The Growth Engine

Marketing teams drown in data but starve for decisions.

The symptoms:

The root cause:

Most dashboards are reporting layers—they tell you what happened.

The Growth Engine is a decision engine—it tells you what to do next, and why.

The Fundamental Question:

Every new campaign starts in "Exploration Mode"—you're paying to gather data, not yet making profit.

The question is:

"When do I have enough data to confidently scale?"

The Growth Engine solves this with statistical rigor. It doesn't guess. It calculates.

What Makes This Different

Traditional Approach:

"This campaign has 30 conversions and 3.2x ROAS. Should I scale?"

Answer: "Maybe? Feels okay?"

Growth Engine Approach:

"This campaign has 30 conversions, but needs 42 to reach statistical significance (95% confidence). It's in Calibration stage with 18% volatility (threshold: 15%). Readiness Score: 72%."

Answer: "Wait. Need 12 more conversions and volatility must drop below 15% before scaling."

The Four Workbenches:

1. Diagnostics (The Brain)

Answers: "Is this campaign statistically ready to act?"

2. Portfolio Balance (The Fund Manager)

Answers: "Where should capital be pushed, fixed, protected, or cut?"

3. Macro Funnel (The Plumber)

Answers: "Where is growth leaking in the user journey?"

4. Market Share (The Opportunity Hunter)

Answers: "What is the cost of inaction in the auction?"

The Key Innovation:

Campaign states are not Google states. They are statistical confidence states.

Google says "Learning" or "Eligible."

We say "Exploration (47% confidence)" or "Exploitation (96% confidence, stable)."

This system doesn't report performance. It governs capital under uncertainty.

Purpose: Explain the statistical lifecycle of a single campaign and prescribe the correct operational posture.

Answers: Should I wait? Should I change? Should I scale? Should I stop?

The Core Model: Minimum Viable Signal (MVS)

The minimum sample size (conversions + spend) required to trust performance metrics within a confidence interval.

Why it matters:

"Wait for X conversions" is arbitrary. A campaign spending $50/day needs fewer conversions to prove itself than one spending $500/day. MVS accounts for both conversion volume and spend efficiency simultaneously.

For Conversion-Based Campaigns:

C_min = (Z² × p × (1 - p)) / ε²

Where:

Example:

Floor applied: Balanced mode requires minimum 10 conversions.

Dynamic Spend Target:

MVS isn't just a conversion count—it's also a spend threshold.

MVS_spend = max(

(C_min × Historical_MVS) / Historical_Conversions,

C_min × Actual_CPA,

0.05 × Cumulative_Spend

)

Example:

Readiness Score (0-100%)

Visual: Circular progress gauge

Readiness = (Current_Conversions / MVS_Target_Conversions) × 100

Color Coding:

The 4 Lifecycle Stages

Stage 1: Exploration (0-70% Readiness)

Buying data. Inefficiency is expected.

Your action: Wait. Don't panic. Let the campaign gather evidence.

Stage 2: Calibration (70-95% Readiness, High Volatility)

Enough data exists, but the signal is unstable. ROAS swings wildly day-to-day.

Your action: Monitor closely. Don't scale yet. Narrow targeting if volatility persists.

Exit criteria: Volatility must drop below threshold for 7 consecutive days.

Stage 3: Exploitation (95%+ Readiness, Low Volatility)

Statistically safe to scale.

Your action: Scale aggressively—increase budget by 30-50%, monitor for 7 days, repeat.

Stage 4: Degradation (ROAS Declining)

Efficiency decay confirmed. Capital is at risk.

Possible causes: Ad fatigue, audience saturation, competitor outbid, seasonal shift, broken landing page.

Stability Score (The Volatility Guard)

Stability = 1 - min(Volatility, 1)

Where: Volatility = Std_Dev(ROAS) / Mean_ROAS (7-day rolling window)

A campaign with 4x ROAS but 30% volatility is riskier than 3x ROAS with 10% volatility. The first could collapse tomorrow. The second is predictable.

Sensitivity Profiles (The Risk Dial)

Conservative (95% Confidence): Z=1.64, Error=20%, Volatility Threshold=25%, Min Conversions=15

Use for: High-stakes campaigns ($5K+ daily budgets), risk-averse clients.

Balanced (90% Confidence) — Default: Z=1.28, Error=30%, Volatility Threshold=40%, Min Conversions=10

Use for: Standard campaigns, most accounts.

Aggressive (85% Confidence): Z=1.04, Error=40%, Volatility Threshold=60%, Min Conversions=5

Use for: Startup mode, testing new markets, small budgets.

The Countdown Widget

Instead of vague "wait and see," you get concrete targets:

🎯 Readiness: 72%

Progress to MVS: ▓▓▓▓▓▓▓▓▓▓▓▓▓░░░░░░░ 72%

Need:

• 4 more conversions (9/13)

• $800 more spend ($2,200/$3,000)

Estimated Time: 3-5 days (based on current pace)

Campaign Maturity Table (The Audit Trail)

A time-indexed ledger showing the statistical state of each campaign on each day.

Columns: Date, Campaign, Stage, Readiness (0-100%), Stability (0-100%), Cumulative Conversions, MVS Target, Cumulative Spend, MVS Spend Target, Shield Status, Exploration Waste

Use cases:

Purpose: Classify campaigns into 4 quadrants and allocate capital strategically.

The 4-Quadrant Framework

Visual: Scatter plot with bubble chart

Quadrant 1: Rockets 🚀 (Top-Right)

High ROAS (≥1.5x) + High Headroom (≥40%)

Translation: A money printer with untapped potential.

Example:

Action: PUSH (Scale aggressively)

The Rocket Diagnostic Modal checks:

Tiered Recommendations:

Quadrant 2: Cash Cows 🐄 (Top-Left)

High ROAS (≥1.5x) + Low Headroom (<40%)

Translation: Steady revenue, but maxed out on impression share.

Action: PROTECT (Defend market share)

Protection strategies:

Quadrant 3: Question Marks ❓ (Bottom-Right)

Low ROAS (<1.5x) + High Headroom (≥40%)

Translation: Inefficient, but there's potential if you fix the problems.

Action: FIX (Diagnose and optimize)

The Question Diagnostic runs 3 checks:

Decision tree:

Quadrant 4: Dogs 🐕 (Bottom-Left)

Low ROAS (<1.5x) + Low Headroom (<40%)

Translation: Burning cash with no upside.

Action: CUT (Pause or salvage)

The Dog Diagnostic calculates a Salvage Score (0-100) based on:

If Salvage Score >60: Cut budget by 70%, fix top 3 waste sources, monitor 7 days.

If Salvage Score <60: Pause immediately.

Portfolio Efficiency Score (0-100)

Composite health metric: Rockets (+15 pts each), Cows (+8 pts), Questions (-2 pts), Dogs (-10 pts).

Hide Brand Control

Toggle to filter out "brand" campaigns that distort the scatter plot (typically 8-12x ROAS, near-100% IS). Focus on growth campaigns without brand noise.

Purpose: Identify where growth is leaking in the user journey and simulate upside from fixing bottlenecks.

The 3 Stages

Stage 1: Impressions (Reach)

How many people saw your ad. No action needed here unless IS Lost is high (that's a Market Share issue, Workbench 4).

Stage 2: Clicks (Creative Capture)

CTR = Clicks / Impressions

Benchmarks: Search Brand: 8-15%, Search Non-Brand: 2-5%, Display: 0.5-1%, Shopping: 1-3%

If CTR is below benchmark → CTR Bottleneck Detected. Your ad isn't compelling. Fix in Creative Lab.

Lift simulation: Improving CTR from 3% to 4.5% on 500,000 impressions = +7,500 clicks/month. At 5% CVR and $50 AOV = +$18,750/month revenue.

Stage 3: Conversions (Offer/LP Fit)

CVR = Conversions / Clicks

Benchmarks: E-Commerce: 2-5%, B2B SaaS: 5-15%, High-Intent Services: 10-25%, Display: 0.5-2%

If CVR is below benchmark → CVR Bottleneck Detected. Landing page or offer is broken.

Diagnosis checklist: Page load >3 seconds? Message mismatch between ad and LP? Confusing CTA? Price shock? Missing trust signals?

Combined Lift Simulation

Current State:

After fixing BOTH bottlenecks:

Result: +125% more conversions without spending a single extra dollar on ads.

The Leakage Waterfall

Waterfall chart showing drop-off at each stage. Green bars = healthy (within benchmark). Red bars = bottleneck (above benchmark drop-off).

Hover on red bar → Tooltip: Current rate, Industry benchmark, Gap, Potential lift if fixed.

Impression Share Context

Connects to Market Share workbench:

Key insight: Fixing rank issues is free money — better QS means more impressions at the same bid. Budget recapture requires additional spend but often runs at 2-4x ROAS.

Purpose: Quantify auction headroom and calculate the cost of reclaiming lost impression share.

The Core Metrics

1. Impression Share (IS): % of available impressions you captured

2. IS Lost (Budget): % lost because your daily budget ran out

Metaphor: You're a vending machine that runs out of stock at 3 PM. Customers keep coming until 8 PM.

3. IS Lost (Rank): % lost because ad rank was too low (bid × Quality Score × Ad Relevance)

The Recapture Cost Calculator

Example scenario: 62% IS, 23% Lost (Budget), 15% Lost (Rank)

Budget Recapture:

Total Available = 500,000 / 0.62 = 806,451 impressions

Lost (Budget) = 806,451 × 0.23 = 185,483 impressions

Additional Clicks = 185,483 × 3% CTR = 5,564

Additional Spend = 5,564 × $8 CPC = $44,512/month

Additional Revenue = 5,564 × 5% CVR × $50 AOV = $13,910/month

ROAS = $13,910 / $44,512 = 3.1x ✅ Profitable

Rank Recapture (Free Money):

Improving QS from 4 to 7 → Ad Rank jumps 75% → IS Lost (Rank) drops from 15% to ~5%

Reclaimed: ~80,645 impressions → ~2,419 clicks → ~121 conversions

Additional spend: $0 (just fix ad relevance and landing page)

Competitive Benchmarking

Search Absolute Top IS: % of time your ad is in Position #1

Position 1 gets 3-5x more clicks than Position 3.

Market Share Trend Chart

3 lines over 90 days: IS (blue), IS Lost Budget (red), IS Lost Rank (orange)

Pattern 1: IS stable, Lost Budget rising → Demand growing, you're not matching with budget

Pattern 2: IS declining, Lost Rank rising → Competitors outbidding or improving QS

Pattern 3: IS rising, both Lost metrics dropping → Optimizations working

Opportunity Summary Card

💰 Total Opportunity

Budget-Constrained Revenue: ~$14K/mo

Rank-Constrained Revenue: ~$6K/mo

Total Revenue Left on Table: ~$20K/mo

Cost to Reclaim (Budget): ~$44.5K/mo

Cost to Reclaim (Rank): $0 (free)

Priority: Fix rank first (free), then budget (paid).

Money spent attempting to learn, with no result.

Type A: Black Holes

Campaigns that spend money but generate zero conversions despite exceeding MVS spend target.

Criteria: Zero conversions AND spent >2× MVS target spend.

Example:

All $6,800 is waste. Learning requires buying signal. If you spend 2× MVS and get zero conversions, you're not learning—you're burning.

Action: Pause immediately.

Type B: Calibration Traps

Campaigns that have conversions but can't stabilize despite excessive spend.

Criteria: Conversions ≥ MVS target, BUT volatility > threshold, AND spent >3× MVS target.

Example:

ROAS swings: Week 1: 4.2x, Week 2: 1.3x, Week 3: 5.8x, Week 4: 0.9x

This isn't a campaign—it's a slot machine.

Justified learning cost: $3,000 (MVS target). Excess: $6,500 (waste).

Action: Pause and pivot, or drastically narrow targeting.

Why This Matters

Without this category, teams justify endless "learning" spend:

"We need more time... Just a few more days... It's still learning..."

The Growth Engine says: "You've spent 3× MVS. The engine has spoken. This isn't learning—it's waste."

Exploration Budget Allocation

Recommended split:

Too much exploration (30%+) = wasting money on unproven bets.

Too little (5%) = missing new opportunities, over-dependent on old campaigns.

15% is the sweet spot for most accounts.

Use Growth Engine when you want to:

Don't use Growth Engine when you want to:

Who Should Use This:

✅ Growth Marketers (strategic budget decisions)

✅ Founders (understanding when to scale)

✅ Analysts (statistical rigor for recommendations)

✅ Campaign Managers (optimizing campaign portfolio)

❌ Executives (use PCC for high-level status)

❌ Junior Practitioners (start with PCC and Creative Lab)

How Often:

Success Metrics:

Ultimate Goal: 80% of budget in Exploitation (proven winners), 20% in disciplined Exploration.