Google Ads операционна система

Google Ads Master Operating System

1. Основната логика

Google Ads account не се строи около campaign types. Строи се около business controls, measurement truth и ясни причини защо всеки campaign съществува.

Редът е:

1. Определи икономиката: допустим CAC/CPA, contribution margin, target ROAS, lead-to-sale rate и стойност на customer.
2. Подреди conversion truth: кое е primary, кое е secondary, каква стойност се изпраща и има ли duplicate tracking.
3. Избери controllable levers: product/service, market, margin, season, customer type, location, device или business objective.
4. Изгради минималната campaign structure, която дава тези levers без да раздробява data.
5. Пусни bottom-of-funnel intent преди upper-funnel expansion.
6. Оптимизирай по S.T.A.B.: Spending and Segmentation, Targeting, Ads and Landing Pages, Bidding.
7. Скалирай само когато conversion volume и trend са стабилни.

Course basis: eCommerce Playbook pages 2-7 and 29-42; Lead Generation Playbook pages 2-8 and 27-42; MGA-20 to MGA-51.

2. Решение за всеки нов campaign

Не създавай campaign, докато няма еднозначен отговор на следните въпроси:

• Каква бизнес функция има?
• Кой budget или bid lever създава?
• Каква аудитория, location, product set или query intent го отличава?
• Коя primary conversion action управлява bidding-а?
• Как ще измерим incrementality спрямо съществуващите campaigns?
• Какво ще спрем, ограничим или изключим, за да няма conflicting optimization?

Ако единствената причина е "да пробваме нещо ново", структурата е слаба.

3. Account architecture

Manager level

• Един company-owned Google Ads Manager account.
• Отделен serving account за различна legal entity, billing currency, time zone или фундаментално различен business model.
• Един account за един market, когато market-ът изисква отделна валута, billing, language operations, feed, legal rules или независим P&L.
• Не раздробявай само защото има много продукти. В един account product families могат да бъдат разделени с campaigns, listing groups, custom labels и asset groups.

Serving account level

• Brand Search е отделен campaign.
• Non-brand Search е отделен по intent и control need, не по всяка отделна keyword variation.
• Shopping/PMax inventory не трябва да има неволно duplicate product targeting със conflicting targets.
• PMax brand traffic се управлява с campaign-level brand exclusions и/или campaign-level negative keywords. Account-level negatives се използват само когато терминът трябва да бъде блокиран в целия account.
• Promotional campaigns са отделни от always-on campaigns, когато промоцията изисква временно различен budget, assets или bidding posture.
• Всеки campaign има един owner, една primary business goal family и записан purpose.

4. Funnel order

eCommerce

1. Search + Shopping за high-intent demand.
2. Separate Brand Search и remarketing, когато има достатъчно audience volume.
3. PMax за new-customer growth и broader inventory след валидно tracking и baseline profitability.
4. Demand Gen/Video за expansion, когато creative supply, budget и measurement могат да го поддържат.

B2B/lead generation

1. Search по high-intent service/problem queries.
2. Brand Search, competitor protection и remarketing.
3. Enhanced conversions for leads + offline qualified lead/sale import.
4. PMax само след надеждна downstream conversion feedback loop.
5. Demand Gen/Video за category education и lead magnet, когато CRM follow-up е готов.

Course basis: eCommerce Playbook pages 12-15 and 31-35; Lead Generation Playbook pages 17-20 and 30-33.

5. Measurement hierarchy

eCommerce

• Primary: purchase with accurate value and currency.
• Secondary: add to cart, begin checkout, view item, newsletter signup.
• Diagnostics: transaction ID coverage, duplicate rate, Ads vs backend revenue variance, enhanced conversion status, consent status.

B2B

• Primary for mature account: qualified lead, sales-qualified lead or closed sale with value.
• Temporary primary for cold start: lead form or qualified call, only until downstream data volume becomes usable.
• Secondary: page view, engaged session, raw contact, download, unqualified call.
• Use a documented conversion hierarchy if exact revenue cannot yet be imported.

Google's account-default goals and primary actions feed the Conversions column and automated bidding. Secondary actions remain observable in All conversions unless deliberately included in an advanced custom goal.

Current source: https://support.google.com/google-ads/answer/4677036

6. Data maturity gates

Gate 0: no trusted data

• No scale decisions.
• Fix tags, deduplication, values, consent, call tracking and CRM IDs.

Gate 1: trusted but low-volume data

• Tight intent and controlled budgets.
• Search terms, negatives, feed quality, ad relevance and landing page work.
• Avoid aggressive tCPA/tROAS constraints.

Gate 2: stable conversion volume

• Course guardrail: at least 30 primary conversions in the last month, weekly volume within roughly 20% variance, no recent bid target change in 6 weeks and no budget change in 2 weeks.
• Treat this as a conservative operating heuristic, not a Google platform rule.

Gate 3: scalable

• Performance holds across multiple windows.
• Search impression share or eligible inventory shows room.
• Marginal CAC/ROAS remains inside business limits.
• Budget can rise without starving higher-priority campaigns.

Course basis: MGA-51, ALL-53, eCommerce Playbook pages 18-22, Lead Generation Playbook pages 21-25.

7. Change discipline

• One meaningful hypothesis per change set.
• Log timestamp, old value, new value, reason, expected effect and evaluation window.
• Do not change budget and bid target simultaneously unless there is an emergency constraint.
• Preserve winners. Restructure dormant or messy accounts in stages.
• For a live converting account, test architecture changes first on mid-tier products/services, not the biggest revenue driver.
• Do not judge an automated campaign from one or two days.

8. Evidence labels used in this system

• COURSE: direct course tactic, benchmark or opinion.
• GOOGLE-CURRENT: verified against current official Google documentation.
• OPERATOR: recommended control rule derived from the course, current platform and business risk.
• ACCOUNT-OBSERVED: measured in the live account/export.
• USER-CONTEXT: reported by the owner but not yet verified in account data.

9. Current 2026 platform delta

• AI Max is an optimization layer inside Search, not a new campaign type. It combines expanded search term matching, asset optimization and Final URL expansion controls.
• Campaign-level negative keywords are available for PMax Search and Shopping inventory.
• PMax now has a channel performance report with cost, conversion and value metrics by channel.
• Demand Gen supports channel controls across YouTube, Discover, Gmail, Maps and Google Display Network.
• Enhanced conversions for web and leads are being unified at account level from April 2026.
• Smart Bidding labels are changing from June 2026, while the underlying behavior stays the same.

Detailed citations: manuals/09-current-platform-delta-2026.md.

Google Account and Ads Account Governance

1. Ownership model

Required structure

• Company-owned Google identity for break-glass administration.
• Named individual users for daily work. Never share one login among several people.
• Company-owned Google Ads Manager account above all serving accounts.
• Serving accounts remain owned by the business, even when an agency or contractor is linked.
• Agency access is through a linked manager, not by handing over the owner login.

Access levels

• Admin: maximum two or three trusted internal owners.
• Standard: operators who create and edit campaigns.
• Read only: analysts, auditors and finance reviewers.
• Email only: notification recipients without data access.
• Remove departed or inactive users immediately.

Google recommends individual access, minimum necessary permissions, manager accounts for multiple accounts, and regular review of related managers.

Current source: https://support.google.com/google-ads/answer/12864492

2. Security baseline

• Enforce 2-Step Verification from the Manager account.
• Restrict allowed email domains when all operators use controlled company domains.
• Review direct users and related managers monthly.
• Keep a recovery method that is not tied to one employee phone.
• Record who can change billing, user access, conversion goals, scripts and linked products.
• Review Change History when spend, ads or access changes unexpectedly.

Manager security mandates can enforce authentication strength and allowed domains across owned sub-accounts.

Current source: https://support.google.com/google-ads/answer/12865295

3. When to create a new serving account

Create a separate account when one or more are true:

• Different legal entity or payment profile.
• Different currency.
• Different reporting time zone.
• Independent P&L and budget authority.
• Regulatory separation is required.
• A market needs isolated conversion goals, feed, data residency process or ownership.

Do not create a new account only because:

• there is a new product category;
• you need a separate campaign budget;
• Search and PMax must be reported separately;
• a new landing page or promotion is launching.

Currency and serving-account time zone are permanently set. Fixing either requires a new account, and campaign copy does not transfer historical data.

Current source: https://support.google.com/google-ads/answer/9842104

4. Account creation checklist

• Legal business name and billing country checked.
• Currency chosen from the P&L currency, not personal convenience.
• Time zone chosen from the operating/reporting day.
• Auto-apply recommendations reviewed and disabled by default until individually approved.
• Payment method and backup method validated.
• Account linked to the correct Manager account.
• GA4, Merchant Center, YouTube, Search Console and Business Profile linked only where needed.
• Data sharing, audience sharing and ownership documented.
• Account-default conversion goals reviewed before the first campaign goes live.
• Naming convention installed.

5. Naming convention

Use names that remain readable in exports:

{market}_{business-line}_{funnel}_{campaign-type}_{intent-or-product}_{goal}_{version}

Examples:

• BG_VETFAMILY_BOF_SEARCH_NONBRAND_ARTHRO_PURCHASE_V1
• BG_VETFAMILY_BOF_SEARCH_BRAND_ALL_PURCHASE_V1
• BG_VETFAMILY_MOF_PMAX_ARTHRO_NEWCUSTOMER_V1
• DE_B2B_BOF_SEARCH_HIGHINTENT_DEMO_V1
• EU_SAAS_TOF_DGEN_PROBLEM-AWARE_MQL_V1

Asset groups/ad groups:

{theme}_{audience-or-product}_{landing-page}_{version}

6. Shared libraries and exclusions

• Account-level negative keywords: universal irrelevance only. They apply across Search and Shopping inventory in Search, PMax, Shopping, App, Smart and Local campaigns.
• Campaign-level negatives: local campaign intent control.
• Brand exclusions: brand entities and misspellings, especially for keeping PMax away from own-brand demand.
• Placement exclusions and content suitability: maintain at the narrowest level that solves the risk without blocking valid inventory.
• Shared budgets: avoid when campaigns have different priority or profitability. A shared budget removes a useful spending lever.

Current sources:

• https://support.google.com/google-ads/answer/11396330
• https://support.google.com/google-ads/answer/14587068

7. Account register

Maintain one row per serving account:

8. Automation and scripts policy

• Every script is read-only by default.
• A write-capable script needs explicit approval, action limit, cooldown, error log and rollback note.
• Store script owner, purpose, schedule, authorized scopes and output destination.
• Never expose a Google Sheet as "anyone can edit" for sensitive account data unless the integration requires it and the risk is explicitly accepted.
• The course's 80/20 smart-bidding integration uses an hourly Google Ads script and an editable web app/sheet. Treat that setup as a third-party data exposure decision, not a default account practice.

Course basis: ALL-53 transcript.

Measurement and Conversion Truth

1. Measurement contract before launch

No campaign launches until this table is complete:

2. Conversion goal rules

• One true outcome should not be counted twice through both Google Ads tag and GA4 import.
• Purchases are primary for eCommerce. Micro-conversions are secondary.
• For lead gen, move bidding downstream as soon as qualified-lead volume is sufficient.
• Use account-default goals only for outcomes appropriate to most campaigns in the account.
• Use campaign-specific goals when a campaign has a genuinely different business outcome.
• Every primary action must have the correct count setting, value behavior, attribution setting and conversion window.
• A raw form submission and a qualified lead are different conversion actions, not two labels for the same event.

Official Google behavior: account-default goals plus primary conversion actions populate the Conversions column and automated bidding. Secondary actions are observe-only in All conversions unless included in a custom goal.

Sources:

• https://support.google.com/google-ads/answer/4677036
• https://support.google.com/google-ads/answer/10995103

3. eCommerce implementation

Required purchase payload:

• transaction ID;
• revenue value;
• currency;
• product/item identifiers where supported;
• first-party customer data for enhanced conversions after consent;
• refund/cancellation handling in the reporting layer.

QA test:

1. Complete one test order from an ad-safe test path.
2. Verify one purchase in the browser/tag debugger.
3. Verify the same transaction ID in GA4/Ads diagnostics.
4. Confirm it is counted once.
5. Compare value and currency with Shopify/backend.
6. Confirm add-to-cart and checkout are secondary.
7. Confirm Ads purchase totals against backend orders over a full reporting window.

4. Lead generation implementation

Capture with every lead where available:

• GCLID, GBRAID or WBRAID;
• timestamp and time zone;
• conversion action name;
• CRM lead ID;
• consent state;
• lead source/campaign metadata;
• downstream stage;
• conversion value or revenue.

Recommended stage ladder:

raw lead -> valid lead -> MQL -> SQL -> opportunity -> won revenue

Do not train bidding on spam, duplicates or unreachable contacts when a downstream qualified event can be imported.

Course basis: MGA-09, Lead Generation Playbook pages 13-16.

5. Enhanced conversions and privacy

• Enhanced conversions use hashed first-party data to improve matching.
• From April 2026, Google is unifying enhanced conversions for web and leads into one account-level setting that accepts user-provided data from tags, Data Manager and APIs.
• For EEA users, ad_user_data consent controls whether user data can be sent for measurement use cases.
• Consent mode implementation must be tested in granted and denied states.
• Do not assume a consent banner automatically implements valid Google consent signals.

Sources:

• https://support.google.com/google-ads/answer/14664077
• https://support.google.com/google-ads/answer/15712870
• https://support.google.com/google-ads/answer/13802165

6. Tracking QA matrix

7. Account audit evidence

Collect:

• Goals summary export.
• Conversion actions with status, source, primary/secondary, count, value and window.
• Diagnostics screenshots/export for enhanced conversions.
• Tag Assistant evidence.
• GA4 key event list and Ads links.
• Backend order/lead totals by day.
• CRM stage mapping and imported offline conversions.
• Consent mode evidence for Bulgaria/EEA traffic.

No conversion setup is marked healthy from a settings screenshot alone. It needs one real event trace and reconciliation with the business source of truth.

Google Ads Optimization Operating Rhythm

1. S.T.A.B. review order

Always review in this order:

1. Spending and Segmentation.
2. Targeting.
3. Ads and Landing Pages.
4. Bidding.
5. Quality control.

Do not start by changing bids when the real problem is broken tracking, irrelevant queries, weak feed titles or a low-converting landing page.

Course basis: MGA-20 to MGA-51 and the full optimization workbook.

2. Weekly control loop

Account safety

• Payment method works.
• No critical policy or account notifications.
• Primary conversions are recording.
• Budget pacing is on plan.
• No unexpected auto-apply or script changes.
• Change History contains no unauthorized edits.

Spend exceptions

• Campaign/ad group/asset group/product with high spend and no result.
• Campaign with good economics but restricted budget or missed demand.
• Sudden cost, conversion or value shift.
• Geographic leakage.
• Search terms requiring negatives.
• Product disapprovals or limited eligibility.

Creative status

• Disapproved/limited ads.
• Broken landing pages.
• Asset groups with missing formats.
• Promotion assets still active after end date.

3. Monthly optimization

Spending and segmentation

• Campaign spend vs results.
• Ad group/asset group/product category spend vs results.
• Decide whether a profitable low-spend unit needs its own lever.
• Identify underperformers that need fix, isolation or exclusion.

Targeting

• Search terms and negatives.
• Exact-match additions from proven queries.
• Keyword CPC, CPA/ROAS, volume and status.
• Auction Insights.
• Location and device.
• Audience observations.
• PMax search themes, brand/URL exclusions and channel report.
• Display/Demand Gen/Video placements.
• Shopping titles, images, product status and competitive pressure.

Ads and landing pages

• RSA test winner by CTR plus conversion economics.
• Replace underperforming PMax/Demand Gen assets.
• Video retention at 25%, 75% and completion alongside conversion metrics.
• Landing-page conversion rate by URL and device.
• Page speed and offer consistency.

Bidding

• Conversion trend across 30/60/90 days.
• Weekly stability.
• Recent budget or target changes.
• Whether tCPA/tROAS is suppressing spend, impressions or conversion volume.
• Whether observed improvement is real or explained by brand mix, price/promo or tracking changes.

4. Quarterly structure review

• Every campaign still has a distinct purpose.
• Brand and non-brand reporting is clean.
• Campaign count is not fragmenting conversion data.
• Product/service taxonomy matches current margin and strategy.
• Locations/languages/markets still need separate levers.
• Customer acquisition settings match new vs existing customer strategy.
• Primary conversions still represent business value.
• Dormant scripts, users, manager links and obsolete negatives are removed or archived.
• Always-on and promotional structures are ready for the next seasonal period.

5. Decision thresholds

Course benchmark layer

The course uses the following diagnostic baselines:

• Search CTR around 7%.
• Search conversion rate 3% for eCommerce and 5% for lead gen.
• Shopping CTR around 0.7-0.9% and conversion rate around 3%.
• Display CTR around 0.6% and conversion rate around 1%.
• Video completion rate around 10-15% and conversion rate around 3%.
• Search impression share below roughly 60-65% as a possible scale signal when economics are good.

These are course heuristics, not universal 2026 benchmarks. Segment by brand/non-brand, market, device and intent before using them.

6. Smart Bidding gate

Course conservative checklist:

• at least 30 primary conversions in the last month;
• no more than about 20% weekly conversion variance across four weeks;
• no bid strategy/target change in the last six weeks;
• no budget change in the last two weeks;
• target based on actual 30-90 day performance;
• after switching, judge conversion outcomes rather than CPC alone.

Operator override:

• Never apply the gate if conversion actions are wrong.
• A campaign can have 30 low-quality leads and still be unfit for Smart Bidding.
• Consider conversion delay and seasonality.
• Use experiments when moving a valuable stable campaign to a materially different strategy.

7. Change log template

Every automated or manual material change gets a row.

8. Incident triggers

Immediate review:

• conversion count/value drops to zero;
• spend rises sharply without conversion volume;
• payment failure;
• site/checkout/form outage;
• Merchant Center mass disapproval;
• account suspension or major policy warning;
• new unauthorized admin/manager;
• wrong geo or language traffic;
• target/budget changed without log;
• duplicate conversion spike.

9. Optimization output format

Every review ends with no more than five prioritized actions:

Recommendations without evidence stay in a hypothesis backlog and are not applied.

Smart Bidding and Scaling

1. Select the optimization objective

eCommerce/value-based

• Maximize conversion value: maximize total value within budget without a ROAS constraint.
• Target ROAS: maximize value while aiming at an efficiency constraint.

Lead volume/value

• Maximize conversions: maximize conversion count within budget.
• Target CPA: maximize conversions around a cost constraint.
• Value-based bidding is preferred when qualified stages or lead values differ materially.

Google is relabeling Target CPA/Target ROAS strategies from June 2026 while retaining the underlying behavior.

Current source: https://support.google.com/google-ads/answer/15099424

2. Prerequisites

• Primary conversion actions represent real value.
• Values and currency are correct.
• No duplicates.
• Conversion delay is understood.
• Recent data represents the offer/market being optimized.
• Budget can support the target.
• Brand and non-brand mix does not mislead the efficiency baseline.

3. Course readiness guardrail

Proceed conservatively when:

• at least 30 primary conversions occurred in the last month;
• weekly conversions were stable within roughly 20% variance for four weeks;
• bidding strategy or target was not changed in the prior six weeks;
• budget was not changed in the prior two weeks.

This is a course operating heuristic, not an official minimum.

Course basis: MGA-50, MGA-51, ALL-53, eCommerce pages 21-22, Lead Gen pages 24-25.

4. Target setting

• Start from achieved 30-90 day performance.
• Do not set tROAS from a wish or tCPA from the finance target if the campaign has never achieved it.
• Tightening a target reduces eligible auctions and may reduce spend/volume.
• Relaxing a target can open more auctions but may reduce average efficiency.
• Evaluate with conversion delay included.
• For B2B, judge by qualified-stage CPA/value, not raw lead CPA.

5. Scaling levers

There are only two fundamental growth paths:

1. Spend more against the demand and audiences already working.
2. Reach colder audiences, broader queries, new channels, products or markets.

Before expansion, improve:

• CTR/query relevance;
• conversion rate;
• feed and offer;
• landing page;
• audience quality;
• product/service economics.

Course basis: ALL-54 and ALL-55.

6. CPC resistance

Course concept: an account reaches CPC resistance when added budget raises CPC materially faster than it raises useful volume. This indicates that the existing high-intent pool is constrained.

Responses:

• improve ad rank/relevance and conversion rate;
• expand adjacent query intent;
• expand market/product inventory;
• add Demand Gen/Video for colder audiences;
• avoid forcing more budget into a saturated campaign solely because it was historically the winner.

7. Budget changes

Course heuristic:

• increase around 20%;
• wait roughly 5-7 days or until conversion metrics recover;
• for small daily budgets, use a practical fixed increment.

Operator rule:

• change budget or target, not both;
• log the change;
• use marginal CPA/ROAS and total profit;
• do not apply a calendar wait blindly when volume/lag requires a different evidence window.

8. Growth and profit cycles

Course "bodybuilder" model:

• Growth cycle: relax efficiency, increase budget/volume, build data/customer base.
• Profit cycle: improve ROAS or CPA after volume expands.
• Align cycles with seasonality and business cash flow.

Do not try to maximize spend and tighten efficiency at the same moment unless offer/CVR/margin improvements create genuine room.

9. Smart bidding change checklist

Before:

• conversion truth verified;
• 30/60/90 day trend reviewed;
• brand/non-brand split reviewed;
• conversion delay reviewed;
• recent promo/price/site changes noted;
• target and budget constraint modeled;
• no conflicting scheduled change.

After:

• annotate the change;
• monitor safety/tracking daily;
• evaluate performance only after enough conversion feedback;
• watch spend, impressions, conversion volume, value and actual target achievement;
• diagnose whether movement is a bidding effect or mix/tracking effect.

10. 80/20 smart-bidding script from course

The course includes a third-party workflow that:

• connects Google Ads data to a Google Sheet via a scheduled script;
• exposes a web app URL to an external 80/20 Agent integration;
• runs a prompt that returns wait/proceed recommendations.

Risk controls before use:

• review script source and scopes;
• identify data exposed through sheet/web app;
• avoid public editable access unless explicitly accepted;
• keep analysis read-only;
• verify any "proceed" recommendation manually;
• never treat a model result as permission to edit the account.

Course basis: ALL-53 transcript.

Campaign Selection Matrix

Decision table

Brand Search

Use when people search the business or product brand. Keep separate even when it looks highly profitable, because it captures demand generated by other channels and organic brand awareness.

Non-brand Search

Default starting point for B2B and lead generation. Strong for eCommerce categories with explicit buying intent. Organize around semantic themes and landing-page match.

Call-focused Search

Use for local/service businesses when calls are the valuable action and staff can answer. Control device, schedule, location radius and call qualification. Review the call report, not only conversion count.

Course basis: MGA-12, MGA-26 and Call-only setup guide.

Dynamic Search

Use for large sites with many service/product pages or changing inventory. Keep in separate ad groups/campaigns, use page rules/feeds and exclude non-commercial sections.

Course basis: MGA-13 and Dynamic Search setup guide.

AI Max Search

Use as an experiment on existing Search when broader matching and asset/URL optimization have a plausible incremental opportunity. It is not a replacement for conversion truth, brand control or landing-page governance.

Current source: https://support.google.com/google-ads/answer/15910366

Standard Shopping

Use for clearer product/query economics, product segmentation and feed-led demand capture. Valuable as a controlled base layer before or alongside PMax.

Course basis: MGA-14, MGA-33 to MGA-37.

Performance Max

Use when cross-channel automation has strong purchase/qualified-value feedback and enough creative/product structure. Define a specific purpose: new customers, product family, market, promotion or economics segment.

Course basis: MGA-15, MGA-38 to MGA-40, eCommerce Playbook.

Display

Prefer remarketing or highly controlled audiences before cold broad reach. Separate from Search. Review placements, apps, devices, audience and creative performance.

Course basis: MGA-16, MGA-17 and Display sheet in the optimization workbook.

Video

Use for education, category creation, founder/product story and sequential remarketing. Optimize hooks, retention, CTA and downstream conversion. A view is not a sale.

Course basis: MGA-18 and Video sheet in the optimization workbook.

Demand Gen

Use when audience, channel and creative tests can be separated. Strong for visual eCommerce remarketing/prospecting and education. For B2B/lead gen, introduce after high-intent Search is stable or saturated.

See blueprints/05-demand-gen-blueprint.md.

Funnel mapping

eCommerce

• TOF: Demand Gen, Video, selective Display.
• MOF: PMax, Video sequence, benefit/problem Search, competitor Search.
• BOF: Shopping, high-intent Search, Brand Search.
• Remarketing: Demand Gen/Display/Shopping where supported and measurable.

B2B/lead gen

• TOF: Demand Gen, Video, education.
• MOF: PMax after offline conversions, benefit/problem Search, competitor Search.
• BOF: high-intent Search, Brand Search, qualified call campaigns.
• Remarketing: Demand Gen/Display with CRM/website audiences.

Course basis: eCommerce Playbook pages 12-15 and Lead Generation Playbook pages 17-20.

Current Google Ads Platform Delta - Verified 2026 Layer

Checked against official Google documentation on 2026-08-18.

Search

AI Max

• AI Max is an optimization layer inside Search campaigns, not a separate campaign type.
• Main feature groups: search term matching and asset optimization.
• It can use broad match, keywordless matching, text customization and Final URL expansion.
• Controls include brand and geographic settings, plus improved reporting.
• Use an AI Max experiment before broad rollout on valuable campaigns.

Sources:

• https://support.google.com/google-ads/answer/15910366
• https://support.google.com/google-ads/answer/15910187
• https://support.google.com/google-ads/answer/16450159

Performance Max

Negative keywords and brand

• Campaign-level negative keywords can block queries on PMax Search and Shopping inventory.
• Account-level negatives apply to relevant Search and Shopping inventory across multiple campaign types.
• Brand exclusions remain the stronger entity-based control for brands, misspellings and foreign scripts.

Sources:

• https://support.google.com/google-ads/answer/15726455
• https://support.google.com/google-ads/answer/11396330
• https://support.google.com/google-ads/answer/14587068

Channel reporting

• PMax channel performance reporting includes cost, interactions, conversions and conversion value across channels.
• It can be segmented by ads using product data and used for channel diagnostics.

Source: https://support.google.com/google-ads/answer/16260130

Final URL expansion

• Final URL expansion is on by default.
• With expansion on, page feeds prioritize but do not restrict URLs.
• With expansion off, destination URLs are limited to page feeds and asset groups.
• URL exclusions and text customization controls must be reviewed for multi-product and regulated sites.

Sources:

• https://support.google.com/google-ads/answer/14337539
• https://support.google.com/google-ads/answer/15995647

Demand Gen

• Channels include YouTube, Discover, Gmail, Maps and optionally Google Display Network.
• Channel selection can be automatic or manually controlled at ad-group level.
• Reporting can be segmented by channel and YouTube format.
• Google recommends all-channel operation for most performance use cases, with manual isolation when creative or strategy specifically requires it.

Sources:

• https://support.google.com/google-ads/answer/15973205
• https://support.google.com/google-ads/answer/13695777

Measurement

• Enhanced conversions for web and leads are moving to a unified account-level setting from April 2026.
• Google Ads can accept user-provided data from website tags, Data Manager and API connections.
• EEA consent signals include ad_storage, ad_user_data, ad_personalization and analytics_storage.

Sources:

• https://support.google.com/google-ads/answer/14664077
• https://support.google.com/google-ads/answer/15712870
• https://support.google.com/google-ads/answer/13802165

Smart Bidding labels

• Starting June 2026, Google is changing labels: "Maximize conversions with Target CPA" becomes "Target CPA", and "Maximize conversion value with Target ROAS" becomes "Target ROAS".
• Google states the underlying bidding behavior does not change.

Source: https://support.google.com/google-ads/answer/15099424

Account administration

• Serving-account time zone and currency remain permanent account-creation choices.
• Manager accounts can enforce 2-Step Verification and allowed domains across owned sub-accounts.
• Named individual access and minimum permissions remain the secure operating model.

Sources:

• https://support.google.com/google-ads/answer/9842104
• https://support.google.com/google-ads/answer/12865295
• https://support.google.com/google-ads/answer/12864492

Feed and AI labeling

• Merchant Center product free-text fields must use one consistent feed language.
• Accurate IDs, GTIN/MPN, variants, prices and availability remain core eligibility inputs.
• Current product specifications include structured title handling for generative-AI-created titles and new AI content label settings rolling out in 2026.

Source: https://support.google.com/merchants/answer/7052112

Search Campaign Blueprint

1. Campaign roles

Brand Search

Purpose:

• Own brand SERP coverage.
• Protect against competitors bidding on the brand.
• Control brand messaging, promotions and sitelinks.
• Report branded demand separately from incremental non-brand acquisition.

Rules:

• Separate campaign and budget.
• Exact and phrase coverage for brand, spelling variants, product-brand combinations and high-risk misspellings.
• Exclude brand from non-brand Search where clean incrementality reporting is required.
• Exclude brand from PMax using campaign-level brand exclusions when PMax should optimize for non-brand/new-customer demand.
• Never present Brand Search ROAS as the acquisition ROAS of the whole account.

Non-brand high intent

• Group by one semantic keyword theme per ad group, not single-keyword ad groups.
• Separate campaigns only when budget, location, margin, service line, landing page or business objective needs independent control.
• Start with high-commercial-intent themes.

Benefit/problem aware

• Queries describing a problem, desired outcome or use case.
• Separate from high-intent product/service queries because CPA, copy and landing page often differ.

Competitor

• Separate budget and reporting.
• Use factual, non-confusing copy and policy-safe trademark handling.
• Judge on qualified outcome, not CTR.

Dynamic Search or AI Max expansion

• Expansion layer after base Search structure, conversion tracking and URL controls are stable.
• Use page feeds and URL exclusions when the site contains support, blog, careers, policies or unrelated product lines.
• Test through an experiment before broad rollout.

2. Keyword research process

Course method:

1. Write the phrases a buyer would use.
2. Enter seed phrases and landing pages into Keyword Planner.
3. Filter for commercial relevance and business fit.
4. Build semantic themes.
5. Start each theme with 1-3 specific long-tail broad-match phrases.
6. Add exact-match queries discovered through search term audits.

The course favors longer broad-match phrases because additional words provide stronger intent context. This is a useful starting heuristic, not a substitute for search term control.

Course basis: MGA-04 and MGA-05, Keyword Research PDF pages 2-22.

3. Ad group design

An ad group passes when:

• every keyword maps to the same intent;
• one landing page is credible for all keywords;
• one RSA message can naturally include the theme;
• search terms should not routinely bleed into another ad group;
• the theme has enough volume to learn without being drowned by a much larger theme.

Do not split synonyms into separate ad groups unless they require a different offer, landing page or budget control.

4. Match type policy

• Exact match: proven converting queries, critical brand terms and control points.
• Phrase match: useful when query intent must stay within a recognizable modifier pattern.
• Broad match: discovery and scale only with trusted conversion data, disciplined negatives and sufficient review frequency.
• AI Max search term matching: a controlled experiment, not an automatic account-wide switch.

Current platform note: AI Max is an optimization layer inside existing Search campaigns. It can use broad match, keywordless matching, text customization and Final URL expansion. Conversion-based Smart Bidding is important for its dynamic matching and asset features.

Source: https://support.google.com/google-ads/answer/15909989

5. RSA construction

Headline inventory should cover:

• keyword/theme relevance;
• differentiated outcome;
• proof or authority;
• concrete number, value or percentage where truthful;
• risk reversal;
• offer;
• brand;
• clear CTA.

Description inventory should cover:

• what the product/service is;
• who it is for;
• why it is different;
• proof;
• what happens after the click;
• CTA.

Rules:

• Every asset must make sense in multiple combinations.
• Pin only when legal, brand or semantic order requires it, or when pinning is the explicit test variable.
• Avoid generic claims like "We are #1" or "Trusted choice" without evidence.
• Keep one major difference between test variants.
• Evaluate CTR together with conversion rate and CPA/ROAS.

Course basis: MGA-06, Ad Copy PDF pages 10-26 and optimisation checklist.

6. Landing page contract

• Intent match between query, ad and page headline.
• One primary action.
• Offer and price/qualification clarity.
• Proof near the decision point.
• Mobile-first interaction and fast load.
• No surprise redirect or irrelevant generic homepage.
• Policy-safe claims.
• Form length appropriate to lead value.
• Thank-you event cannot fire on page reload or direct URL entry.

7. Settings checklist

• Correct campaign goal and primary conversion actions.
• Search campaign type.
• Network selections deliberately reviewed. Do not mix Display expansion into Search.
• Location option set deliberately. For local/Bulgaria acquisition, default operating policy is Presence unless there is a clear reason to target interest in the location.
• One language strategy per campaign when copy and landing experience differ.
• Ad schedule only when business availability or measured economics require it.
• Search partners separated in reporting and retained only if quality holds.
• Automatically created assets/text customization reviewed against brand and regulatory risk.
• Brand exclusions and negative lists applied at the correct scope.
• Full asset coverage: sitelinks, callouts, structured snippets, images, price/promotion/call/location where relevant.

Google's default geo setting includes people who show interest in the location. Presence narrows delivery to people in or regularly in the target location.

Source: https://support.google.com/google-ads/answer/9376662

8. Launch budget

Course heuristic: a starting daily budget around 10 times estimated CPC can collect enough click data. Treat this as a planning aid, not a universal minimum.

Operator rule:

• Budget is derived from required conversion learning and affordable downside.
• If expected clicks cannot plausibly produce a conversion inside the evaluation window, consolidate themes or increase the test budget.
• Do not spread a small budget across many campaigns.

9. Search optimization cadence

Weekly

• Spend vs result exceptions.
• Search terms and negatives.
• Disapprovals and broken assets.
• Budget pacing.
• Conversion action health.

Monthly

• Campaign/ad-group segmentation opportunities.
• Exact-match additions from proven terms.
• Keyword CPA/ROAS, CPC and status.
• Auction Insights.
• Location, device and audience observations.
• RSA test result and next one-variable test.
• Asset performance.
• Bid target constraints.

Quarterly

• Day/hour patterns.
• Landing-page conversion by URL.
• Demographics where legally and statistically appropriate.
• Brand incrementality and PMax/Search overlap.
• Structure simplification or controlled segmentation.

Course basis: MGA-20 to MGA-32, MGA-44 to MGA-46 and Search sheet in the optimisation workbook.

10. Search kill/scale decisions

Scale when:

• conversion tracking is trusted;
• unit economics are inside target;
• there is missed eligible demand or impression share;
• search terms remain relevant;
• results hold across enough data and more than one short window.

Fix before scaling when:

• CTR is weak because relevance/copy is weak;
• landing page conversion is weak;
• negatives are incomplete;
• geographic leakage exists;
• bidding is constrained by unrealistic tCPA/tROAS;
• Brand Search is hiding weak non-brand acquisition.

Pause or isolate when:

• the theme repeatedly spends past its evidence-based loss limit;
• search intent is structurally wrong;
• the service/product cannot convert profitably after ad and landing-page fixes;
• the campaign has no distinct purpose.

Performance Max, Shopping and Multi-Product Blueprint

1. Strategic role

Shopping provides product-level demand capture and clearer query/product control. PMax expands across Google's inventory and can add reach, creative formats and automation. Neither fixes a weak feed, broken measurement or unprofitable unit economics.

For a dormant or new eCommerce account:

1. Restore purchase tracking and Merchant Center health.
2. Re-establish Brand Search plus controlled non-brand Search/Shopping demand.
3. Reintroduce PMax after product-level economics and conversion feedback are trustworthy.
4. Add Demand Gen or Video only after creative and budget are not starving bottom-funnel demand.

2. Campaign split decision tree

Create a separate Shopping/PMax campaign only when it creates a necessary lever:

• market or language;
• materially different margin/COGS;
• seasonality;
• inventory availability;
• hero vs low-performing products;
• new-customer vs all-customer goal;
• promotional vs always-on budget;
• distinct URL/product family control;
• separate tROAS/business target supported by enough data.

Do not split only by every menu category if each new campaign becomes data-starved.

Course basis: eCommerce Playbook pages 2-7 and 37-42.

3. Product performance taxonomy

Hero

• High value/ROAS and enough volume.
• Protect budget and inventory.
• Consider isolating other products away so Heroes are not budget-starved.

Sidekick

• Good efficiency but low spend.
• Test whether performance survives more budget in a controlled campaign or product segment.

Zombie

• Too little traffic/data to classify.
• Isolate with a controlled budget and improve feed/query alignment.

Villain

• High spend and persistently poor economics.
• Exclude or isolate under a strict budget while fixing title, price, page, offer or query match.

Course basis: eCommerce Playbook pages 23-24.

4. Feed operating system

Required product truth

• stable unique ID/SKU;
• accurate title and description;
• price and availability matching landing page;
• correct GTIN/MPN/brand where applicable;
• variant attributes such as size, color and item group ID;
• high-quality images without avoidable policy conflicts;
• shipping and returns data;
• correct language and market feed setup.

Google permits up to 150 characters for product title and requires free-form attributes in one consistent feed language. Incorrect or missing identifiers and variant attributes can cause limited eligibility or disapprovals.

Current source: https://support.google.com/merchants/answer/7052112

Course title formula

Use the most important information first and include relevant query modifiers:

product type + core attribute/use case + brand + variant modifiers

Examples of modifiers:

• audience;
• size;
• material;
• color;
• species/breed/age where policy-safe and relevant;
• pack size;
• benefit or use case when factual.

Search term reviews should feed title testing. Do not stuff unrelated keywords.

Course basis: MGA-33 to MGA-37 and Shopping sheet in the optimisation workbook.

5. Merchant Center QA

Daily/weekly

• Disapprovals and limited products.
• Price/availability mismatch.
• Feed/API failures.
• Account warnings and policy deadlines.

Monthly

• Product clicks, CTR, conversion rate and value/cost.
• Competitor price pressure and product visibility.
• Store quality: speed, image quality, shipping, returns, reviews and payment options.
• Product title and image tests.
• Inventory and margin changes.

6. PMax setup

Campaign

• One explicit business purpose.
• Correct conversion goals.
• Product subset defined through listing groups/feed labels.
• New-customer setting chosen deliberately, not automatically.
• Brand exclusion policy documented.
• Campaign-level negative keywords added for unwanted Search/Shopping queries.
• Final URL expansion and text customization explicitly reviewed.
• URL exclusions for blog, support, legal, account, jobs, unrelated products and out-of-stock destinations.
• Geographic Presence rule checked.
• Bid strategy based on actual data maturity.

Asset groups

• One coherent product/service theme.
• Matching product listing group.
• Matching landing page family.
• Search themes relevant to that exact product/service set.
• Dedicated images, video, headlines and descriptions.
• No generic asset group that mixes unrelated product categories.

Creative supply

• Brand name and logo through brand guidelines.
• Multiple aspect ratios.
• At least one purpose-built video. Do not rely on auto-generated video as the creative strategy.
• Text variations that remain factual in any combination.
• Seasonal assets planned before the peak period.

Google's current PMax creative guidance recommends at least 7 image assets, including a 1200x1200 image, and says low-rated assets should be replaced before deletion.

Current source: https://support.google.com/google-ads/answer/14528221

7. Brand control

For separate Brand Search reporting:

• Use campaign-level brand exclusions in PMax for own brand and relevant sub-brands.
• Use PMax campaign-level negative keywords when exact query control is needed.
• Do not use account-level brand negatives if Brand Search still needs to serve those terms.
• Review misspellings and foreign-script brand variants through brand exclusion behavior.
• Decide whether Shopping ads may serve on excluded-brand searches based on reporting and incrementality goal.

Google identifies brand exclusions as the stronger brand-entity control because they can cover misspellings and foreign scripts. Campaign-level negatives are available for PMax Search and Shopping inventory.

Sources:

• https://support.google.com/google-ads/answer/14587068
• https://support.google.com/google-ads/answer/15726455

8. Final URL expansion policy

Turn on when

• site taxonomy is clean;
• all indexed commercial pages are appropriate;
• conversion goals are consistent;
• URL exclusions are complete;
• dynamic text is safe for the brand and regulated category.

Restrict or turn off when

• multiple unrelated business lines share a domain;
• legal/medical/regulated copy needs exact control;
• pages have mismatched offers or currencies;
• the site contains large editorial/support sections;
• asset groups must stay inside a specific product family.

With expansion on, page feeds prioritize URLs but do not strictly limit delivery. With expansion off, PMax uses URLs supplied through page feeds and asset groups.

Current source: https://support.google.com/google-ads/answer/14337539

9. PMax optimization

Weekly

• Spend vs purchase/value.
• Product disapprovals.
• Search term insights and campaign negatives.
• Asset-group status.
• Tracking and budget pacing.

Monthly

• Product performance taxonomy.
• Feed titles/images.
• Location and audience signals.
• Asset-level performance and replacements.
• URL and landing-page performance.
• Placement and channel performance reports.
• Target ROAS/CPA constraint diagnostics.

Quarterly

• Brand/non-brand contribution.
• New vs existing customer economics.
• Campaign consolidation vs segmentation.
• Hero/Sidekick/Zombie/Villain migration.
• Promotional vs always-on separation.
• Incrementality test plan.

10. Channel performance

PMax channel reporting now shows campaign-level delivery across Google channels with impressions, clicks, cost, conversions and conversion value. Use it to diagnose creative coverage and channel mix. Do not automatically split a channel merely because its isolated CPA differs; assess whether it contributes to total campaign efficiency and available volume.

Current source: https://support.google.com/google-ads/answer/16260130

11. Scaling

Course guardrails:

• Increase budget around 20% after results recover and stabilize, often using a 5-7 day wait.
• Set tROAS from achieved 30-90 day performance, not from the aspirational target.
• Move toward a higher tROAS in steps.
• Separate growth periods from profit-efficiency periods.

Operator application:

• Change either budget or target first.
• Evaluate marginal ROAS, not only blended ROAS.
• Preserve Brand Search separation so PMax cannot manufacture an apparently high ROAS by harvesting brand demand.
• Use contribution margin and new-customer economics, not revenue ROAS alone.

Course basis: eCommerce Playbook pages 18-22 and ALL-54.

12. Recommended multi-product skeleton

Small catalog

• Brand Search.
• Non-brand Search by major intent themes.
• Shopping with ad/listing groups by product family or economics.
• One PMax for new-customer/non-brand growth, brand excluded.
• Optional remarketing/Demand Gen only if budget and creative support it.

Large catalog

• Brand Search.
• Non-brand Search for priority categories and use cases.
• Shopping catch-all or controlled base layer.
• PMax campaigns by genuine budget/margin/market lever.
• Custom labels for margin, season, hero status, price band and inventory.
• Strict URL and listing group alignment between PMax campaign and asset groups.
• One brand-level Demand Gen structure unless product families need distinct creative/audience strategy.

B2B and Lead Generation Blueprint

1. Core principle

B2B Google Ads succeeds when bidding learns from qualified pipeline, not when the account maximizes cheap form submissions.

Before media:

• define ICP;
• define disqualifiers;
• define primary markets;
• calculate lead-to-qualified, qualified-to-sale and average deal value;
• set an allowable qualified-lead CPA;
• connect click IDs to CRM stages;
• prepare a landing page for one buyer problem and one next action.

2. Initial campaign structure

Search campaigns

• Brand.
• High-intent category/service.
• Problem/benefit-aware.
• Competitor, isolated.
• Location or language split only where economics/operations differ.
• Call-focused campaign only when calls are valuable, staffed and measurable.

Campaigns should be separated by the lever that needs independent budget or operation: service line, location, device/call availability, season or margin.

Course basis: Lead Generation Playbook pages 2-6 and 29-42.

3. Query strategy

Prioritize:

• product/category queries;
• "software/platform/service for X";
• use-case and integration queries;
• switching/replacement queries;
• problem-aware queries with commercial intent;
• competitor alternatives where policy and economics allow.

Exclude or isolate:

• jobs, salary and careers;
• free/template/course/definition intent unless part of a lead magnet campaign;
• consumer intent in B2B-only offers;
• support/login/current-customer intent;
• irrelevant geographies;
• academic research intent.

4. Conversion hierarchy

Preferred bidding event order:

1. Closed/won revenue with actual value.
2. Sales-qualified opportunity with modeled value.
3. Qualified lead/MQL with modeled value.
4. Valid lead form or qualified call during cold start.

Never keep raw lead as the permanent optimization goal if the CRM can distinguish quality.

Suggested modeled value:

stage value = probability of sale at stage x expected gross profit per sale

Example:

• Expected gross profit per deal: EUR 8,000.
• Qualified lead to sale: 10%.
• Qualified-lead value sent for bidding: EUR 800.

Recalibrate from actual cohort data.

Course basis: Lead Generation Playbook pages 13-16.

5. Landing page

• Identify role/company fit in the hero section.
• State the business outcome, not a vague feature list.
• Show proof: logos, quantified results, relevant certification, case study.
• Explain implementation friction and time-to-value.
• Pre-qualify with company size, market, need or budget when sales capacity is limited.
• Keep the CTA aligned with buying stage: demo, assessment, quote, audit or lead magnet.
• Store click IDs in hidden fields and CRM.

6. Ad copy

• Use ICP language and pain.
• State a concrete differentiator competitors cannot copy.
• Use truthful numerical proof.
• Match CTA to buyer stage.
• Avoid unprovable "#1", "best" and generic trust claims.
• Use distinct ads for technical buyer, economic buyer and operator only when their query/landing journey differs.

7. PMax gate for lead gen

Do not launch PMax for lead generation until:

• spam controls work;
• enhanced conversions for leads are active;
• offline qualified stages import reliably;
• primary goals exclude low-value micro-conversions;
• URL expansion cannot reach support, jobs, blog or consumer pages;
• there is enough downstream signal to evaluate lead quality.

PMax can then be used as a broader mid-funnel/new-customer layer, with campaign-specific goals and strong URL/brand controls.

8. Demand Gen role

Use when:

• the market needs education;
• there is a strong report, webinar, comparison, calculator or case study;
• creative exists for YouTube/Discover/Gmail/Maps/Display inventory;
• nurture and CRM follow-up are ready;
• success is evaluated downstream, not only by cheap downloads.

Demand Gen now supports manual channel selection at ad-group level or automatic use of all Google channels. Start broad unless the creative or strategy genuinely requires a surface-specific test.

Source: https://support.google.com/google-ads/answer/15973205

9. B2B budget logic

• Concentrate on the highest-value intent before opening more funnel stages.
• A campaign needs enough budget to create a plausible number of qualified outcomes, not just clicks.
• Split markets only when each split can still learn.
• Use portfolio bidding only when campaigns share the same economic target and conversion quality.
• High-CPC markets require stronger qualification and value import, not automatic abandonment.

10. B2B weekly scorecard

11. Scale gate

Scale only when:

• qualified conversion data is stable;
• sales capacity can handle more volume;
• marginal qualified-lead CPA is acceptable;
• CRM feedback arrives fast enough for bidding;
• new spend is not merely increasing raw/spam leads;
• query and geo expansion remain within ICP.

Bulgaria and International Market Adaptation

1. What stays constant

• One accountable owner and clean access.
• Accurate primary conversions and values.
• Brand separated from non-brand acquisition.
• Campaigns built around business levers.
• Search term and feed discipline.
• Controlled bidding changes.
• Local policy, language, price, shipping and landing-page truth.

2. Bulgaria operating model

Account

• Currency: choose the reporting/billing currency deliberately at account creation. It cannot be changed later.
• Time zone: Europe/Sofia for a Bulgaria-operated account unless group reporting requires a different new account design.
• Language: Bulgarian copy and Bulgarian landing pages. Add English only when the actual buyer behavior and landing experience justify it.
• Location option: default to Presence for domestic products/services. The Google default also includes users showing interest in Bulgaria.
• Use Bulgarian spelling, transliterations and common mixed-language product terms in query research.

Search

• Build Bulgarian semantic themes manually. Keyword Planner volume will be thinner and noisy.
• Include local commercial modifiers: price, order, delivery, city, online, brand/category terms in Bulgarian and common Latin spellings.
• Maintain transliteration negatives and positives separately when they behave differently.
• Do not over-segment a low-volume market. Fewer campaigns/ad groups preserve learning.
• Brand defense matters if marketplaces, resellers or competitors appear on brand queries.

eCommerce/PMax

• Feed title and landing page language must match the market.
• Price, shipping time, cash-on-delivery terms, returns and trust proof must be explicit.
• Product IDs remain stable across updates.
• Segment only major product/margin levers because total conversion volume is smaller.
• PMax brand exclusion is essential when Brand Search must be reported independently.

B2B

• Search volume can be too low for narrow job-title or industry segmentation.
• Combine close problem themes and qualify on the landing page/form.
• Import qualified stages because raw lead volume is small and each bad signal has larger influence.

3. International single-market launch

Each new country gets a market readiness gate:

• local demand and CPC research;
• local offer and price competitiveness;
• delivery, returns and payment support;
• native-quality copy;
• local landing pages and policies;
• conversion tracking with correct value/currency;
• customer support capacity;
• Merchant Center eligibility/feed language;
• legal and ad-policy review;
• break-even economics including VAT, duties, shipping and returns.

Do not launch one campaign across countries with materially different economics or languages.

4. Account vs campaign by country

Separate account

• different currency or time zone;
• different legal entity/payment profile;
• separate operating team/P&L;
• radically different conversion hierarchy or compliance;
• independent Merchant Center/account ownership.

Same account, separate campaign

• same currency, ownership, tracking and business model;
• country still needs separate budget, language, feed, price or target;
• consolidated account data remains useful.

Shared campaign across countries

Use only when:

• language and offer are identical;
• unit economics are close;
• shipping/service levels are close;
• one target makes commercial sense;
• country-level reporting is monitored and budget control is not required.

5. Language architecture

• One campaign per customer-facing language when ad copy and landing pages differ.
• Do not rely on Google to translate ads.
• Use local search behavior, not literal translation of Bulgarian/English keywords.
• Keep brand spelling variants and local scripts in the brand-control map.
• Product feed free-form fields should use one consistent language per feed.

Current Merchant Center source: https://support.google.com/merchants/answer/7052112

6. Geo controls

• Presence: domestic/local acquisition and most country-specific eCommerce.
• Presence or interest: only when people outside the market can genuinely convert, such as travel, relocation, exports or destination services.
• Exclude locations that generate non-serviceable traffic.
• Inspect matched user location, not only targeted location.
• Use city/region split only when budget, margin, service capacity or messaging needs the lever.

Current source: https://support.google.com/google-ads/answer/9376662

7. International expansion sequence

1. Prove one offer in one market with Search/Shopping.
2. Localize landing page, feed and support before broad media.
3. Protect/measure Brand Search separately.
4. Establish non-brand unit economics.
5. Add PMax with market-specific brand, feed and URL controls.
6. Add Demand Gen/Video after localized creative exists.
7. Compare markets by contribution margin and new-customer CAC, not platform ROAS alone.

8. Cross-market scorecard

Kill a market test when the fully loaded economics fail after query, offer and landing-page corrections. Do not keep it alive because blended international ROAS hides the loss.

Demand Gen Blueprint

1. Use case gate

Demand Gen is a creative and audience expansion system, not the first campaign for an unproven account.

Good use cases:

• eCommerce prospecting and remarketing with product feed;
• educational brands using proof, tutorials and founder content;
• B2B/lead gen after Search reaches CPC/volume resistance;
• promotion of lead magnets, trials, webinars and community offers;
• creative learning across visual Google inventory.

Do not launch when:

• Search/Shopping demand is underfunded;
• conversion tracking is weak;
• there is no creative testing capacity;
• remarketing/prospecting goals are mixed with no budget control;
• downstream lead or LTV tracking is absent.

Course basis: ALL-55 transcript.

2. Three-variable testing model

Every ad group test should make its variables visible:

1. Audience: customer list/lookalike, website visitors, cart abandoners, custom segment, interest/demographic.
2. Channel/placement: Shorts, in-stream, in-feed, Discover, Gmail, Maps or Display Network.
3. Creative: concept, hook, actor, setting, format, image, thumbnail, CTA.

Do not place every audience and format in one ad group. You lose the ability to learn which combination works.

Naming:

{prospecting-or-remarketing}_{audience}_{channel}_{creative-theme}_{version}

Examples:

• RMKT_CART-ABANDONERS_SHORTS_PRODUCT-REVIEW_V1
• PROSPECT_LOOKALIKE_SHORTS_UGC-PROOF_V2
• PROSPECT_CUSTOM-SEGMENT_INFEED_FOUNDER-STORY_V1

3. Campaign structure

Separate campaigns when budget control is needed for:

• prospecting vs remarketing;
• new customer vs existing customer;
• market/language;
• product feed vs non-feed creative;
• different business goals.

Use ad groups to test audience/channel combinations inside a coherent campaign goal.

Course basis: ALL-55 transcript.

4. Creative by YouTube format

Shorts

• Vertical, native-feeling, simple production.
• Course heuristic: 15-30 seconds, with up to 60 seconds possible.
• UGC, customer testimonial or product voiceover often fits the surface better than polished "ad-looking" work.
• Show the outcome/product quickly.
• CTA before the end when the concept needs it.

In-stream

• The opening must earn attention before the skip point.
• Brand early for recall.
• Problem and solution introduced quickly.
• Test actor, age, gender, setting and tone separately.

In-feed

• Behaves closer to chosen content.
• Thumbnail and title do the click work.
• Strong for long-form education, founder story, tutorial and social proof.
• Optimize overlay, title and thumbnail as separate test variables.

Image/Discover/Gmail/Display

• Test product vs lifestyle, message, proof, color and offer.
• Inspect app/site inventory and quality through available channel/content reports and account-level suitability controls.

5. Current channel controls

Demand Gen currently supports:

• YouTube in-stream;
• YouTube in-feed;
• YouTube Shorts;
• Discover;
• Gmail;
• Maps;
• Google Display Network.

Selection is at ad-group level. Use All Google channels for broad performance optimization unless creative or strategy requires a deliberate channel-specific test. Channel reporting is available at campaign, ad-group and ad level.

Current source: https://support.google.com/google-ads/answer/15973205

6. Audience architecture

• Remarketing: visitors, product viewers, cart abandoners, non-converters.
• Customer data: buyers, subscribers, qualified leads.
• Lookalike: source list must represent the desired business outcome, not mixed low-quality leads.
• Custom segments: relevant websites, search behavior, apps and YouTube content.
• Google interests/demographics: use as separate hypotheses when possible.
• Exclusions: purchasers/converters from acquisition campaigns, age or other groups only when legal, relevant and supported by evidence.

7. Product feed use

For eCommerce:

• product tiles can support Shorts and other visual formats;
• feed titles, price, availability, images and landing pages must be correct;
• map product set to campaign goal;
• separate remarketing/new-customer economics in reporting.

8. Initial budget and learning

Course heuristic: start Demand Gen with roughly 5-10% of account budget while audience, channel and creative fit are still being solved.

Operator rule:

• Do not starve profitable Search/Shopping.
• Set a loss limit tied to qualified outcomes or incremental account effect.
• Evaluate direct, view-through and total account effects carefully, without using view-through conversions as automatic proof of incrementality.

9. Optimization

Audience

• Compare cost, conversions and value/cost.
• Pause or isolate high-spend low-value audience tests.
• Separate remarketing if it materially outperforms prospecting and needs its own budget.

Channel

• Review channel and YouTube-format performance.
• Keep automatic multi-channel operation when it produces the best total outcome.
• Isolate a channel only for a clear creative/strategic reason or controlled test.

Creative

• Start with 3-4 meaningful concepts, not cosmetic variants.
• Replace losers with a new variation of the winning hook/actor/setting/format.
• Review video retention, CTR, conversion rate, CPA and value/cost together.

Bidding

• Use the same conversion-volume/stability checks as other automated campaigns.
• Do not tighten tCPA/tROAS while changing creative, audience and budget simultaneously.

10. Course case-study learnings

These are examples, not transferable benchmarks:

• One eCommerce remarketing test found Shorts plus product feed profitable while Display did not; prospecting remained weaker.
• One personal brand/community test found image-based Display inventory stronger than Shorts.
• The lesson is not "Shorts wins" or "Display wins". The repeatable insight is to separate and test audience, channel and creative combinations.

Course basis: ALL-55 transcript.

Dormant Google Ads Account Restart

Current known context

Status: USER-CONTEXT, not yet verified in Google Ads.

• Advertising has been inactive since near the end of 2025.
• Performance Max was the main sales driver.
• Brand Search existed separately.
• Brand was excluded from other campaigns to reduce overlap and misleading performance.
• Planned next use cases include B2B and multi-product eCommerce.

Restart verdict rule

Do not simply enable every old campaign. Do not rebuild everything in a new account before preserving history.

Choose one of three outcomes per campaign:

• REACTIVATE: structure, tracking target and offer are still valid.
• REBUILD-IN-PLACE: historical learning is useful, but targeting/assets/settings need a controlled replacement.
• ARCHIVE: campaign has no current purpose, incompatible goals or irreparable structure.

Phase 1: preserve evidence

Before any edit:

• Export campaign, ad group, asset group, keyword, search term and product data.
• Save Change History around the last active period and shutdown.
• Capture conversion goal settings and diagnostics.
• Export 30/60/90 days before pause, last full active quarter, full 2025 and comparable prior-year periods.
• Save Merchant Center diagnostics and product history.
• Save all current users, managers, scripts, auto-apply settings and linked products.
• Record current website, feed and offer differences from 2025.

Phase 2: measurement gate

The account remains paused until:

• primary purchase/qualified-lead actions are correct;
• duplicates are removed;
• values and currency are correct;
• one real test conversion is observed end-to-end;
• enhanced conversions and consent mode are checked;
• Merchant Center product approval and landing destinations are healthy;
• backend reconciliation is documented.

Phase 3: historical diagnosis

Brand truth

• Separate Brand Search cost, conversions and value.
• Measure whether PMax served own-brand queries despite exclusions.
• Recalculate non-brand PMax/Search performance.
• Check if a change in brand demand, price, promo or organic traffic explains apparent ROAS.

PMax truth

• Channel mix.
• Search term/category insights.
• Product-level performance.
• New vs existing customers.
• Asset groups and URL expansion.
• Brand and negative keyword controls.
• Actual target ROAS/CPA vs achieved results.
• Feed/offer changes since the active period.

Search truth

• Brand vs non-brand.
• Query relevance and negative history.
• Keyword match-type drift.
• Landing-page conversion.
• Location/device/network quality.
• Auction changes.
• Ad and asset performance.

Phase 4: restart design

Minimum eCommerce restart

1. Brand Search.
2. Non-brand Search for proven high-intent categories.
3. Shopping or controlled product layer.
4. PMax for defined non-brand/new-customer purpose with brand exclusion.

Do not add Demand Gen until bottom-funnel campaigns, creative supply and measurement are stable.

Minimum B2B restart/new account

1. Brand Search.
2. High-intent Search by service/problem theme.
3. CRM click-ID capture and offline qualified conversion import.
4. Remarketing after audience volume exists.
5. PMax/Demand Gen only after downstream lead quality is visible.

Phase 5: launch control

• One campaign family at a time.
• Confirm spend, queries, geo, conversions and landing pages after launch.
• No simultaneous budget and target change.
• Daily safety checks during restart, but no short-window performance panic.
• Compare with the same conversion definition and attribution context as historical data.
• Log every change.

Stop conditions

Keep campaigns paused when any are true:

• tracking is untrusted;
• site/checkout/form is broken;
• primary goal optimizes for a micro-conversion;
• Merchant Center has material disapprovals;
• budget owner/allowable CAC is undefined;
• brand isolation is not implemented;
• geo/language/market offer is not ready;
• CRM cannot distinguish qualified leads for B2B.

Google Ads Audit Data Request

Preferred access

Read-only Google Ads Manager access to the relevant serving accounts. Read-only Merchant Center and GA4 access where applicable.

No live edits are required for the audit.

Google Ads exports

Date ranges:

• last 90 active days before pause;
• full 2025;
• full 2024 where available;
• current date for settings and status.

Export with segments where supported:

• Campaigns.
• Ad groups.
• Ads and assets.
• Keywords.
• Search terms.
• Landing pages.
• Locations/user locations.
• Devices.
• Networks.
• Audiences/demographics.
• Conversion action.
• New vs returning/new-customer reporting.
• Products/listing groups.
• PMax asset groups, search term insights and channel performance.
• Auction Insights.
• Change History.

Core columns:

• status;
• budget;
• bid strategy and target;
• impressions;
• clicks/interactions;
• cost;
• conversions;
• conversion value;
• all conversions;
• all conversion value;
• CTR;
• average CPC;
• conversion rate;
• cost/conversion;
• value/cost;
• search impression share;
• lost IS budget/rank;
• click share where available.

Conversion evidence

• Goals summary.
• Every conversion action with source, status, primary/secondary, count, value, window and attribution.
• Enhanced conversion diagnostics.
• GA4 linked-property details.
• Tag/GTM container export or screenshots.
• One backend order/lead reconciliation sample.
• Consent mode implementation evidence.

Merchant Center

• Account and data-source list.
• Product status export.
• Diagnostics/issues.
• Shipping, returns and tax setup.
• Product data export with ID, title, description, GTIN/MPN, brand, price, availability, image, product type and custom labels.
• Store quality report.

Business data

eCommerce

• gross revenue;
• discounts;
• refunds/cancellations;
• COGS or contribution margin by product/category;
• shipping subsidy;
• payment fees;
• new vs repeat customers;
• LTV window used for decisions;
• out-of-stock periods;
• promotions and price changes.

B2B

• raw leads;
• valid/qualified/MQL/SQL counts;
• opportunities and pipeline value;
• won deals and revenue;
• lead-to-stage conversion rates;
• average sales cycle;
• sales capacity and serviceable geographies;
• spam/duplicate/unreachable rates.

Audit output

• 74-control health score.
• Historical performance truth with brand and conversion-definition caveats.
• Wasted-spend map.
• Tracking and data quality findings.
• Campaign-by-campaign REACTIVATE / REBUILD-IN-PLACE / ARCHIVE verdict.
• Restart build sheet with budgets, goals, exclusions and launch gates.
• Separate eCommerce and B2B architecture where required.

74-Control Google Ads Audit Scorecard

Scoring

For each check:

• 2: verified healthy.
• 1: partial, uncertain or needs improvement.
• 0: failed.
• N/A: genuinely not applicable; remove it from denominator.

Weighted health score:

sum(category score ratio x category weight)

Category weights:

A. Conversion tracking - 12 checks

1. Primary actions match business outcomes.
2. Micro-conversions are secondary.
3. No duplicate Ads/GA4 counting for the same event.
4. Purchase/lead event fires only on true completion.
5. Transaction or lead IDs support deduplication.
6. Conversion value and currency are correct.
7. Count setting matches the action.
8. Conversion windows are documented and appropriate.
9. Enhanced conversions are active and healthy.
10. Consent mode signals are correct for EEA traffic.
11. Offline qualified lead/sale import works where required.
12. Ads totals are reconciled with backend/CRM.

B. Wasted spend - 12 checks

13. High-spend zero-result campaigns identified.
14. High-spend zero-result ad/asset groups identified.
15. High-spend zero-result products identified.
16. Irrelevant search terms excluded.
17. Brand spend separated and not hiding acquisition performance.
18. PMax brand traffic controlled.
19. Geographic leakage excluded.
20. Poor network/channel/placement traffic diagnosed.
21. Devices with material quality problems diagnosed.
22. Expired promotions and obsolete ads stopped.
23. Broken/out-of-stock landing destinations excluded.
24. Budget is not trapped in campaigns without a valid purpose.

C. Structure - 12 checks

25. Every campaign has a documented purpose.
26. Brand Search is separate.
27. Non-brand campaigns map to business levers.
28. Markets/languages are separated where needed.
29. Product/service themes are coherent.
30. Campaign count does not fragment data unnecessarily.
31. Duplicate product targeting does not create conflicting optimization.
32. PMax asset groups align with products, URLs and creative.
33. Always-on and promotional structures are controlled.
34. Conversion goals match campaign purpose.
35. New-customer settings match business strategy.
36. Shared budgets/portfolio strategies only join economically compatible campaigns.

D. Keywords and queries - 12 checks

37. Keyword themes reflect current buyer intent.
38. Match types are deliberate.
39. Broad match has trusted conversion feedback and review discipline.
40. Search terms are reviewed regularly.
41. Universal negatives are applied at correct scope.
42. Campaign negatives protect intent boundaries.
43. Exact-match winners are captured.
44. Keyword/ad/landing relevance is coherent.
45. Low quality/status warnings are investigated.
46. Auction Insights changes are reviewed.
47. AI Max/DSA expansion has URL and brand controls.
48. PMax search themes and negative controls are coherent.

E. Ads, assets and landing pages - 14 checks

49. RSAs have complete, differentiated assets.
50. Claims and offers are factual and policy-safe.
51. Ad tests isolate one major variable.
52. CTR and conversion economics are reviewed together.
53. Sitelinks are relevant and complete.
54. Callouts and structured snippets are complete.
55. Image, price, promotion, call and location assets are used where relevant.
56. PMax has advertiser-created images in required formats.
57. PMax has purpose-built video.
58. PMax brand guidelines and text customization are reviewed.
59. Demand Gen/Video creative matches channel format.
60. Disapproved/limited ads are resolved.
61. Landing pages match query and ad intent.
62. Mobile speed, form/checkout and conversion event work end-to-end.

F. Settings and governance - 12 checks

63. Company owns the account and Manager hierarchy.
64. Named-user access follows least privilege.
65. 2-Step Verification/security mandates are active.
66. Related managers and inactive users are reviewed.
67. Currency/time zone/account legal entity are correct.
68. Billing and backup payment are healthy.
69. Auto-apply recommendations are controlled.
70. Scripts and automations have owner, purpose and logs.
71. Location option is deliberate.
72. Language/network/ad schedule settings are deliberate.
73. Linked GA4/GMC/YouTube/CRM properties are correct.
74. Material changes are logged and unauthorized changes are absent.

Verdict bands

• 85-100: structurally launchable after issue-specific fixes.
• 70-84: restart in controlled stages.
• 50-69: material rebuild-in-place before scale.
• <50: tracking/governance/structure failure; remain paused.

The score never overrides a critical fail. Broken conversion truth, wrong billing/ownership, policy suspension or non-functional checkout keeps the account paused regardless of numeric score.

Google Ads Operator Training Path

Цел

След завършване операторът трябва да може да:

• създаде правилна Google Ads account hierarchy;
• валидира conversion truth;
• построи Brand и non-brand Search;
• подреди Shopping feed и PMax;
• създаде B2B lead-quality feedback loop;
• управлява Demand Gen tests;
• оптимизира по S.T.A.B.;
• направи evidence-based audit и restart.

Module 1: Account governance

Материали:

• manuals/01-account-governance.md
• Course lessons MGA-02 to MGA-03.

Практика:

• Нарисувай MCC hierarchy.
• Попълни account register.
• Определи owner, users, currency, time zone и linked products.

Pass condition:

• Няма shared login.
• Serving accounts имат ясна business/legal причина.
• Security baseline е документиран.

Module 2: Measurement truth

Материали:

• manuals/02-measurement-and-conversion-truth.md
• MGA-08, MGA-09.

Практика:

• Създай conversion map.
• Маркирай primary/secondary.
• Проследи една test purchase/lead транзакция end-to-end.

Pass condition:

• Няма duplicate primary conversions.
• Value/currency/IDs са валидни.
• Backend reconciliation е доказан.

Module 3: Keyword and intent design

Материали:

• blueprints/01-search-campaign-blueprint.md
• MGA-04, MGA-05.

Практика:

• Изгради keyword themes за един продукт и една B2B услуга.
• Отдели brand, high intent, benefit/problem и competitor.
• Създай initial negative map.

Pass condition:

• Всяка theme има еднакъв intent, ad message и landing page.
• Няма SKAG fragmentation.

Module 4: Search build

Материали:

• blueprints/01-search-campaign-blueprint.md
• MGA-06, MGA-07, MGA-10 to MGA-13.

Практика:

• Напиши 2 RSA variants с една major difference.
• Попълни assets.
• Изгради Search settings sheet.

Pass condition:

• Brand campaign е отделен.
• Geo, network, language, goal и URL controls са изрично избрани.

Module 5: Shopping and feed

Материали:

• blueprints/02-pmax-shopping-multiproduct-blueprint.md
• MGA-14, MGA-33 to MGA-37.

Практика:

• Класифицирай продукти като Hero, Sidekick, Zombie, Villain.
• Напиши optimized titles.
• Попълни feed QA checklist.

Pass condition:

• Product IDs, prices, availability и titles са точни.
• Campaign split-ът следва margin/market/performance lever.

Module 6: Performance Max

Материали:

• blueprints/02-pmax-shopping-multiproduct-blueprint.md
• MGA-15, MGA-38 to MGA-40.

Практика:

• Проектира PMax purpose, listing groups, asset groups и search themes.
• Направи brand/negative/URL exclusion map.
• Създай creative inventory.

Pass condition:

• PMax не harvest-ва brand без да е отчетено.
• Asset group, product set и URL family са подравнени.

Module 7: Demand Gen and creative expansion

Материали:

• blueprints/05-demand-gen-blueprint.md
• MGA-16 to MGA-19, ALL-55.

Практика:

• Създай test matrix audience x channel x creative.
• Напиши naming за ad groups.
• Отдели prospecting и remarketing budget.

Pass condition:

• Всеки test изолира learnable variables.
• Creative format съответства на surface-а.

Module 8: B2B lead generation

Материали:

• blueprints/03-b2b-lead-generation-blueprint.md
• ALL-57 and Lead Generation Playbook PDF.

Практика:

• Изгради ICP/query map.
• Изчисли modeled stage values.
• Проектира offline conversion import.

Pass condition:

• Raw lead не остава permanent optimization target.
• Qualified stage и CRM attribution са ясни.

Module 9: Smart Bidding and scale

Материали:

• manuals/04-smart-bidding-and-scaling.md
• MGA-50, MGA-51, ALL-53, ALL-54.

Практика:

• Оцени 3 campaigns по readiness gate.
• Избери budget или target change, не двете.
• Създай growth/profit cycle.

Pass condition:

• Target идва от achieved performance.
• Conversion quality и lag са включени.

Module 10: Audit and restart

Материали:

• audit-kit/00-dormant-account-restart.md
• audit-kit/01-data-request.md
• audit-kit/02-audit-scorecard.md
• MGA-20 to MGA-49.

Практика:

• Попълни 74-control audit.
• Издай Reactivate/Rebuild-in-place/Archive verdict.
• Подреди top five fixes.

Pass condition:

• Всяка finding има evidence.
• Critical fail overrides numeric score.
• Няма live change без approval.

Certification exercise

Предадени artifacts:

1. Account register and hierarchy.
2. Conversion map with one verified test event.
3. Search build sheet.
4. PMax/feed build sheet.
5. 74-control audit and restart decision.

Операторът минава, когато artifacts са проверени срещу реален account/export и business source of truth.