Google Ads Optimization Playbook

A practitioner-level guide to advanced Google Ads management, covering the techniques that separate efficient, ROI-positive accounts from bloated spend with poor returns.

---

1. Keyword Diagnosis Before Campaign Launch

Ad Preview and Diagnosis Tool

Before adding any keyword to a live campaign, test it using the Ad Preview and Diagnosis Tool (found under Tools and Settings then Planning).

What this tells you:

• Whether your ad is actually showing for the keyword
• What competitor ads appear for that term
• Whether the keyword matches actual search intent (the results page reveals what Google thinks users want when they search this term)
• Geographic variations in ad display

Why this matters: Many advertisers add keywords based on volume data alone without checking whether the search results page matches their intent. A keyword with 10,000 monthly searches is worthless if Google interprets the intent as informational and your landing page is transactional.

Process:

1. Enter the keyword in the Ad Preview tool
2. Set the geographic location to your target market
3. Set the device type (mobile vs desktop)
4. Review the organic and paid results that appear
5. Ask: "Would someone searching this term genuinely want what I am offering?"
6. If the results page is dominated by informational content (blog posts, Wikipedia, how-to guides), your commercial landing page will have poor Quality Score and high CPC

Keyword Intent Classification

Classify every keyword before adding it:

Intent Type	Signals	Expected CPC	Conversion Likelihood
Navigational	Brand names, specific products	Low-Medium	High (if your brand)
Informational	How to, what is, guide, tutorial	Low	Low
Commercial Investigation	Best, review, comparison, vs	Medium	Medium-High
Transactional	Buy, pricing, demo, free trial	High	High

Focus budget on commercial investigation and transactional keywords. Use informational keywords only in separate campaigns with content-focused landing pages and lower bids.

---

2. Quality Score Engineering

The 1 Ad Group Per Keyword Structure

For campaigns with fewer than 50 keywords, the single keyword ad group (SKAG) structure is the most reliable way to engineer high Quality Scores from launch.

How it works:

• Create one ad group per keyword
• Name the ad group exactly matching the keyword
• Write ad headlines that include the exact keyword
• Direct to a landing page with the keyword in the H1 and URL

Why this works: Quality Score is composed of three factors:

1. Expected CTR (historical performance of the keyword)
2. Ad relevance (how closely your ad matches the keyword)
3. Landing page experience (how relevant and useful the landing page is)

When your ad group contains a single keyword, and your ad headline includes that exact keyword, you achieve near-perfect ad relevance. This virtually guarantees a starting Quality Score of 4-6 out of 10, which is above the median and means you pay less per click from day one.

Naming convention:

• Keyword: b2b marketing software
• Ad group name: b2b marketing software
• Ad headline 1: B2B Marketing Software
• Landing page H1: B2B Marketing Software for Growing Teams
• Landing page URL: /b2b-marketing-software

When NOT to use SKAGs: If you have 200+ keywords, the management overhead becomes unsustainable. In that case, group tightly related keywords (3-5 per ad group) by theme and use dynamic keyword insertion.

Quality Score Improvement Levers

Quality Score	Action
1-3	Pause keyword -- landing page or intent mismatch is severe
4-5	Improve ad relevance and landing page alignment
6-7	Good baseline -- optimize CTR with ad copy testing
8-10	Excellent -- scale spend on these keywords

---

3. Dynamic Keyword Insertion

On Ads

Dynamic Keyword Insertion (DKI) automatically inserts the user's search term into your ad headline or description.

Syntax: {KeyWord:Default Text}

• {KeyWord:} capitalizes Each Word
• {Keyword:} capitalizes First word only
• {keyword:} no capitalization
• {KEYWORD:} ALL CAPS

Example:

• Headline: Best {KeyWord:Marketing Software}
• User searches "email automation tool"
• Ad displays: Best Email Automation Tool

Guardrails:

• Always set a sensible default text in case the search term is too long (30 character headline limit, 90 character description limit)
• Review your search terms report weekly to ensure DKI is not inserting irrelevant or embarrassing terms
• Do not use DKI with broad match keywords -- the inserted terms will be too unpredictable

On Landing Pages

This is where most advertisers miss a significant optimization. Use URL parameters to dynamically change landing page headlines to match the keyword that triggered the ad.

Implementation:

1. Append {keyword} parameter to your final URL: example.com/solution?kw={keyword}
2. Use JavaScript on the landing page to read the URL parameter and replace the H1 text
3. The landing page headline now matches the exact search term

Impact: This improves Quality Score's landing page experience component and increases on-page conversion rates by 15-30% in testing, because the visitor sees exact continuity from their search query through the ad to the landing page.

---

4. GCLID Offline Conversion Tracking

Why This Matters for B2B

Most B2B conversions do not happen online. A form submission is not a sale -- it is a lead. The actual conversion (closed deal, signed contract) happens weeks or months later in a CRM. Without offline conversion tracking, Google Ads optimizes for form fills, not revenue.

Implementation Steps

1. Enable auto-tagging in Google Ads settings (this appends the GCLID parameter to your URLs automatically)
2. Capture the GCLID on your landing page using a hidden form field -- when the form is submitted, the GCLID is stored alongside the lead in your CRM
3. Map conversions in your CRM -- when a lead closes as won, export the GCLID with the conversion value and date
4. Upload to Google Ads via the Conversions section (manual upload or automated via API)

Automation option: Connect your CRM to Google Ads using Zapier or a native integration to automatically push offline conversions when deal stages change. This keeps your data fresh without manual exports.

Using Offline Conversions for Bidding

Once you have 90+ days of offline conversion data uploaded, switch your bid strategy to Target ROAS or Maximize Conversion Value. The algorithm will now optimize for actual revenue, not just form submissions. This single change can transform campaign ROI because the algorithm learns which keyword-audience-time combinations produce real customers, not just leads.

---

5. Target ROAS Setup and Calculation

Calculating Your Target ROAS

Formula: Target ROAS = (Revenue from Ads / Cost of Ads) x 100

Example:

• Average deal value: $50,000
• Close rate from Google Ads leads: 10%
• Expected revenue per lead: $5,000
• Target CPA (what you can afford per lead): $500
• Target ROAS: ($5,000 / $500) x 100 = 1,000%

Implementation Guidelines

• Do not set Target ROAS immediately on a new campaign. Run with Maximize Clicks or Maximize Conversions for 4-6 weeks to build conversion history.
• Minimum conversion volume: You need at least 15 conversions in the last 30 days before Target ROAS bidding can function. For offline conversions, this means waiting until enough data has been uploaded.
• Start conservative: Set your initial Target ROAS 20% below your actual target. If your real target is 1,000%, set it to 800%. This gives the algorithm room to learn without immediately restricting delivery.
• Adjust gradually: Increase Target ROAS by 10-20% increments every 2 weeks. Aggressive jumps cause the algorithm to severely restrict impressions.

CPA Simulator

Use the CPA Simulator (found in keyword-level columns) to model how raising or lowering your target CPA affects projected conversions and spend. This helps you find the efficiency frontier -- the point where increasing CPA no longer produces proportional conversion gains.

How to access: Add the "CPA Simulator" columns to your keyword view. It shows estimated conversions at various CPA levels so you can identify diminishing returns.

---

6. Audience Strategies

Customer Match

Upload your contact list (emails, phone numbers, mailing addresses) to target known contacts across Google Search, Gmail, YouTube, and Display.

Use cases:

• Branded awareness: Show ads to your CRM contacts when they search for your brand or competitors
• Upsell campaigns: Target existing customers with upgrade offers on YouTube and Gmail
• Exclusion: Exclude existing customers from acquisition campaigns to avoid wasted spend
• Similar audiences: Google automatically creates "similar to Customer Match" audiences you can target

Requirements:

• Account must have a good history of policy compliance
• Minimum 1,000 active members in the list
• Lists must be updated at least every 90 days

Custom Affinity Audiences

Build audiences based on interests, URLs, and apps your target customers engage with.

For B2B, define custom affinity audiences using:

• URLs of industry publications your buyers read
• URLs of competitor websites
• Keywords related to your buyers' professional interests
• Apps your target audience uses

Competitor Keyword Targeting

Bidding on competitor brand keywords is a legitimate and often highly profitable strategy.

Advanced tactic: When targeting competitor keywords, exclude geographic areas where the competitor has physical offices. Employees at those offices frequently search their own company name, and they are not prospects. They inflate your impression count without converting. Most competitors neglect to check Auction Insights, so they will not notice you are bidding on their terms.

Exclusion method:

1. Identify competitor office locations
2. Create location exclusions for those cities or zip codes
3. This eliminates employee searches and focuses spend on actual prospects evaluating the competitor

Gmail Ads with Competitor Targeting

Gmail ads appear in the Promotions tab of Gmail. Target users who receive emails from competitors by using competitor domain names as keywords in a custom audience.

Setup:

1. Create a Display campaign (Gmail ads are now part of Display/Discovery)
2. Use custom segments with competitor URLs and keywords
3. Write ad copy that positions against competitor pain points
4. Link to a comparison or switching page

---

7. Campaign Structure and Strategy

Performance Max Campaigns

Performance Max (PMax) campaigns use automation to deliver ads across all Google surfaces (Search, Display, YouTube, Gmail, Maps, Discover).

When to use PMax:

• Ecommerce with product feeds (strongest use case)
• Lead generation when you have strong offline conversion data feeding back
• When you want incremental reach beyond Search

When to avoid PMax:

• New accounts with no conversion history
• When you need granular keyword-level control
• When your conversion tracking is incomplete (PMax with bad data optimizes toward bad outcomes)

PMax best practices:

• Provide the highest quality creative assets possible (images, videos, headlines, descriptions)
• Create separate asset groups for distinct audience segments
• Add audience signals to guide the algorithm (Customer Match lists, custom segments)
• Run PMax alongside standard Search campaigns -- PMax does not fully replace Search for high-intent keywords

Attribution Model Selection

Available models:

• Last click (default, but least informative)
• First click
• Linear
• Time decay
• Position-based
• Data-driven (recommended when you have enough conversion volume)

Recommendation: Switch to data-driven attribution as soon as your account qualifies (requires 300+ conversions and 3,000+ ad interactions in the last 30 days). Data-driven attribution uses machine learning to distribute credit based on actual conversion path analysis, giving you a more accurate picture of which keywords and campaigns truly drive results.

The Bid Strategy Alignment Fix

Symptom: Impressions suddenly drop despite no changes to budget, keywords, or targeting.

Diagnosis: Check whether your Marketing Objective (set at the campaign level) aligns with your bid strategy. A mismatch causes the algorithm to restrict delivery.

Common mismatch: Campaign objective set to "Leads" but bid strategy set to "Maximize Clicks." The system detects conflicting signals and throttles delivery.

Fix:

1. Go to Campaign Settings
2. Verify the campaign objective
3. Ensure the bid strategy matches (Leads objective should use Maximize Conversions or Target CPA, not Maximize Clicks)
4. If you changed bid strategies recently, allow 7-14 days for the learning phase to stabilize before making further changes

---

8. Campaign Experiments

A/B Testing Ad Strategies

Google Ads Experiments let you split traffic between your original campaign and a test variant.

Setup:

1. Go to the campaign you want to test
2. Click Experiments then Create Experiment
3. Choose the variable to test (bid strategy, ad copy, landing page, audience)
4. Set the traffic split (50/50 for fastest statistical significance)
5. Set the experiment duration (minimum 2-4 weeks, ideally until you reach 100+ conversions per variant)
6. Monitor results and apply the winner

What to test:

• Bid strategy changes (Manual CPC vs Target CPA vs Maximize Conversions)
• Landing page variations
• Ad copy themes (benefit-driven vs feature-driven vs social proof)
• Audience targeting additions
• Match type changes

Location of Interest Targeting

Default setting: Google targets people "in or who show interest in" your target locations. This means someone in another country researching your target city will see your ads.

For local businesses: Change to "Presence: People in or regularly in your target locations" to avoid irrelevant clicks.

Radius experimentation: Test different radius sizes around your target locations:

• Start with a wider radius (50+ miles) to gather data
• Analyze conversion rates by distance
• Narrow the radius to the distance bands that convert profitably
• Layer bid adjustments by distance (increase bids for closer, decrease for further)

---

9. Automated Rules and Scripts

Essential Automated Rules

Set up these rules to catch problems early:

1. Pause keywords with 0 conversions and 200+ clicks (check weekly)
2. Increase budget by 15% when campaign is limited by budget and CPA is below target (check daily)
3. Pause ads with CTR below 1% after 1,000+ impressions (check weekly)
4. Alert when daily spend exceeds 150% of daily budget (check every 6 hours)
5. Pause keywords with Quality Score below 3 (check weekly)

Useful Scripts

• Search terms mining script: Automatically reviews search terms and adds high-performing queries as exact match keywords
• Bid adjustment by hour of day: Analyzes conversion data by hour and adjusts bids automatically
• Budget pacing script: Ensures monthly budget is spent evenly rather than front-loaded
• Broken URL checker: Flags ads pointing to pages returning 404 or 500 errors

---

10. Connecting Paid and Organic

Using Top-Converting Paid Keywords for SEO

Your Google Ads search terms report is a goldmine for organic keyword strategy.

Process:

1. Export your search terms report for the last 90 days
2. Filter for terms with conversion rate above your account average
3. Check current organic rankings for these terms
4. Prioritize terms where you rank on page 2-3 (close to page 1 but not there yet)
5. Create or optimize content targeting these terms
6. As organic rankings improve, reduce paid spend on those terms

This creates a flywheel: paid validates which keywords convert, organic captures that traffic for free, paid budget shifts to new test keywords.

Retargeting with Discount Offers

For ecommerce or SaaS with free trials, retarget website visitors who did not convert with a discount or extended trial offer.

Structure:

• Audience: Website visitors in the last 14 days who did not convert
• Exclude: Users who visited fewer than 2 pages (likely accidental clicks)
• Creative: Lead with the discount/offer, include urgency (limited time)
• Frequency cap: 3-5 impressions per user per day across Display
• Duration: 14-30 days (longer retargeting windows have diminishing returns)

---

11. Taking Over Underperforming Accounts

The Agency Audit Pattern

A common pattern: an agency running campaigns at high spend with high CPL and unqualified leads. The typical fixes when taking over an underperforming account:

1. Audit search terms report -- often reveals significant spend going to irrelevant queries
2. Add negative keywords aggressively -- build a comprehensive negative keyword list from wasted spend
3. Restructure match types -- agencies often over-rely on broad match for volume
4. Fix conversion tracking -- frequently misconfigured (counting page views as conversions, double-counting, missing offline conversions)
5. Reduce geographic targeting -- agencies often target too broadly to inflate impression numbers
6. Rebuild landing pages -- agencies rarely optimize landing pages because it is outside their scope

Typical result: Reducing wasted spend while producing better results with positive ROI by eliminating waste and focusing budget on the keywords and audiences that actually convert. High spend often masks inefficiency — more spend does not equal more results when targeting and tracking are broken.

---

12. Keyword Research Deep Dive

Keyword research is not a one-time exercise you do before launching campaigns. It is an ongoing process that should inform budget allocation, content creation, and competitive strategy across both paid and organic channels.

Seasonal Search Volume Analysis

Every market has seasonality. Ignoring it means you either overspend during slow periods or underspend when demand peaks.

How to identify seasonal patterns:

1. Run your core keywords through Google Trends to see 12-month and 5-year patterns
2. Look for predictable spikes — holiday periods, industry events, fiscal year timelines, back-to-school, tax season, budget planning cycles
3. Note the ramp-up period before each spike. Demand does not spike overnight — it builds over 2-4 weeks before the peak

How to capitalize on seasonality:

• Increase budgets 2-3 weeks before a predicted demand spike. If you wait until demand peaks, competitors have already claimed the impression share.
• Pre-build seasonal ad copy and landing pages so they are ready to deploy immediately. Rushing creative during peak demand leads to mediocre performance.
• Some markets see 3-5x normal search volume during peak periods. The CPCs may increase, but so does conversion intent. A keyword that costs $5 during off-season might cost $12 during peak — but if conversion rates also double, the peak CPC is actually more efficient.
• Reduce budgets 1-2 weeks after the peak passes. Trailing demand converts at lower rates because the urgency is gone.

Negative Keyword Discovery Process

The most expensive mistake in Google Ads is not failing to find good keywords — it is failing to exclude bad ones. Negative keywords should be identified during research, not after you have wasted budget.

Proactive negative keyword research:

1. When building your keyword list, pay attention to related terms that show high volume but irrelevant intent
2. Common negative keyword categories that apply to almost every B2B account: customer service, jobs, careers, hiring, cancel, cancellation, phone number, help desk, login, sign in, down, outage, free (if you are not running free offers)
3. Add these as campaign-level negative keywords before the first ad runs
4. For ecommerce: add competitor brand names you do not want to appear for, product names you do not carry, and service terms you do not offer

Ongoing negative keyword maintenance:

• Review the Search Terms Report weekly for the first 3 months of any campaign
• Move to bi-weekly review once the account is mature
• Export converting and non-converting search terms separately — terms that drive impressions but never convert are negative keyword candidates
• Build a shared negative keyword list that applies across all campaigns to prevent wasted spend on known irrelevant terms

Competitor Keyword Categories

When researching keywords in a competitive market, organize your findings into categories based on who or what the searcher is looking for. Each category has different economics.

Category mapping:

Category	Example Keywords	Typical CPC	Intent Level
Brand terms	Your own brand name, misspellings	Low	Very High
Competitor terms	Competitor brand names, "vs" queries	Medium-High	High
Category terms	Generic product/service category	High	Medium-High
Platform/aggregator terms	Marketplace or directory names	Medium	Medium
Creator/influencer terms	People associated with your space	Low	Low-Medium
Informational terms	How-to, what is, guide, tutorial	Low	Low

Why this categorization matters: Budget should flow to categories with the best ROI, not the highest volume. Brand terms convert at the highest rate but have limited volume. Category terms have the most volume but the highest competition. Competitor terms sit in a sweet spot — high intent (they already understand the category) with competition limited to you and the named competitor.

Test before committing: Run small-budget tests across all categories for 2-4 weeks. Measure cost per conversion, not just CPC or CTR. A $15 CPC keyword that converts at 8% is more valuable than a $3 CPC keyword that converts at 0.5%.

CPC as an SEO Difficulty Proxy

The cost-per-click advertisers pay for a keyword in Google Ads is a reliable indicator of how competitive the organic landscape is for that same keyword.

The logic: Keywords with high CPCs attract that level of spending because they convert well and generate revenue. The same commercial intent that makes a keyword expensive in paid search makes it attractive to organic competitors. More companies competing organically means more content, more backlinks, and more authority concentrated on page 1.

How to use this insight:

• When planning organic content, sort your keyword opportunities by estimated CPC
• Prioritize creating content for keywords where CPC is moderate (commercially valuable but not dominated by deep-pocketed competitors)
• Very high CPC keywords may require 6-12 months of dedicated SEO effort to crack page 1. Factor this into your content calendar.
• Very low CPC keywords often have low commercial intent — they may drive traffic but not conversions. Create this content for brand awareness, not lead generation.

Budget Forecasting from Keyword Data

Keyword research data enables realistic budget planning rather than arbitrary spend targets.

Building a 6-month forecast:

1. List your target keywords with monthly search volumes and estimated CPCs
2. Estimate your expected click-through rate by position (position 1-3: 5-15%, position 4-7: 2-5%)
3. Apply your historical conversion rate to estimate monthly conversions per keyword
4. Multiply estimated clicks by CPC to get monthly spend per keyword
5. Sum across all keywords for total monthly budget requirement
6. Build three scenarios: conservative (lower impression share), moderate (target impression share), aggressive (maximum impression share)

Account for people costs: Even if the founder or a team member manages the campaigns themselves, that time has value. Failing to account for the hours spent on campaign management biases your ROI calculations. A channel that looks "free" because nobody is paying an agency may actually have a higher effective cost than a channel where you pay for managed service but spend zero internal hours.

Track time by channel: Log hours spent managing each marketing channel monthly. When you combine ad spend plus time cost, the true channel ROI often looks very different from the ad-spend-only calculation. This data prevents the common mistake of over-investing in "cheap" channels that consume enormous amounts of time.

Connecting Google Ads to Claude Code via MCP

The Model Context Protocol (MCP) lets you connect your Google Ads account directly to Claude Code, turning ad optimization into a conversation instead of a CSV export workflow. Once configured, Claude can read account data, run queries against your campaigns, and surface recommendations in the same session where you discuss strategy.

How to connect:

• Install an MCP server that supports the Google Ads API (community servers exist; some require a developer token and OAuth credentials from your Google Ads manager account)
• Register the server in your Claude Code configuration so it is available as a data source in your sessions
• Scope access to the specific accounts you want Claude to see — avoid granting broader permissions than necessary

What this unlocks in practice:

• Search terms analysis: Ask Claude to pull the last 30 days of search terms across a campaign, cluster them by intent, and flag high-spend terms with zero conversions as negative keyword candidates
• Keyword performance review: Have Claude rank keywords by cost per conversion, highlight ones drifting above your target CPA, and suggest bid adjustments or match-type changes
• Wasted spend audit: Request a weekly scan for auto-applied recommendations, low-quality-score keywords, and placements on irrelevant sites — then discuss which ones to disable
• Campaign-level diagnostics: Ask "why did conversions drop this week?" and let Claude correlate impression share, auction insights, quality score shifts, and search volume changes
• Optimization backlog: Turn the findings into a prioritized action list you can execute directly, with Claude drafting the negative keyword lists, ad copy variations, or bid adjustments inline

The point is not to automate decisions — it is to compress the loop between asking a question about your account and having the data in front of you. What used to take a 20-minute CSV pull and a pivot table becomes a follow-up question.