Google Ads & Shopping

If you're building a new category, most of your energy usually goes into creating demand. If you're selling into an established category with active search intent, Google may be the lead channel from day one. The skill is knowing which game you're playing and allocating accordingly.

The Role of Google in Your Channel Mix

The "Meta creates, Google captures" framing is useful, but it's a spectrum, not a binary. A brand inventing a new product category might put 80% of budget into Meta to create awareness, then use Google to catch the search intent that follows. A supplements brand where thousands of people are already searching "best magnesium supplement" every month might put 60% into Google from day one, because the demand already exists.

Most DTC brands sit somewhere in between and shift the balance as they scale. What matters is understanding where your customers are in their journey. If they don't know your product exists, you need to create demand first. If they're already searching for what you sell, capture that intent before spending on awareness.

Meta = Demand Generation

• Finds customers who don't know they want you
• Top-of-funnel awareness
• Leads early budget for novel or category-creating products

Google = Demand Capture

• Captures customers who already want you
• Mid/bottom-funnel intent
• Leads early budget for products with existing search demand

For most DTC brands at $1M-$10M, Google represents 20-40% of paid media spend. One common mistake: treating Google as one channel. It's four distinct channels under one platform.

The Four Pillars of Google Ads for DTC

The numbering is a reference order, not a build order: most brands start with Shopping, add Brand Search only when competitors bid on their name, then layer in Non-Brand and PMax as the data justifies each.

ROAS ranges vary significantly by category, margin structure, and stage. Use as directional starting points.

Recommended Account Structure

These percentages shift by stage. Early brands typically start with Shopping only, add Brand Search when competitors bid on their terms, then layer in Non-Brand and PMax as conversion volume grows. Don't launch all five campaign types on day one. Build the account as the data justifies each addition.

Google Ads Benchmarks for DTC

Reference points to sanity-check your account against, not targets to chase. Every one of these moves with category, margin, brand strength and stage, so use them to spot something badly out of line, then dig into why. Where two of your numbers disagree with these, trust your own contribution-margin maths over any benchmark.

The two that matter most for where you point budget: Shopping's CPC advantage (why it's your core) and branded search's sub-1.0x incrementality (why it's defence, not growth).

Worked Example: Building the Account Over Six Months

(Inputs: $10K/month Google budget, 50% contribution margin before marketing, so break-even ROAS = 2.0x. Middle-of-the-road numbers; swap in your own.)

Months 1-2: Shopping only. Full $10K into Standard Shopping with the feed optimised first. Expect ~5x ROAS on a solid feed. You're profitable from day one and the account is generating the conversion data everything else needs.

Month 3: Add Brand Search (10%). Only because competitors have started bidding on your name - this is defence, not growth. $9,000 Shopping + $1,000 Brand, low bids, no chasing impression share. Brand runs ~12x but it's mostly harvesting demand you created elsewhere: most of those customers were coming anyway, so don't pay to re-acquire them at scale. Blended: (90% x 5) + (10% x 12) = 5.7x.

Months 4-5: Add Non-Brand Search (25%). With 50+ conversions/month and stable economics: $6,500 Shopping + $1,000 Brand + $2,500 Non-Brand. Non-brand starts at ~2x while it learns - barely break-even, but it's the only genuinely incremental search spend in the account, which is why it gets 2.5x the brand budget. Blended: (65% x 5) + (10% x 12) + (25% x 2) = 4.95x. The blended number DROPPED and that's correct: you're buying real new customers instead of polishing the average.

Month 6: Test PMax. Only with 30-50+ conversions/month, a clean feed, and video assets ready. Shift half the non-brand budget across. Watch for cannibalisation: pause PMax for two weeks after a fortnight of running - if revenue barely moves, it was harvesting, not adding.

The pattern: every genuinely incremental addition lowers blended ROAS while increasing real growth - defensive brand spend is the exception, polishing the blend while adding nothing new. Judge each campaign against its own job, and judge the account on nCAC and MER, not the blend.

1. Brand Search: The Trap You Need to Avoid

What It Looks Like

• 10-30x ROAS
• Lowest CPC in your account
• "Best performing" campaign

What's Actually Happening

• Paying for customers who were coming anyway
• Taxing demand you already created
• Every dollar here is a dollar not finding new customers

Brand search decision framework:

• Nobody bidding on your terms? Skip brand search. Your organic listing is free.
• Competitors bidding? Minimal campaign with a low bid. Don't chase impression share - make the store the obvious choice instead.
• Resellers bidding? Address the root cause with exclusive products and a better buying journey.

The Reseller Strategy: Win on Experience, Not Bids

Don't waste budget fighting resellers in Google's auction. Make the .com the only place to get the full experience: exclusive products, colourways, bundles, better warranty and loyalty programmes, guided product selectors, and post-purchase onboarding. When the store offers things nobody else has, the need to bid on your own brand name drops away.

This connects to your broader distribution strategy. See Section 23: Marketplaces & Wholesale for the full picture and Section 11: Website & Conversion for building the direct experience that makes this work.

2. Non-Brand Search

Generic, category-level terms (e.g., "phone mount for bike"). Expensive. CPCs run $2-$6 typical, can hit $8+ in highly competitive categories, with 1-3% conversion rates. The exact numbers vary widely by category, brand strength, landing page quality and intent level of the keyword. Needs strong unit economics.

When Non-Brand Works

• AOV $50+
• Product solves a searchable problem
• Strong landing page with proven CVR
• You've maxed Shopping first

When to Skip It

• Low AOV that can't absorb $2-$8 CPCs
• Product category people don't search for
• No dedicated landing pages per keyword cluster
• Shopping still has room to scale

Start with long-tail, high-intent keywords. Exact/phrase match over broad. Comprehensive negative keyword lists from day one. Dedicated landing pages per keyword cluster.

ROAS targets: 3-5x is traditional, but growth-stage brands with proven 90-day payback can justify 1-2x on genuinely new customers.

3. Google Shopping

The single most important Google channel for DTC. Product image ads convert better than text because the user sees product, price, and brand before clicking.

Feed quality is everything:

Feed tools: Feedonomics (gold standard, premium), DataFeedWatch (mid-tier), Shopify's native Google channel (free, fine for <500 SKUs).

4. Performance Max (PMax)

Google's automated cross-channel campaign. Runs across Search, Shopping, Display, YouTube, Gmail, and Discover. Powerful but opaque. You give up control for scale.

Common PMax mistakes: Launching before you have conversion data. Not providing video (Google auto-generates awful ones). Including brand search and counting it as "PMax performance." Setting ROAS targets too aggressively from day one. One massive asset group instead of segmenting by product category.

While PMax is largely automated, you can add negative keywords at the account level to prevent brand term cannibalisation. Worth doing if your brand search ROAS looks artificially high.

YouTube Ads for DTC

For most DTC brands under $5M, YouTube as a standalone channel is likely premature. Higher creative costs, longer testing cycles, murky attribution. Works best when your product benefits from demonstration, you're spending $50K+/month total, and you have existing video content to repurpose.

Formats: Skippable in-stream (15-60s, pay after 30s), non-skippable (15s max), Shorts ads (vertical, growing), video action campaigns (legacy - replaced by Demand Gen, below).

Creative: Hook in 3 seconds, product within 5 seconds, problem-agitate-solve framework, 15-30 seconds sweet spot, founder-led or UGC outperforms polished spots.

Targeting: Custom intent audiences (based on search terms - YouTube's killer feature), in-market, remarketing, competitor channel placements.

Typical ranges: Cost per View (CPV) $0.03-$0.10, view rate 25-35%, CTR 0.5-1.5%, CPM $8-$20. Results vary by creative quality and targeting.

Demand Gen: The Upper-Funnel Successor

If YouTube is where you run video, Demand Gen is how you run it now. It replaced Video Action Campaigns and pulls your creative across YouTube (in-stream, in-feed, Shorts), Discover and Gmail in one campaign, optimised for action rather than views. The pitch for DTC: it's the closest Google gets to a Meta-style prospecting engine, finding people who aren't searching for you yet, with Google's intent and lookalike data underneath. That makes it the natural test when Meta prospecting plateaus or you want to reduce single-platform dependency.

Two things make or break it. First, attach your product feed so Demand Gen can run shoppable, product-aware ads rather than generic brand video. Second, feed it image assets as well as video, image-plus-video consistently beats video alone here. And segment by placement: Shorts, in-feed and Gmail behave like different channels, so don't read them as one blended number. Same rule as everywhere else in this section: judge it on incremental new customers and blended MER, not its in-platform ROAS, because view-through conversions flatter it.

Google-Specific Attribution

Use Google Ads native conversion tracking (not just GA4 imports). Set up enhanced conversions. Compare Google Ads reported ROAS against your blended ROAS from Shopify. UTM (Urchin Tracking Module) parameters on everything.

Conversion Tracking Infrastructure

None of the bidding in this section works if the conversion signal feeding it is broken. Smart Bidding, Target ROAS, PMax, AI Max, they all optimise towards the conversions you report. Report fewer than really happened (the default state now that cookies are dying) and Google bids timidly and starves your best campaigns. Three pieces of plumbing recover most of that lost signal.

Enhanced conversions. When a customer buys, you pass Google a hashed (one-way encrypted) version of their email or phone from the checkout. Google matches it to a signed-in account and recovers conversions that cookie loss would have dropped. It's a setting plus a small tag change, and it's the single highest-leverage tracking fix. Turn it on first.

Server-side tracking. Instead of firing the conversion tag from the browser (where ad blockers, ITP and consent prompts eat it), you route it through a server container (server-side GTM or your platform's equivalent). More reliable, more durable, and it's the foundation enhanced conversions sit best on. More setup, but worth it once you're spending real money.

Consent Mode v2. If you sell to the EU/EEA or UK, this isn't optional, Google requires it to keep using ad data for those regions, and without it your remarketing and modelled conversions there degrade. It passes the user's consent choice to Google so it can model the conversions it's no longer allowed to observe directly. The underlying consent and cookie obligations live in Section 9: Compliance & Regulatory; this is the Google-Ads-specific plumbing that sits on top of them.

Tools: GA4 for cross-channel journeys, Looker Studio for dashboarding, Triple Whale/Northbeam/Rockerbox for multi-channel view at $20K+/month spend.

For Shopping and PMax, feed health is the foundation everything else sits on. Check your Merchant Center diagnostics regularly.

PMax's automation makes it easy to lose sight of who you're actually acquiring. Keep an eye on new vs. returning customer mix.

Shopping Feed Management

Your Shopping feed is the foundation of every Shopping and Performance Max campaign. Google can only show what's in the feed - and the quality of what's in the feed directly determines your impression share, click-through rate, and cost per click. A mediocre feed with a great bidding strategy will always lose to a great feed with a mediocre bidding strategy.

Most brands set up their Merchant Center feed once, connect it via an app, and never look at it again. That's leaving performance on the table.

Title Optimisation

Product titles are the single highest-leverage element in your feed. Google uses them to match search queries to your products. The default titles pulled from Shopify are almost always too short, too generic, or structured wrong.

The formula: Brand + Product Type + Key Attribute + Colour/Size

You're not writing ad copy here. You're giving Google the information it needs to match your product to the right queries. Front-load the most important terms - Google truncates after about 70 characters in most placements.

Product Type Taxonomy

Google provides a default product category taxonomy, but you should also define your own product_type field using your internal hierarchy. This gives you more control over how products are grouped in campaigns and lets you create more granular ad groups.

Your own taxonomy should reflect how customers think about your products, not how your warehouse organises them. "Phone Cases > iPhone > iPhone 16 Pro" is more useful for bidding than a flat list of SKUs.

Custom Labels

Custom labels are the most underused feature in Shopping feeds. You get five fields (custom_label_0 through custom_label_4) that let you segment products by any criteria you choose. Use them for:

• Margin tier: High, medium, low - so you can bid more aggressively on high-margin products
• Bestseller status: Top 20% by revenue - these deserve higher bids and priority
• Lifecycle: New launch, core range, clearance, seasonal
• Price band: Helps with bid adjustments across different AOV products
• Promotional status: On sale, bundle, limited edition

These labels don't show to customers. They're purely for your campaign structure and bid management. Without them, you're bidding the same on a clearance item with 10% margin as on your hero product with 65% margin.

Bid on Profit, Not Revenue (POAS)

Target ROAS bids on revenue. That's the trap. Tell Google to hit a 4x and it will happily pour budget into cheap, low-margin SKUs that convert easily and starve the high-margin products that actually fund the business. Two products at the same ROAS are not the same sale, and revenue can't tell them apart. Profit can.

POAS (Profit On Ad Spend) fixes this by feeding Google contribution margin as the conversion value instead of order revenue. You calculate margin per product (selling price minus COGS, shipping, transaction fees, returns) and pass that number through the conversion tag, so Smart Bidding optimises towards profit dollars, not top-line. Tools like Profitmetrics, Tracklution and Dotidot wire this into the feed and the tag for you. The halfway house if you can't pass live margin: split the feed into margin tiers with custom labels (you're already doing this, see above) and set a differentiated tROAS per tier, a tighter target on the thin-margin tier, looser where the margin can carry it.

Feed Audit Cadence

• Weekly: Check Merchant Center for disapprovals and warnings. Fix immediately - disapproved products aren't showing at all.
• Monthly: Review title performance. Are high-impression products converting? Are low-impression products buried because of poor titles?
• Quarterly: Full feed audit. Check images, descriptions, attributes, custom labels. Update seasonal labels. Remove discontinued products.

Feed Management at Scale

At Quad Lock, we used DataFeedWatch to manage feeds across multiple regional Shopify stores with region-specific pricing, titles, and product availability. Once you're running Shopping across more than two or three markets, managing feeds manually through Shopify's native Merchant Center integration becomes unsustainable. Feed management tools let you create rules, transformations, and regional overrides without touching your source product data.

Advanced Feed Operations

As your catalogue grows, these become increasingly important:

Supplemental feeds: Use for promotional prices, seasonal attributes, or data corrections without touching your primary feed. This is how you run sale annotations or add temporary attributes without risking your core feed stability.

Variant handling: Ensure colour/size variants have unique, descriptive titles - not just "Blue - S/M/L" but "[Brand] Linen Midi Dress - Sky Blue - Size M". Each variant competes independently in the auction. Generic variant titles lose.

Stock suppression: Automatically suppress out-of-stock or low-inventory items from the feed. Wasted clicks on unavailable products burn budget and hurt your quality score. Most feed management tools can handle this with rules.

Sale annotations: Use the sale_price and sale_price_effective_date fields to trigger "Sale" badges in Shopping results. These significantly improve CTR during promotional periods and cost nothing to implement.

Search Query Management

Search query review is the most underrated discipline in Google Ads. It's not glamorous, and there's no automation that replaces a human looking at what people actually typed before they saw your ad. But it's where you find wasted spend, new keyword opportunities, and early signals of cannibalisation.

Negative Keyword Structure

Build your negative keyword lists in layers:

• Account-level shared lists: Universal negatives that apply everywhere - competitor brands you don't want to bid on, irrelevant industries, "free," "jobs," "DIY," etc.
• Campaign-level negatives: Terms that are wrong for a specific campaign but might be fine elsewhere. A brand campaign should negative out generic product terms to keep traffic clean.
• Ad group-level negatives: Precision sculpting within a campaign. Use these to prevent ad groups from competing with each other for the same queries.

Start building negatives from day one. Every week of unreviewed queries is money spent on irrelevant traffic.

Match Type Strategy

Google has been pushing broad match + Smart Bidding as the default, and for brands with enough conversion data, it works. Broad match gives the algorithm room to find queries you'd never think to target. Smart Bidding adjusts bids in real time based on conversion probability.

But broad match without Smart Bidding is a budget fire. And Smart Bidding without enough data (fewer than 30 conversions per month in a campaign) makes unreliable decisions. If you're early stage or in a low-volume category, exact and phrase match with manual bidding still gives you more control.

The practical approach: start tight (exact/phrase), expand to broad match in campaigns with enough data, and review queries weekly to catch anything the algorithm gets wrong.

AI Max for Search

AI Max is Google's opt-in setting that bolts PMax-style automation onto regular Search campaigns. Switch it on and it does three things: search-term matching (it finds queries beyond your keywords, like broad match on steroids), text customisation (it rewrites and assembles headlines per query), and final-URL expansion (it can send traffic to whichever page on your site it thinks fits the query best). Google's own headline is roughly 14% more conversions at a similar CPA, rising further for campaigns that were heavily exact and phrase match before. Treat that as a vendor number under ideal conditions, not a guarantee: independent retail data shows a much wider real-world spread, with plenty of accounts seeing flat or worse CPA. Test it, measure it, don't assume the lift.

The thing nobody tells you: AI Max makes your guardrails more important, not less. The moment you let Google match broader queries and pick its own landing pages, your brand-defence and waste-control discipline becomes the only thing standing between you and budget leaking into junk traffic. Tighten the brand negatives, lock URL expansion to the pages you actually want ranking, and put search-query review on a weekly cadence (the rest of this Search Query Management section is now non-negotiable, not optional). Google has signalled this is the direction of travel for Search, with auto-upgrades arriving from late 2026, so learn it on your terms now rather than waking up to it switched on for you.

Mining Queries for Opportunities

Your search query report isn't just a list of waste to eliminate. It's a source of new keywords, new ad group ideas, and even product development signals:

• High-converting queries you're not explicitly bidding on, create dedicated ad groups for them
• Comparison queries ("X vs Y"), these convert at 2-3x the rate of generic terms and signal high purchase intent
• Long-tail queries with purchase intent, "best waterproof phone mount for motorbike" is more valuable than "phone mount"
• Queries revealing customer language you hadn't considered, use their words in your ad copy

Brand Queries in Search Reports

When reviewing search queries, separate brand from non-brand. The brand search trap covered earlier applies here too, high ROAS on brand terms doesn't mean high incrementality. Filter brand queries out of your performance analysis to see what Google is actually doing for you.

Review Cadence

• Weekly: High-spend campaigns. Look for new negatives, wasted spend, emerging query patterns.
• Monthly: All active campaigns. Update shared negative lists. Mine for new keyword opportunities.

Performance Max Governance

Performance Max is Google's black-box campaign type. It runs across Search, Shopping, Display, YouTube, Gmail, and Discover simultaneously, using Google's AI to allocate budget and target users across all those surfaces. You hand Google your product feed, some creative assets, and a conversion goal, and it does the rest.

When it works, it works well. When it doesn't, you have almost no visibility into why.

When PMax Works

• Sufficient conversion volume: 30-50+ conversions per month minimum. Below that, the algorithm doesn't have enough data to optimise reliably.
• Strong creative assets: PMax assembles ads dynamically from the assets you provide. Bad images, weak headlines, and generic descriptions mean bad ads across every surface.
• Clean product feed: Everything in the feed management section above applies double here.

When Standard Shopping Still Wins

• Low conversion volume: PMax needs data density to learn. Standard Shopping with manual bidding gives you more control early on.
• Brand protection concerns: PMax bids on your brand terms by default. If you're already running brand campaigns, PMax cannibalises that traffic and claims the credit.
• Need for transparency: If you need to explain to a board or investors exactly where your Google spend is going, PMax's reporting limitations are a real problem.

Shopping vs PMax: Decision Framework

Most brands should start with Standard Shopping. Move to PMax when you have 30-50+ conversions per month consistently, strong creative assets, and a clean feed. Many scaled brands run both: Standard Shopping for control on core products, PMax for incremental discovery.

Cannibalisation: The Biggest Risk

Left unchecked, PMax will eat your existing brand search and standard Shopping volume. It targets the lowest-hanging fruit first, which is usually people who already know your brand. Your PMax ROAS looks incredible, while your other campaigns suffer, and your blended efficiency stays flat or declines. A high PMax ROAS is not proof of incrementality. It can equally mean PMax is harvesting demand you'd already created elsewhere.

Guardrails to Set

• Brand exclusions: Exclude your brand terms from PMax (this is now available in the settings). Force PMax to find incremental traffic.
• Asset group structure: Don't dump everything into one asset group. Segment by product category so you can read performance at a meaningful level.
• Audience signals: These aren't targeting, they're hints to the algorithm about who to start with. Use your customer lists, high-value segments, and website visitors as signals.
• Monitor new vs. returning customer mix: If PMax is mostly converting existing customers, it's remarketing, not prospecting.

Reading PMax Reporting

Google's reporting for PMax is deliberately limited. You can see asset group performance, some audience breakdowns, and conversion data - but you can't see search queries (beyond a limited insights report), placement-level data, or how budget was split across channels.

This means you can't optimise PMax the way you optimise standard campaigns. Instead, evaluate it at the blended level. Is your overall MER improving with PMax on? Is your nCAC stable or improving? Is new customer acquisition actually growing, or is PMax just claiming credit for existing demand?

Meta and Google will likely be your two primary channels for a long time. But eventually you'll hit diminishing returns or want to reduce platform dependency. Section 16: Other Marketing Channels covers when and how to diversify.

Section 15 Checklist