Google Ads in 2026
For the person who keeps wasting money and wants to stop. Tracking, structure, bidding, keywords, PMax, AI Max — everything that matters.

Should You Even Be Here?
Google Ads captures demand — it doesn't create it
The 3-Question Filter
Do people actively search for what you sell?
No competitor ads for your keywords = no demand or no profitable way. Both bad. Check demand →
Can your margins absorb lead cost?
Avg CPC: $4.22 (legal: $9.21, dental: $6.69). 20 clicks per lead at $200/job = math fails.
Can you survive 6-8 weeks of learning losses?
One bad month threatens cash flow? Start with referrals and SEO.
Wrong For You If
Need demand creation, not capture
Offer is undifferentiated
Nobody answers the phone
Product is impulse/discovery
Consider Local Service Ads
Pay per lead, not click. Trust signals built in.
Best for: plumber, electrician, lawyer, dentist, HVAC. Many run both.
Conversion Tracking — Fix This First
Everything else depends on this
Automation is brutally literal. It scales whatever you tell it matters. If page views = conversions, it optimizes for page views.
What to Track as Primary
| Type | Primary | Secondary Only |
|---|---|---|
| E-commerce | Purchase | Add to cart, begin checkout |
| Lead gen | Qualified form submission | Page view, PDF download |
| Service biz | Phone call >60s | Contact page view |
| SaaS | Trial signup / demo | Pricing page view |
5 Tracking Mistakes That Burn Money
Double-counting
Ads tag + GA4 import both Primary. CPA looks half of reality. Use ONE source.
Low-value as Primary
'Contact page view' teaches algo to find browsers, not buyers.
Wrong counting
Leads: 'One' (per click). E-commerce: 'Every' (each purchase). Backwards = inflated.
Missing Conversion Linker
GTM Conversion Linker must fire on ALL pages. Without it, cross-domain breaks.
No offline feedback
Track 'form submitted' but never tell Google which became customers.
Enhanced Conversions for Leads (2026)
Hashed first-party data for better attribution. 5-15% more reported conversions. Settings > Enhanced Conversions > verify Active in Diagnostics. Target: 50%+ enhanced.
Campaign Structure
2-4 campaigns for $1K-$3K/month
Highest-intent commercial keywords.
Protects from competitors. Separate metrics.
After 1K+ visitors/month.
60+ conv/month, brand exclusions, strong creative.
The plumber test
Emergency calls: 40% conv, avg $300. Remodel leads: 5% conv, avg $15K. One CPA target = algo starves high-value remodels. Split by economics.
Search Ads Step by Step
RSAs, assets, good vs bad
Good Headlines (Plumber)
Emergency Plumber — Available Now
Licensed & Insured Since 2005
Same-Day Service Guaranteed
4.9 Stars on Google Reviews
No Call-Out Fee
Drain Cleaning Starting at $99
Each tests different value: urgency, trust, price, proof
Bad Headlines
Best Plumber in Denver
Denver's Top Plumber
#1 Plumber in Denver CO
Premium Plumbing Denver
Denver Plumbing Experts
Same thing 5 ways. Nothing to test. Ad strength: "Poor"
Essential Ad Assets
Service pages, pricing, reviews. Not homepage.
'Free Estimates' 'Licensed' 'Same-Day'
Clickable phone on mobile. Track as conversion.
Your work, team. No text overlays.
Services: Drain Cleaning, Water Heaters...
In-ad form. Test vs landing page.
Bidding Strategy
Graduate based on volume, not hope
New account, testing, tight budget
Clean tracking verified. Broken? Stay Manual.
Similar values. Start slightly above current CPA.
Varied values. Passing accurate value data.
The key rule
Don't graduate until signals are clean. More automation + bad inputs = faster waste. Learning: 7-14 days Search, 4-6 weeks PMax. No changes during learning.
Keywords & Match Types
Broad match truth
Same meaning. Tightest control, highest conv rate.
Includes meaning. Moderate control, adds context.
Related searches. Widest net, lowest control.
Strategy by Budget
Exact match top 10-15 keywords. Add phrase match winners.
Phrase default. Exact for top terms. Test broad on 2-3 with negatives.
Phrase core + broad expansion, Smart Bidding, weekly search term review.
Broad + Smart Bidding viable with solid tracking + offline feedback.
Negative keywords — most overlooked fundamental
Account-level (1,000 terms): free, cheap, DIY, how to, jobs, salary, careers, reddit, youtube. Weekly: Search Terms Report → irrelevant → add as negatives. Saves more money than any other optimization.
Performance Max
62% of clicks in Feb 2026
Works
E-commerce with product feed
Local biz with 60+ monthly conversions
Saturated branded + non-branded Search
High-quality creative (images, video)
Fails
First campaign in new account
No conversion quality signals
Budget under $1K/month
No brand exclusions (10-42% cannibalization)
Setup essentials
Brand exclusions first. 15 headlines, 5 descriptions, 7+ images, 1+ custom video (40% better than auto-generated). Upload customer list. Separate asset groups per service. Start Max Conversion Value (no ROAS target) for 4-6 weeks.
AI Max for Search
The 2025-2026 feature
Expands beyond keywords using broad match + keywordless tech
Gemini generates headlines from your landing pages
Sends users to most relevant page automatically
Set Text Guidelines BEFORE enabling
Term exclusions (25/campaign): words AI cannot use. Messaging restrictions (40/campaign): concept-level rules. Google claims 14-27% more conversions — their data, not independently verified.
Metrics That Matter
Stop obsessing over vanity numbers
Vanity Metrics
Impressions
Goes up with spend. Means nothing about business results.
Clicks
More clicks ≠ more customers.
CTR alone
High CTR + zero conversions = wrong people.
CPC alone
$2 click that never converts vs $10 click at 20% = $50/lead. $10 wins.
Real Metrics
Cost Per Conversion
Below max allowable CAC
Conversion Rate
3-5% Search. <2% = landing page problem
ROAS
Varies by margin. 3:1 = $3 per $1 spent
Search Terms Report
Weekly. Source of negative keywords
Cost Per Qualified Lead
Needs offline feedback. The metric that matters most
The report nobody reads
Search Terms Report shows actual queries triggering your ads. In 2026, Google hides many under "Other" — makes negatives harder but more important.
10 Mistakes That Waste Money
From most to least common
No conversion tracking (or wrong tracking)
Sending traffic to homepage instead of landing page
Too many campaigns with too little data
No negative keywords (silent budget killer)
Wrong match types for budget level
Ignoring mobile (5+ sec load = wasted clicks)
Set-and-forget — drift without weekly attention
Optimizing for lead volume not revenue
Changing everything during learning period
Running PMax too early without data or brand exclusions
Google Reps vs. Reality
Their KPI is your spend
"Switch to broad match"
Only works with mature measurement, strong negatives, 30+ monthly conversions.
"Turn on Performance Max"
Without data, tracking, brand exclusions, PMax burns money.
"Increase your budget"
'Limited by Budget' is a status, not a diagnosis. Fix targeting first.
"Apply all recommendations"
Optimization Score measures Google's adoption, not your performance.
"Enable auto-apply"
Gives Google permission to change your account without asking.
"Expand to Search Partners"
Quality varies wildly. Not appropriate for small budgets.
How to handle the call
Listen politely. Never change anything during the call. Ask: "Does this help me get paying customers, or primarily increase my spend?"
What Changed in 2026
Actual changes, not repackaged advice
| Change | Impact | Action |
|---|---|---|
| AI Max for Search | 14-27% lift claimed | Test on established campaigns with text guidelines |
| Text Guidelines | Brand control for AI copy | 25 term exclusions + 40 restrictions/campaign |
| Campaign Total Budgets | Fixed budget for flight | Use for seasonal or promotions |
| Budget Pacing Change | 30.4x daily/month | Monitor monthly spend carefully |
| PMax: 62% of clicks | Dominant campaign type | Not relevant for <$3K/month |
| Demand Gen absorbs VAC | One visual type | Discovery + Video Action merged |
Deprecated
Expanded Text Ads (Jun 2022)
Enhanced CPC (Mar 2025)
Call-Only Ads (Feb 2026→2027)
Discovery → Demand Gen
Video Action → Demand Gen
Still Works
Manual CPC (easier to find now)
Exact Match
Phrase Match
Standard Search Campaigns
Negative Keywords (1K acct-level)