Robin Tesselaar | Google Ads + Merchant Fix

How to Run Google Ads for E-Commerce (Beginners Tutorial 2026) If you want to grow an e-commerce or Shopify store with Google Ads, you first need to understand how Google Ads actually works and how to set it up the right way. In this Google Ads beginners tutorial, I walk you step by step through creating your first Google Ads campaign, setting up your account, and avoiding the most common beginner mistakes. This complete Google Ads for e-commerce guide shows you exactly how to run Google Ads in 2026, even if you’re starting from zero.

AI Max is not a new campaign type it's a feature suite layered onto existing search campaigns three capabilities: - keywordless intent targeting beyond landing page content - dynamic ad copy tailored to individual search contexts - AI-selected landing pages, not the one you specified the full suite delivers roughly 7% more conversions versus search term matching alone what most accounts haven't configured yet: AI Brief - gemini-powered steering for messaging rules, audience targets, and matching restrictions if you're running AI Max without it, you're handing the wheel to google entirely text disclaimers in final URL expansion: regulated industries can now maintain required disclosure text even when AI selects the landing page AI Max for shopping: dynamic shopping ads that answer conversational queries and capture long-tail searches standard shopping misses the risks most accounts aren't managing: landing page control: final URL expansion sends users to AI-selected pages, which can undermine conversion funnels and any A/B tests running in the background search term opacity: the query range is broader than keyword campaigns, negative keyword hygiene becomes more critical here, not less the auto-upgrade deadline: DSA, ACA, and campaign-level broad match get auto-upgraded to AI Max starting september 2026 advertisers who don't configure settings proactively will have them mirrored from legacy setup the right way to test: - 50/50 traffic split via google's experiments feature - 30 days minimum before scaling - monitor search term reports daily in weeks one and two the accounts that will benefit most are the ones that configure it deliberately before september not the ones that get upgraded on google's timeline

that moment you realize scaling wasn’t the problem, but your system was:

I built a complete Smart Bidding operating system for 2026. Here's what's inside: Step 0: Tracking and value integrity Step 1: Strategy selection Step 2: Matching strategy to your real business Step 3: Learning phase rules (what not to touch) Step 4: Setting targets that don't collapse spend Step 5: Portfolio bidding for low-volume accounts Step 6: How to test without breaking production Step 7: Seasonality adjustments before promos Step 8: Margin data, LTV signals, and value rules Step 9: Phase-by-phase transition roadmap Step 10: Weekly review that preserves stability Want access? Comment "google," and I will send it over. (Must be following)

in 2026, google AI autonomously manages: - bidding - audience selection - creative asset generation including video - URL routing - inventory allocation across channels and now with AI Brief, powered by gemini: - messaging direction - matching boundaries - audience persona targeting the advertiser's job has structurally changed what AI cannot decide for you: conversion architecture - what counts as a valuable conversion, what CRM data to import, how to structure value rules account structure: how many campaigns, how to segment, when to consolidate vs split brand voice - AI Brief lets you set messaging restrictions, but the strategic direction has to come from a human incremental measurement - whether PMax or AI Max is driving net-new customers or just capturing demand you would have won anyway the model won't tell you any of this on its own AI Brief gives you three steering mechanisms: - messaging guidelines: what ads should and shouldn't say - matching guidelines: what searches to prioritize or avoid - audience guidelines: who to reach and with what message variation it only works if you have a clear strategic position to feed into it

five things to stop doing in google ads in 2026: 1. phrase match as a primary strategy AI-powered broad match now reads intent better than phrase match did semantically the only cases where phrase match still earns its place: - highly regulated categories - brand terms where exact phrase control is legally or competitively critical everywhere else, migrate to exact or broad 2. defaulting to standard shopping without testing PMax the 2024 ad rank update removed PMax's auction priority over standard shopping standard shopping gives you cleaner attribution and better brand safety PMax gives you scale and cross-channel reach the right answer is testing both 3. GA4 as primary conversion action GA4 attributes conversions to the date the conversion happened, not the ad click date that misalignment starves smart bidding of the real-time signal it needs native google ads conversion tags are the standard GA4 is secondary 4. letting PMax capture branded queries PMax gravitates toward branded searches by default — easiest conversions, highest reported ROAS the problem: those conversions were largely going to happen anyway separate branded traffic into dedicated brand search campaigns to see the true incremental contribution of PMax 5. over-pinning RSA headlines pinning headlines feels like control what it actually does is prevent google's system from finding which combinations perform fewer, higher-quality assets with minimal pinning outperform stuffed, heavily-pinned ad groups the pattern across all five: tactics that felt like control but were actually hiding signal from the system doing the bidding clean inputs beat clever constraints

7 things breaking most google ads accounts right now: 1. GA4 as primary conversion action GA4 imports don't fire in real time smart bidding is starved of the freshness it needs to bid correctly native google ads conversion tags are the standard GA4 should be secondary 2. no consent mode when users decline cookies and consent mode isn't configured those conversion paths disappear from smart bidding signals entirely not a compliance issue a performance issue 3. no offline conversions lead gen accounts that only fire on form submissions are teaching google to optimize for volume not for revenue 4. phrase match reliance phrase match is caught between two better options: — exact match for tight semantic control — broad match with smart bidding for scale and intent reach most phrase match keywords need to move to one or the other there's no compelling reason to stay in the middle 5. call-only ads still running google began phasing these out in january 2026 new call-only ads can't be created existing ones stop serving entirely by february 2027 the replacement is RSAs with call assets if you haven't migrated yet, the clock is running 6. PMax transparency tools not used google added real controls in early 2026: -first-party audience exclusions to shift spend toward net-new customers -budget reporting with end-of-month projections -full demographic reporting by age and gender -network segmentation in placement reports accounts ignoring these are spending budget on existing customers and calling it growth 7. AI voice-overs already running on your video assets google quietly enabled automatic AI-generated voice-overs on PMax video assets in early 2026 opt-out deadline was march 20 if you missed it, check your PMax video assets now there may already be AI voice-overs running out of brand tone the common thread across all 7: these aren't bidding problems they're infrastructure problems smart bidding performs to the quality of the signals it receives fix the signals first

here's what most Google Ads accounts are missing right now: google's AI Overviews the results on the search page only serve ads from broad match and keywordless targeting exact match and phrase match are architecturally excluded AI Max, which graduated from beta in april 2026, takes this further: - keywordless targeting based on real-time intent signals - dynamic copy tailored to the individual query - AI-selected landing pages, not the one you specified - accounts using it see roughly 7% more conversions at similar CPA starting september 2026, DSA campaigns and campaign-level broad match campaigns get auto-upgraded to AI Max whether you're ready or not what actually changed: old model: keyword triggers bid 2026 model: predicted conversion probability per query triggers bid old model: you pick the destination 2026 model: AI routes to the best page old model: you write the copy 2026 model: AI Brief guides dynamic generation the marketers who understand these updates will outperform the ones who don't

I created the best keyword strategy in 2026. Here's what's inside: • Step 1: what match types mean in 2026 • Step 2: the match type decision matrix • Step 3: four-layer negative keyword architecture • Step 4: weekly negative keyword system • Step 5: search themes for PMAX • Step 6: what AI Max actually is • Step 7: ad group structure • Step 8: DSA campaign audit • Step 9: DSA to AI Max migration • Step 10: the 2026 keyword strategy Want access? Comment "keyword," and I will send it over. (must be following)

google updated merchant center specs on april 14. enforcement starts june 30. what that means in practice: - disapproval rates up ~18% already in Q1 - up to 20% of products are at risk in a typical feed - PMax performance down 35–60% with a weak feed PMax is now the only format for new campaigns. a broken feed is no longer a compliance issue. it's a campaign killer.

I created the best keyword strategy in 2026. Here's what's inside: • Step 1: what match types mean in 2026 • Step 2: the match type decision matrix • Step 3: four-layer negative keyword architecture • Step 4: weekly negative keyword system • Step 5: search themes for PMAX • Step 6: what AI Max actually is • Step 7: ad group structure • Step 8: DSA campaign audit • Step 9: DSA to AI Max migration • Step 10: the 2026 keyword strategy Want access? Comment "keyword," and I will send it over. (must be following)

the average shopify store uses 12 of Google's 170 supported attributes. that gap is the difference between showing up and dominating. most accounts fix tier 1 attributes and call the feed complete. it isn't. tier 1: disapproval or severe visibility loss: title, gtin, price, availability, colors, brand, gender. a price mismatch triggers immediate disapproval. a missing gtin on a branded product reduces visibility before it disappears. tier 2: reduced impression share: description, google product category, product type, additional images, color, size, gender, age group. most accounts stop here. tier 3: competitive differentiation: sale price with effective date triggers price drop badging. lifestyle images feed pmax display placements. short title controls how products appear on display. video_link is new as of april 2026, eligible to serve from june 30. tier 4: ai commerce readiness: product q&a data, compatible accessories, sustainability attributes, real-time inventory count. these don't affect traditional shopping yet. but they determine how often your products get recommended in ai mode and agentic surfaces. the average shopify store uses 12 of google's 170 supported attributes. the gap between tier 1 compliance and tier 4 readiness is the gap between showing up and winning.

google is shutting down the legacy content api on august 18, 2026. if your product feed runs through a spreadsheet upload or old integration, your shopping campaigns go dark that day. the replacement is the merchant api. it's been live since 2024. the operational difference matters: old system: feed updates once a day new system: near-instant sync that gap means paying for clicks on out-of-stock products. every hour the feed doesn't update, you're spending on dead inventory. complex migrations need to be build early to take off properly. august 18 is closer than it looks.

seeing demand gen image ads stuck in review for 7+ days across multiple accounts right now. this isn't an isolated issue. google confirmed it. no fix timeline given. what's likely behind it: google force-upgraded all remaining video action campaigns to demand gen in april. every one of those campaigns, many auto-created by google, hit the review queue at the same time. the system is backed up. the operational problem isn't the delay itself. it's what teams do while they wait: - clone campaigns that haven't served yet - swap assets before review resolves - change bids on campaigns with no data - kill campaigns that never got a fair run that's how you create performance problems that have nothing to do with the original issue. if your demand gen ads are stuck right now: don't touch them. the worst thing you can do to a campaign that hasn't served is treat it like one that failed.

most new google ads accounts fail in the first 30 days. mostly because they launched before the foundation was ready. 2 things have to be locked in before a single campaign goes live. first: conversion tracking. in 2026, standard click-based tracking misses 30-50% of actual conversions. a complete stack looks like this: - ga4 + google ads link (baseline) - enhanced conversions (recovers cookie-blocked conversions) - consent mode v2 (models conversions from users who declined tracking) if your tracking is broken, smart bidding is blind. it's not optimising toward sales. it's optimising toward whatever signals it can find. second: account-level negative keywords. before the first campaign goes live and not after. informational intent, competitor names, job-related terms, low-intent modifiers. these apply globally the moment you add them. every irrelevant click in week one teaches the algorithm the wrong thing. that lesson compounds. the accounts that scale well at 90 days almost always have the same thing in common: we build the foundation from day one. clean tracking. tight exclusions. nothing else. rushed approach is the enemy in the first 30 days.

I made $20+ million from Google Ads. And I noticed that 90% of PMax accounts are optimized completely wrong. So I built a free course to help you: What it covers: • Module 0: clean data baseline • Module 1: where signals live inside Insights • Module 2: 7-day vs 28-day read • Module 3: silent eligibility blocks • Module 4: performance outliers • Module 5: bidding problem or feed problem • Module 6: search term themes • Module 7: negatives and brand exclusions • Module 8: bidding strategy by conversion volume • Module 9: product feed titles • Module 10: weekly schedule Want access? Comment "pmax," and I will send it over. (must be following)

90% of ecom accounts launch pmax wrong. they add every asset google asks for. headlines, descriptions, images, video. then wonder why spend goes to youtube and display instead of shopping. for a new account, that's the wrong order. the correct first campaign is feed-only pmax. when pmax has nothing to serve on youtube or gmail, it concentrates on shopping. highest intent. clearest conversion signal. cleanest data for the algorithm to learn from. without creative assets, google also can't auto-generate visuals that damage your brand before you've earned a single sale. the setup matters too: - start on maximise conversion, not value - new accounts don't have the 50+ conversions needs more time - budget atleast 60-100 euro daily - don't touch it for 2-3 weeks the accounts that fail in month one usually made the same mistake: they gave the algorithm too many options before it had any data. pmax doesn't need creative freedom (yet) in week one. it needs a clean signal. feed-only gives it that.

new google ads account. you launch everything at once. search, pmax, full assets, broad match, smart bidding. then wonder why the first 30 days burned budget with nothing to show. here's the order that actually works: days 1–14: feed-only pmax only. no search. no creative assets. no exceptions. let shopping collect signal before you touch anything else. check search terms daily. add negatives at campaign level. don't change bids. don't rewrite ads. don't panic. days 14–30: add branded search. brand terms convert 3–5× higher than non-brand. it's the fastest way to feed the algorithm clean conversion data without wasting spend on cold traffic. manual cpc or target impression share only. smart bidding on brand terms in a new account starves on data and stops serving. day 30+: nothing else until you have 30–50 conversions. that's not a preference. that's the minimum google's bidding algorithms need to build a reliable model. launch non-brand search before that threshold and smart bidding optimises toward noise. the accounts that scale well at 90 days almost always did the same thing: every big change resets the learning phase. every reset costs you time you can't buy back.