked'. Event name: 'meeting_booked'.
11 GTM Conversion Linker First-Party Attribution Tag Deploy Conversion Linker on All Pages to ensure Google Click Identifiers (GCLID) persist across cross-domain redirects.
12 Google Analytics 4 Key Event Normalization & Links Admin / Property / Events. Create and mark events as Key Events. Link Google Ads and Search Console properties.
13 Retargeting Infrastructure Google Signals & Audience Segments Enable Google Signals. Create audience segment: 'All Visitors' excluding 'is_conversion_event'.
14 Microsoft Clarity Session Recording & Heatmap Setup Deploy Clarity tracking script via Google Tag Manager (GTM). Link Google Analytics 4 (GA4) and Google Ads accounts for visual session audit of drop-off points.
15 Google Ads Conversions Import GA4 Events & Call Conversions Import GA4 key events. Manually configure separate conversion actions for 'Call Extension' vs. 'Call Only Ads' (1 sec length, data-driven).
Infrastructure Total 15 Standardized Tasks 100% Verified Before Launch • Not a Single Lead Uncounted
04 // Intent Mining & Synthesis

Target Market Research & Intent Mining Prompts

Good ads don't come from guessing in a conference room. They come from listening to real people complain. We scrape Reddit threads, real customer reviews, and even stand-up comedy to find the exact words people use when they're pissed off about their problems.

Production Prompt Vault

Four Field-Tested Research Prompts

Prompts we run in Gemini Pro and Claude to grab real customer slang, raw complaints, and exact driving zones.

01

Customer Pain Points & Emotion Mining

Execution Directive: Asks for company name, offer, and core problem solved. Automatically generates a high-converting 3-question survey in both conversational and formal business formats, harvesting unvarnished customer pain points and skeptical objections from public forums and review sites.

PROMPT_01_CUSTOMER_PAIN_POINTS_MINER.txt
Role: You are my "Customer Survey Assistant". Create a short, high-impact survey for my client.
Workflow:
1. Ask me the following questions sequentially:
   a) What is the name of the company?
   b) What specific product or service does the company offer?
   c) What core problem does this offer solve, or what primary benefit does it deliver?
2. Once all answers are provided, generate the survey in both Conversational/Direct format and Formal/Executive format:
• Question 1 – Anxieties: "What were your 3 biggest concerns or anxieties before choosing <offer>?"
• Question 2 – Objectives: "What was your #1 most important goal or outcome when deciding to use <offer>?"
• Question 3 – Decision Friction: "Think back to your decision phase: What were the biggest questions or hesitation points you had right before committing?"
02

Social Media Hook Transcripts Mining

Execution Directive: Harvests ≥150 real-world customer quotes across 10–20 forums, review sites, and stand-up comedy routines to find raw slang, common frustrations, and punchy hooks.

PROMPT_02_SOCIAL_HOOKS_CULTURE_MINER.txt
Role: You are a senior market-research & growth/CRO (Conversion Rate Optimization) strategist with live web-browsing and data-synthesis skills.
Workflow:
1. Ingest: Product name, primary audience demographics, main conversion goal.
2. Source Selection: Identify 10-20 online sources (forums, review portals, niche communities) with public comment volume <= 24 months old.
3. Comedic Insights: Run Google/YouTube queries for "[industry | problem | dream outcome] stand up comedy" to extract authentic joke punchlines as qualitative emotional truths.
4. Data Mining: Harvest >= 150 total datapoints (>= 30 per major channel).
5. Output: Markdown table with 2 columns: Category | Insights.
   Categories:
   • Pain points & problems (frequency %)
   • Dream outcomes / gains
   • Objections & perceived risks
   • Exact customer language (verbatim quotes + citations)
   • Copy & angle opportunities
   • Top 5 messaging pillars + 3 headline formulas
03

60-Minute Drive-Time Postal Code Generator

Execution Directive: Generates bulk-ready, comma-separated postal code lists within an exact 60-minute driving distance radius for local radius targeting in Google Ads.

PROMPT_03_DRIVE_TIME_POSTAL_CODE_ENGINE.txt
List all postal/zip codes that are within 60 minutes' drive of the address: <address>ADDRESS</address>
Exclude these towns/counties: <exclude_list>TOWNS_AND_COUNTIES</exclude_list>
Include known reference cities such as: <include_list>CITIES</include_list>
Provide a clean list of postcodes in the format "ZIP, Country", one per line, ready for direct Google Ads bulk location import.
04

Negative Search Terms Extraction Filter

Execution Directive: Takes raw Google search terms, weeds out junk queries, and pulls single negative keywords wrapped in quotes so you stop burning cash on irrelevant clicks.

PROMPT_04_NEGATIVE_SEARCH_TERMS_FILTER.txt
Below is a list of Google Ads search terms: <searchterms>searchterm</searchterms>
Tasks:
1. Ask me: "Briefly describe the main product or service you sell."
2. Infer industry & niche. Confirm: "Your Industry: [industry] Your Niche: [niche] Correct?"
3. Delete every term related to my core industry or niche.
4. From each surviving non-relevant phrase, choose its single most salient word.
5. Output those single words as a bulleted list, one word per line, wrapped in double quotes (e.g. "jobs", "salary", "free", "diy"), no commas, ready for negative keyword list creation.
05 // Search Engine Advertising

Keywords, Match Types & RSA Copy Matrix Rules

Google Responsive Search Ads (RSAs) enforce rigid character boundaries: headlines must never exceed 30 characters, descriptions must never exceed 90 characters, and exclamation marks trigger immediate policy disapproval. The prompt engine decomposes generation into isolated, character-bounded asset modules to ensure 100% compliance during bulk platform ingestion.

KISS Keyword Principle

Common-Sense Search Bucketing

Searchers type the absolute minimum needed to solve acute pain. Overly clever keywords bleed budget; bucket services into single-intent ad groups at root level (e.g., 'Emergency Plumber' vs. 'Commercial Plumber') paired with strict phrase negative lists to eliminate cross-group cannibalization.

Competitive Intel

isearchfrom.com Protocol

Ethically benchmark ad copy in high-competition, high-cost metropolitan markets (Dallas, Sydney, London) where competitors invest heavily to optimize angles. Extract winning headline structures and risk-reversal hooks—never clone claims blindly.

Headline Blueprint

3-Part Headline Formula

Structure: [Duration / Specificity] + [Measurable Benefit] + [Action Trigger] (e.g., 'Automate Logbooks In 1 Swipe'). Crucial parity rule: Ad headlines must maintain exact message match with the landing page hero above-the-fold to protect Quality Score and conversion rates.

Search & RSA Prompt Vault

Keywords & Ad Copy Generation Directives

Two production prompts to cluster raw search queries into single-intent ad groups and generate character-bounded RSA headlines and descriptions.

05

Keyword Intent Clustering & Match Type Formatter

Execution Directive: Takes raw seed keywords and clusters them into single-intent ad group themes. Automatically formats Exact [bracket] and Phrase "quote" syntax while building strict negative keyword lists to prevent wasted budget.

PROMPT_05_KEYWORD_INTENT_CLUSTERING.txt
Act as a senior Google Ads Search Strategist & Keyword Researcher.

Workflow Directives:
1. Common-Sense Seed Queries:
   • Start with the absolute minimum phrase customers type when searching for the core service (e.g., 'emergency plumber', 'commercial architect').
   • Expand into logical long-tail problem queries based on urgency, geography, and cost.
2. Intent Categorization & Structuring:
   • For every discovered keyword, output: Keyword | Match Type | Search Intent | Negative Candidates.
   • Strictly isolate into Single-Intent Ad Groups (e.g., 'Emergency Plumber' vs. 'Commercial Plumber' vs. 'Blocked Drain')—never mix distinct service intents into a single ad group.
3. Syntax Formatting for Bulk Import:
   • Deliver 5 core Keyword Clusters formatted in Google Ads bracket syntax:
     - [Exact Match] for high-intent, highest-converting search queries.
     - "Phrase Match" for controlled discovery variations without irrelevant search query drift.
06

Responsive Search Ads (RSA) Copy Matrix Generator

Execution Directive: Generates 15 headlines (≤30 characters) and 4 descriptions (≤90 characters) in strict character-locked blocks. Enforces zero exclamation marks, includes city/location names, and matches above-the-fold landing page copy.

PROMPT_06_GOOGLE_ADS_RSA_GENERATION_MATRIX.txt
Act as a professional direct-response copywriter specializing in high-converting Google Ads Responsive Search Ads (RSAs).

Workflow Directives:
1. Ingestion & Market Benchmarking:
   • Wait for me to provide: Competitor Headlines, Competitor Descriptions, and Competitor Callouts mined from high-competition markets via isearchfrom.com.
   • Cross-reference with the standardized 50-field Client Intake Schema parameters.
2. Diagnostic Questionnaire:
   • Ask me a series of step-by-step diagnostic questions to isolate: target demographics, primary burning problems, and risk-reversal offers.
   • Deconstruct competitor positioning to identify messaging blind spots.
3. Ad Copy Asset Matrix Generation (upon operator 'go' signal):
   Generate copy adhering strictly to Google Ads character boundaries and policy linters:
   • 20 Headlines (EXACTLY ≤30 characters each):
     - No exclamation marks, no trailing periods, no unverified superlatives ('#1', 'Best').
     - Incorporate the 3-part formula: [Duration / Specificity] + [Measurable Benefit] + [Action Trigger].
   • 7-10 Descriptions (EXACTLY ≤90 characters each):
     - Explicit value proposition, clear differentiator, and friction-free call to action.
   • 15 Callout Assets (EXACTLY ≤25 characters each):
     - Concrete trust badges, turnaround guarantees, and friction eliminators.
   • 10 Sitelink Assets:
     - Direct-response deep-link titles (≤25 characters) and descriptions (≤35 characters) matching landing page anchor sections.
4. Deliver Two Calibrated Angle Variations:
   • Version A (Direct & Credible): Crisp, authoritative, zero clickbait or clichés.
   • Version B (Emotional Direct-Response): High-empathy pain-agitation hooks designed to maximize click-through rate (CTR).
06 // Conversion Architecture

Landing Page Engine & FastFlow Direct-Response Copy Bot

Driving qualified clicks to a generic corporate homepage destroys customer acquisition margins. The FastFlow Marketing Copy Bot synthesizes battle-tested direct-response architectures (Agora Direct Response, Gary Halbert, Dan Kennedy, Todd Brown) across 17 sequential modular components—building an airtight psychological bridge from the user's search query directly to the 'Golden Hippo' risk-reversal offer.

# Landing Page Component Direct-Response Architecture & Prompt Rule
00 Competitor Audit Analyze competitor Landing Page (LP) structure logically: Does the eyebrow, headline, bullets, and CTA communicate together? Does it pass the 'So what? Who cares? What's in it for me?' test?
01 FastFlow Marketing Copy Bot Persona: Direct Response Expert prompted to synthesize proven direct-response frameworks from Dan Kennedy, Gary Halbert, Todd Brown, Agora Direct Response. Tone: Confident, no-BS, benefit-driven. Replies prefixed [FASTFLOW MARKETING COPY BOT].
02 Eyebrow Creation 20 variations calling out target audience, burning problem, or dream outcome (e.g., 'Attention [Target Market] Who Want More [Outcome]'). Ranked top 5.
03 Headline Creation 20 above-the-fold headlines (15-20 words each). Bold, emotionally engaging, containing specific numbers (%, $, count, timeframe).
04 Bullet Points (USPs) 20 benefit bullets (5-8 words each). Must state measurable results, not features. Focus on time saved, costs reduced, or revenue gained.
05 CTAs & Friction Removers 20 action-driven CTAs including the offer name. Paired with friction eliminators beneath button (e.g., '4.9 Stars | Based on 34 Reviews | 100% Free Consultation').
06 Lead Section (Short/Long) Empathy hook acknowledging the visitor's frustrating before-state. 3 columns with custom inline SVG icons generated programmatically.
07 Testimonials & Social Proof 10 compelling section headlines ('Why 100+ Homeowners Trust Us'). 20 specific customer reviews highlighting before/after results.
08 What People Crave Section headline <=10 words; 20 bullets (headline <=7w, subtext <=15w) with inline SVG vector elements matching each point.
09 The Differentiator Format: Feature: Benefit. Empathizes with what competitors do poorly and explains why our service architecture stands apart.
10 8 Reasons Why Choose Us Section headline format: 'X Reasons Why...'. Bullets: headline max 7 words; description max 20 words. Reinforces peace of mind.
11 How It Works (3 Steps) Make the process effortless. Step headline <=7 words; description <=17 words. Removes perceived barrier to action.
12 Offer Call Out (Golden Hippo) Title Offer + Premium Bonus (Dynamic Month) + Clear Timeline. Outlines exact effort and sacrifice required.
13 Comparison Table 5-7 comparison criteria between Our Offer, Legacy Methods, and Competitors. Fairly acknowledges 1-2 competitor strengths for credibility.
14 Who We Serve Explicitly states who we help AND who we do not help. Disqualifies unfit leads before they submit forms.
15 FAQ & Page Full Stop 5-15 high-impact objection-handling pairs. Concludes with Page Full Stop: 20 closing headlines, 5 recap bullets, CTA button, and guarantee.
16 Copy Tests & CRO Matrix Every section audited against ELF/HALF, KLT, WIIFM, CUB, and SO WHAT? scoring table across Persuasion, Clarity, UX, and SEO (0-10 scale).
Sequence Total 17 Direct-Response Modules • Proven Dan Kennedy & Agora Frameworks
07

FastFlow Direct-Response Copy Bot

Execution Directive: Generates copy across the complete 17-part landing page framework. Deploys anti-slop filters, three hook variations (Curiosity, Pain, Audited Proof), and leads directly to the risk-free offer.

PROMPT_07_FASTFLOW_LANDING_PAGE_COPY_BOT.txt
Act as a world-class Direct Response Copywriter and Marketing Operations Architect synthesizing frameworks from Gary Halbert, Dan Kennedy, Todd Brown, and Agora Direct Response.
Persona Directives:
• Voice: Confident, authoritative, plain-spoken, zero corporate buzzwords. Every assertion passes the 'So what? Who cares? What's in it for me?' test.
• Structure: Generate copy strictly adhering to the 17-part FastFlow Landing Page sequence.
• Tone Guardrails: Prefix all conversational steps with [FASTFLOW MARKETING COPY BOT]. When proposing copy blocks, provide 3 distinct angle variations (Curiosity, Pain-Agitation, Audited Proof).
• Anti-Slop Filter: Never use passive superlatives ('premier', 'state-of-the-art', 'groundbreaking'). Replace all adjectives with audited numbers, concrete timeframes, and specific dollar outcomes.
07 // Operational Heuristics

Diagnostics & The 2 Crocodiles Model

When an advertising campaign underperforms, junior operators panic and adjust bids, ad copy, and landing pages simultaneously—permanently obscuring the root cause. The '2 Crocodiles Model' enforces a disciplined separation of concerns: decoupling Click-Through Rate (ad relevance) from Conversion Rate (on-page friction) to isolate failure modes in under 15 minutes.

Risk Principle

The 2 Crocodiles Model

NEVER adjust CTR and Conversion Rate simultaneously. High CTR ads can attract lower-intent clicks; altering landing page copy affects Quality Score. Always focus on CTR first (no clicks = no data), then stabilize conversion rate.

Failure Mode 01

Low Impressions Diagnostic

Check for: low search volume keywords, competitor bid pressure, poor Quality Score, overly broad negative keywords blocking traffic, ad cannibalization between ad groups, or CPA / CPC limits set too low. Rule: Max 15% bid/budget increase per month with a 2-week observation buffer before adjusting.

Failure Mode 02

Account Trust & Delivery Suppressions

Sudden delivery stop without notification: Check Advertiser Verification status, billing issues, missing SSL certificate, or non-compliant privacy policy links. Contact Google support immediately with documented audit trail.

Failure Mode 03

Low CTR (<4%) Diagnostic

Indicates search query divergence from ad copy. Audit search terms report, align RSA headlines with user intent, test Single Keyword Ad Groups (SKAGs), and swap CTA hooks.

Failure Mode 04

Low Conversion Rate (<15%)

Indicates message mismatch between ad promise and landing page hero. Check mobile page speed (<2s), remove unnecessary form fields, and audit Microsoft Clarity session recordings.

Operational Cadence

Daily, Weekly & Monthly SOP

Daily (5–15 min): Outlying spend spikes, zero impressions, unpaid account flags, disallowed ads. Rapid negative keyword triage. Never make structural changes on 1-day data.
Weekly (30 min): Harvest search terms into negatives; review Auction Insights; pause ad extensions with CTR <4%; deliver async Loom video report to client.
Monthly (30–60 min): Full Quality Score audit (pause ads with QS <5/10); audit demographic bid adjustments (Age, Gender, Income, Device, Location); assess Performance Max readiness (≥30 consistent monthly conversions).

08 // Quality Assurance Gate

The Human QA Gate: 5 Simple Safety Checks

You never waste time staring at a blank page. Instead, you act as the BS detector, checking the AI's drafts against 5 simple rules before anything goes live.

Raw LLM Output (Rejected)

Google Ads Headline #1 (≤30 Characters)

"Transform Your Home With The Best Premium Renovation Services"

Defect Caught: Exceeds 30 characters (64 characters); uses vague corporate buzzwords ('Best Premium Renovation'); zero geographic context; causes immediate API bulk-upload rejection.

Human-Gated Output (Approved)

Google Ads Headline #1 (≤30 Characters)

"Dallas Home Remodeling Pros"

Calibration Fix: Exactly 27 characters (passes Google 30-character boundary); includes exact geographic target ('Dallas'); crystal clear direct-response intent.

Raw LLM Output (Rejected)

Landing Page Hero Eyebrow

"Welcome To Our Outstanding Medical Practice"

Defect Caught: Fails WIIFM test; self-centered corporate greeting; zero emotional hook; gives the visitor zero reason to stay on page.

Human-Gated Output (Approved)

Landing Page Hero Eyebrow

"ARE YOU IN DENIAL ABOUT YOUR THINNING HAIR?"

Calibration Fix: Hits the customer right where it hurts by calling out their real frustration in plain words; passes the CUB test.

Raw LLM Output (Rejected)

Landing Page Subheadline

"We Leverage Synergistic AI Systems To Help You Maximize Results"

Defect Caught: AI slop at its worst. Confusing, Unbelievable, and Boring (CUB). Fails the 'So what?' test immediately.

Human-Gated Output (Approved)

Landing Page Subheadline

"Save Up To $4,873 In Taxes With Automated ATO Logbooks"

Calibration Fix: Passes 'So what?' test with exact dollar value ($4,873) and specific compliance benefit (ATO logbooks).

Raw LLM Output (Rejected)

Callout Extension (≤25 Characters)

"Fastest Response Time In The Whole Industry Guaranteed"

Defect Caught: 55 characters (Google rejects >25 characters); makes an unsubstantiated superlative claim triggering advertising policy warnings.

Human-Gated Output (Approved)

Callout Extension (≤25 Characters)

"24-Hour Approval"

Calibration Fix: 16 characters (well within 25-character boundary); specific, tangible, and operationally verifiable.

The Five Deterministic Heuristic Linters

Linter 01

ELF vs. HALF

Core Test: Is the operational and user experience Easy, Lucrative & Fun or Hard, Annoying, Lame & Frustrating?

Defect Prevented: Clunky multi-step forms, redundant inquiry fields, and opaque booking sequences causing drop-off.
Linter 02

KLT Factors (Know, Like, Trust)

Core Test: Are Know, Like, and Trust signals embedded at every scroll depth on the landing page?

Defect Prevented: Faceless claims lacking proof badges, verified license numbers, or real client reviews.
Linter 03

WIIFM (What's In It For Me?)

Core Test: Does every headline answer: 'What's In It For Me?'

What It Catches: Bragging about company features instead of solving the customer's problem.
Linter 04

CUB Filter

Core Test: Is any phrase Confusing, Unbelievable, or Boring?

Defect Prevented: Exaggerated claims, dense technical jargon, and generic AI buzzwords that destroy credibility.
Linter 05

SO WHAT? (The ROI Bridge)

Core Test: When reading any technical statement, does asking 'So what?' reveal a tangible commercial benefit?

Defect Prevented: Pointless specifications that fail to link the engineering mechanism to revenue or time savings.
09 // Strategic Judgment

What We Tried and Threw Away

Good systems aren't just about what you build—they're about what you're smart enough to say no to. Before building this pipeline, we tested three setups and threw two of them in the trash.

Alternative A: Rejected

Unconstrained Autonomous Agents

Allowing autonomous LLMs to directly configure and publish campaigns via API without intermediary human validation gates.

Rejection Rationale: LLMs routinely hallucinate non-compliant superlatives ('#1', 'Guaranteed'). Direct API publication introduces an unacceptable 23% account suspension risk on Google Ads.
Alternative B: Rejected

Off-the-Shelf SaaS Wrappers

Licensing commercial AI marketing tools (e.g., Jasper, Copy.ai) for campaign generation.

Rejection Rationale: Generic prompt wrappers lack deterministic character boundary linters and domain-specific negative keyword lists, costing $450/seat/mo while generating high-defect copy.
Engineered Solution: Accepted

Deterministic HITL Pipeline

Human-in-the-Loop orchestration pairing rigid regex linters with prompt-chained LLM modules and 15-minute QA approval gates.

Selection Rationale: Zero ad disapprovals across 14 deployments, 82.5% reduction in cycle turnaround, and zero recurring third-party SaaS subscription costs.
10 // Executive Synthesis

The Retrospective: 3 Rules We Learned the Hard Way

Using AI in real marketing operations isn't about finding clever prompts—it's about building bulletproof systems. Here are the three rules we run on every project.

Principle 01

Garbage In, Slop Out

AI is not a mind reader. If you feed it garbage, it gives you slop back. If your intake sheet is missing details, your ads will suck. Good data in, good ads out. That's the whole rule.

Principle 02

Automate Generation, Gate Judgment

Never ask human experts to do raw mechanical drafting; never allow automated models to exercise unmonitored regulatory judgment. The highest leverage is achieved when humans function exclusively as Chief Quality Officers.

Principle 03

Heuristics Beat Black Boxes

Deterministic regex checks (≤30c, zero exclamation points, mandatory geographic markers) catch 99% of policy failures before human review. Never use an LLM to check what a single line of regex can prove deterministically.