Skill 3: CPA Spike Diagnosis
When CPA spikes, this skill runs a 7-day comparison to find the root cause. Investigate across seven dimensions:
- Which campaigns had the biggest CPA increase?
- Within those campaigns, which ad groups and keywords drove the spike?
- Which search terms got clicks but no conversions?
- Did impression share change (competitive pressure)?
- Did conversion rate drop (landing page or offer problem)?
- Did CPC increase (auction pressure)?
- Were there recent account changes (budget shifts, new keywords, bid changes)?
Deliverable: root causes ranked by severity with specific fixes for each.
Generate a report designed for a founder or marketing manager to read in under 2 minutes:
- Executive summary — 3 sentences: what happened, why, and what you're doing about it
- Metrics table — spend, impressions, clicks, CTR, CPC, conversions, CPA, ROAS — all with week-over-week change
- Top 3 winning campaigns — what's working and why
- Top 3 struggling campaigns — diagnosis and planned response
- Search term changes — new high-volume converters and new wasters added to negatives
- Prioritized action items for the coming week
No jargon. Write for someone who approves budgets, not someone who manages bids.
Section 2: Search Terms & Keywords
Skill 5: Search Term Mining
Analyze all search terms from the last 30 days and categorize into three buckets:
- "Add as keywords" — high-performing terms not yet added as exact-match keywords
- "Negate immediately" — high-spend, zero-conversion terms and irrelevant queries
- "Review" — borderline cases that need human judgment
Also identify: high CTR but low conversion rate terms (intent mismatch), and patterns in converting terms (common words, phrases, modifiers that appear repeatedly).
Deliverable: estimated monthly savings from the recommended negations.
Skill 6: Negative Keyword Discovery
Pull 60 days of search term data and identify terms that should be negated. Categorize by type:
- Irrelevant intent — informational queries (how-to, reviews, free, DIY)
- Wrong product/service — queries for things you don't sell
- Competitor brand terms — unless you're intentionally bidding on them
- Geographic mismatch — out-of-target locations
- Job-related — people looking for employment, not your product
- Waste patterns — common word stems that consistently don't convert
Organize into campaign-level vs. account-level negatives. Recommend exact match vs. phrase match for each. Prioritize by spend impact.
Pull all keywords with significant spend and categorize into four performance buckets:
- Stars — high volume, low CPA, high ROAS → increase bids and budget
- Opportunities — decent metrics with room for optimization → test match types, improve ad relevance
- Money pits — high spend, poor CPA or zero conversions → pause, restructure, or lower bids
- Long tail gold — low volume but excellent CPA → expand with similar variations
For each keyword, include: text, match type, impressions, clicks, CTR, CPC, conversions, CPA, ROAS, and quality score.
Section 3: Quality Score & Ad Relevance
Skill 8: Quality Score Audit
Pull quality score data for all keywords and analyze:
- Distribution across the 1–10 scale
- Impression-weighted average quality score
- Keywords with QS below 5 that have more than $20 in spend (last 30 days)
For each low-QS keyword, break down the three components — ad relevance, expected CTR, and landing page experience — and identify which is below average. Recommend specific fixes for each.
Key benchmark: each quality score point above average reduces CPC by approximately 16%. Estimate the CPC savings available from fixing the identified issues.
Analyze all active ads and identify:
- Top 5 by CTR — what headline structures are winning (question, benefit, number, urgency)?
- Top 5 by conversion rate — what messaging converts (price/discount vs. value-focused)?
- Bottom 5 by CTR — why aren't these getting clicked?
- Bottom 5 by conversion rate — these attract clicks but don't convert — why?
Look for patterns: which CTAs perform best, which headline structures consistently win, whether price-driven or value-driven messaging converts better.
Deliverable: 5 new ad copy variations based on the winning patterns found in your account data.
Section 4: Bidding & Budget
Skill 10: Budget Allocation Optimizer
For each campaign, pull: daily budget, actual spend, conversions, CPA, ROAS, impression share, and impression share lost to budget and rank.
Optimization logic:
- Shift budget from high-CPA campaigns to low-CPA campaigns
- Increase budget on campaigns losing impression share due to budget
- Reduce budget on campaigns with diminishing returns
- Keep total account spend the same (budget-neutral reallocation)
Deliverable: before/after comparison showing projected impact on total conversions and average CPA.
Skill 11: Bid Strategy Analysis
For each campaign, analyze: current bid strategy type (manual CPC, maximize conversions, target CPA, target ROAS, etc.), the target setting, actual performance against the target, and whether the strategy is meeting its goal.
Identify:
- Strategies missing their target by more than 20%
- Campaigns that should switch strategy types
- Targets that are too aggressive (limiting spend) or too loose (allowing waste)
Deliverable: specific strategy changes per campaign with projected performance impact.
Section 5: Audience & Targeting
Pull 30 days of data split by device (desktop, mobile, tablet) for each campaign. Report: spend, impressions, clicks, CTR, CPC, conversions, CPA, ROAS.
Identify:
- Which device converts best and at what CPA
- Mobile CPA more than 2x desktop (landing page problem or bid adjustment needed)
- Tablet spending significant budget with minimal conversions
Deliverable: specific device bid adjustments per campaign.
Pull the top 20 locations by spend with per-location metrics: impressions, clicks, CTR, CPC, conversions, CPA, ROAS.
Identify:
- Top 5 locations by CPA efficiency (scale these)
- Bottom 5 locations by CPA (reduce or exclude)
- High-spend, zero-conversion locations (exclude)
- Locations that justify bid adjustments
Deliverable: specific geographic bid adjustments and exclusion recommendations.
Skill 14: Audience Segment Analysis
Analyze 30 days of data across all audience segments: in-market, affinity, remarketing, customer match, and any custom audiences.
For each segment report: impressions, clicks, CTR, CPC, conversions, CPA, ROAS.
Answer:
- Which segments convert best?
- Which have high spend but poor CPA?
- How does remarketing compare to prospecting?
- Which segments should be added, excluded, or have bids adjusted?
Deliverable: audience bid adjustments and new targeting opportunities.
Pull 30 days of shopping performance broken down by: product, product type, brand, custom label, and category.
Identify:
- Top 10 products by ROAS (your hero SKUs)
- Bottom 10 by ROAS with significant spend (drag on account)
- High impressions but low CTR (listing quality problem — title, image, price)
- High CTR but low conversion rate (landing page or pricing problem)
- Zero-impression products (feed or bidding issue)
Recommendations: products to exclude, products that deserve dedicated campaigns, and feed optimization priorities.
Analyze the top 50 products by spend and evaluate each on:
- Title quality — is it descriptive with key attributes? (brand, product type, features, size, color)
- Description quality — does it include relevant search terms?
- Category accuracy — is the product mapped to the most specific Google category?
- Image quality — resolution, background, clarity
- Price competitiveness — how does it compare to competitors?
Title rewrite format: [Brand] + [Product Type] + [Key Attribute 1] + [Key Attribute 2] + [Size/Count]
Deliverable: specific title rewrites for the top 10 underperforming products.
Note: your Google Shopping feed is also what AI engines like ChatGPT and Perplexity pull from when recommending products — optimizing it for Shopping also optimizes it for agentic commerce.
Section 7: Competitor Analysis
Skill 17: Auction Insights Analyzer
Pull auction insights data for the top 5 campaigns by spend over the last 30 days.
Per competitor, analyze: impression share, overlap rate, position above rate, top-of-page rate, and outranking share.
Identify:
- Which competitors consistently outrank you
- Competitors gaining impression share (emerging threats)
- Whether you're losing share due to budget or rank
- New competitors entering the auction
Deliverable: specific bid increases, ad quality improvements, and budget adjustments to respond.
Skill 18: Competitor Ad Copy Research
Using the competitor domains from auction insights, research their current ad strategy:
- What headlines are they running?
- What offers or promotions are live?
- What CTAs do they use?
- Where do their ads land?
- How is their messaging differentiated from yours?
Gap analysis: what do competitors offer that your current ads don't address?
Deliverable: 5 specific ad copy improvements based on competitive gaps.
Section 8: Reporting & Briefs
Skill 19: Monthly Executive Summary
Compare this month vs. last month and produce a one-page executive summary:
- Headline result — one sentence on ROI direction and magnitude
- Metrics table — spend, conversions, CPA, ROAS, CTR — all with month-over-month change
- Two biggest performance drivers — what moved the needle most
- Budget efficiency score — percentage of spend going to converting vs. non-converting campaigns
- One specific recommendation with expected impact
Audience: founder, CMO, or client stakeholder. One page maximum.
Generate a full client report with 8 sections:
- Account overview with month-over-month trends
- Campaign-by-campaign summary
- Top wins — what worked and why
- Areas of concern — underperformance and your response
- Search term highlights — best new terms, worst negated terms
- Competitive landscape changes
- Recommendations and next steps
- Budget utilization summary
Tone: professional, confident, specific. No vague "we're continuing to optimize" statements — every observation needs a number and an action.
Section 9: Testing & Optimization
Skill 21: Next Test Recommender
Based on current account data, recommend the top 5 tests to run in the next 2 weeks. For each test, provide:
- What to test — ad copy, keywords, bid strategy, audience, or landing page
- Why it matters — the performance data supporting the hypothesis
- How to set it up — specific implementation steps
- Success definition — target metric and threshold for a winner
- Test duration — how long to run before evaluating
Prioritize by estimated impact on CPA or ROAS. Every recommendation must be data-driven and account-specific — no generic best practices.
Setup checklist
This is the same optimization stack we run for paid media clients at Fade Digital. If you want us to run it for you instead of doing it yourself, get in touch.
