Win the auction on Google Ads.
Google Ads runs ~80% of global PPC spend on top of 8.5 billion daily searches and a partner network that reaches the rest of the open web. Quality Score, Smart Bidding, and Performance Max have replaced manual bid management for almost every account that's still profitable. The Capconvert Ad Program for Google is built around how the auction actually clears in 2026 — not how it cleared in 2018.
How Google Ads decides whose ad runs.
Every Google query triggers a real-time auction. Eligible advertisers compete on Ad Rank — your bid multiplied by your Quality Score, plus a few minor signals (ad extensions, context, expected impact of formats). The highest Ad Rank that clears the auction floor wins the top slot. The next-highest takes the second slot. The rest don't show.
In practice, Quality Score has a larger impact on cost than your bid. A Quality Score of 9 versus 5 cuts your CPC roughly in half for the same position. Quality Score is the cheapest CPC lever in the entire account — and the work to move it (ad copy quality, landing-page experience, expected CTR) is invisible from the outside.
The stack behind every auction.
Google Ads doesn't run a single algorithm — it runs a stack of interconnected systems. Optimizing for the auction means optimizing for each layer.
Quality Score. A 1–10 score per keyword, derived from expected click-through rate, ad relevance, and landing-page experience. The single largest cost lever you have. Reported in the UI but reconstituted at auction time per query and per device.
Smart Bidding. The ML bidding stack (tCPA, tROAS, Maximize Conversions, Maximize Conversion Value). Trained on your account's first-party conversion data plus aggregate signals across Google's ad graph. Quality of the conversion signal feeding it is the load-bearing input — a cleaner pixel out-performs a tighter bid strategy.
Performance Max. Goal-based campaigns that span Search, Shopping, Display, YouTube, Discover, Gmail, and Maps in a single asset bundle. Now the default for most accounts. Strong on aggregate ROAS but weak on transparency and brand-term defense — both managed via account-structure discipline.
AI Max for Search. Newer (2025) campaign type that lets the model expand your keyword targeting and rewrite ad creative within guardrails. Most of the +60 AI improvements Google shipped in 2025 land here.
Demand Gen. Visual-led campaigns running across YouTube Shorts, Discover, and Gmail. Replaces the legacy Discovery campaigns and overlaps with Performance Max on inventory but with more creative control.
What drives Ad Rank, weighted.
Google publishes the Ad Rank formula but doesn't publish weights. Across thousands of accounts and tens of millions in tracked spend, the hierarchy below holds consistently. These are the levers we work on for every Google Ads engagement.
The lever most agencies ignore.
Quality Score is the single biggest variance generator between profitable Google Ads accounts and unprofitable ones running the same campaigns at the same bids. Two advertisers can target identical keywords with identical budgets and watch one pay $1.20 per click while the other pays $3.40 — Quality Score is the difference, every time.
Three components, weighted roughly equally: Expected CTR (Google's prediction of how often searchers will click your ad for this query), Ad Relevance (how closely the ad's content matches the keyword and the inferred intent), and Landing Page Experience (page speed, message match, mobile usability, content depth). Each component is reported on a 1–10 scale; the composite Quality Score is what drives auction discount.
The work that moves it is unglamorous: rewriting ads to match the actual query (not just the broad keyword), structuring ad groups tightly enough that one ad serves one intent cluster, fixing landing-page Core Web Vitals, deleting RSAs that drag the average down, and writing ad copy that earns clicks rather than chases volume. None of it shows up in the dashboard until the next auction cycle. All of it compounds.
What the ML actually rewards in ad assets.
Modern Google Ads creative isn't a static ad. It's an asset bundle the ML assembles per query and audience — Responsive Search Ads pull headlines and descriptions from a pool, Performance Max blends headlines with images, video, and descriptions across surfaces. The optimization unit is the asset, not the ad.
RSA pinning discipline. Most accounts over-pin (locking down every headline) or under-pin (letting the ML serve combinations that lose Quality Score). The right pattern is pinning brand and offer headlines to position 1, leaving 8–12 unpinned headlines for the ML to test, and writing every description as if it could appear without the others.
Asset density. Google rewards full asset bundles — 15 headlines, 4 descriptions, 4 logos, 5+ images, 1+ video for Performance Max. Sparse bundles ship lower Quality Score and limit campaign reach. We default to filling every slot.
Brand-safety markers. AI-generated copy without disclosure, claims that fail Google's policy filter, missing trademark approvals — all silent rejections that cap campaign delivery without obvious error messages. We audit the policy stack on every onboarding.
Clean conversion signal is the load-bearing input.
Smart Bidding is only as good as the conversion signal it's optimizing against. Most underperforming Google Ads accounts have the same root cause: dirty, incomplete, or misattributed conversion data. The auction layer is fine. The signal feeding it is broken.
Server-side tagging. Browser-side pixels lose 15–35% of conversions to ad blockers, ITP, and consent rejections. Server-side tagging via Google Tag Manager Server or Stape captures conversions browser-side trackers miss and feeds them back to Google with full attribution.
Enhanced Conversions. First-party data (email, phone, address) hashed and matched to Google's user graph. Recovers attribution that pure cookie tracking loses. Now table-stakes — accounts running without it are bidding blind in iOS-heavy verticals.
Conversion-value rules. Different conversion actions (purchase, lead, demo request) deserve different values. Smart Bidding optimizes against the value signal — feeding it constant $1 across every conversion action is the most common rookie mistake we surface in audits.
Customer-list re-ingestion. Customer Match lists feed back into audience signals, lookalikes, and exclusion targeting. Refreshed weekly, they become a meaningful Smart Bidding input over a quarter.
How we run Google Ads for clients.
Every Google Ads engagement runs through the same five-step methodology, refined across the $103M+ of ad spend Capconvert manages annually. We don't reinvent the playbook per engagement — we reinvent the inputs.
Account audit. Quality Score gap analysis, conversion signal health check, Performance Max transparency review, search-term harvest, account-structure assessment. Output is a remediation roadmap with projected CPC delta per fix.
Tracking foundation. Server-side tagging deployed, Enhanced Conversions wired in, conversion values calibrated, Customer Match feeds connected. The signal layer comes online before bidding optimization touches anything.
Campaign architecture. Brand defense isolated, branded vs. non-branded segmentation enforced, Performance Max scoped with asset groups by buyer intent, Search campaigns for high-CPC priority terms outside PMax, Shopping segmented by margin tier.
Creative system. Full RSA bundles deployed (15 headlines, 4 descriptions), Performance Max asset groups populated to capacity, A/B/n testing wired into the publish cadence, quarterly creative refreshes on volume accounts.
Bid & budget management. tROAS or tCPA targets calibrated to LTV, budget pacing reviewed weekly, search-term reports actioned, audience signals refreshed monthly. Smart Bidding does the per-auction work; we do the strategy work.