8+ years growing brands on KPIs, now with AI
Google Shopping That Moves Your MER
Feed-first structure, brand isolation, and PMax discipline that lift blended revenue, not just platform-reported numbers.
Google Ads Partner · 8+ years performance marketing · Meta & TikTok
Get your free Growth Gap Analysis
Senior strategist review · response within one business day
Prefer to reach out directly?
The Challenge
The Google Shopping Problem Nobody Talks About in Your Weekly Standup
Your Performance Max campaign is reporting a 7x ROAS. Your Standard Shopping is at 5.2x. The account looks healthy. Then you open Shopify, divide total revenue by total ad spend, and your MER is 2.6x. Somewhere between the platform report and the business reality, more than half the story disappeared.
That gap is not a rounding error. It is structural. PMax is built to claim credit; it captures branded search, retargeting, and discovery traffic, then reports a blended number that makes all three look like prospecting wins. When you isolate non-brand incremental performance, the real number is often 2x–4x, not the 7x on the dashboard.
Meanwhile, your Merchant Center feed is quietly doing damage you can't see. A GTIN mismatch doesn't just disapprove one SKU; it can trigger account-level warnings. A price discrepancy between your feed and your landing page pulls products from auction without a single notification. And your 40 new SKUs from last season's launch? PMax decided they were too risky to spend on, so they've never gotten a single impression. They're zombies now.
You're sophisticated enough to know something is wrong. You see the platform ROAS and the MER diverging. You know PMax is a black box with a marketing budget. What you need is an operator who knows exactly where the levers are, and has the discipline to use them.
The Opportunity
The Brands Winning on Shopping Right Now Are Doing Three Things You're Probably Not
Since October 2024, Google no longer auto-prioritizes PMax over Standard Shopping. Ad Rank now determines which campaign serves, which means a well-structured Standard Shopping campaign can beat your PMax campaign for the same query. Most brands and agencies haven't rebuilt their structure around this. That is a real, exploitable gap.
The operators pulling the strongest blended ROAS numbers right now are running a deliberate hybrid: Standard Shopping on top-revenue SKUs with tight product group segmentation, PMax handling long-tail and discovery, brand traffic isolated into its own Standard Shopping campaign with negatives enforced everywhere else. The 70/30 or 80/20 PMax-to-Standard split isn't a default; it's a decision made per account based on conversion volume and margin tier.
Feed optimization is the highest-leverage, most-ignored work in Shopping. Google's own data shows a 20% average increase in clicks when correct GTINs are added. Title structure (specifically what lives in the first 70 characters of a 150-character limit) determines whether your product appears for high-intent queries or gets buried behind brands who bothered to front-load brand, product type, and the differentiating attribute. Most feeds are written like product descriptions. The best feeds are written like keyword strategies.
There is also a profit-first bidding layer that most accounts are not running: custom labels segmented by margin tier, Conversion Value Rules that weight higher-margin SKUs in Smart Bidding signals, and POAS-aware bid structures that stop optimizing for revenue and start optimizing for contribution. The brands that figure this out stop celebrating high ROAS on low-margin SKUs.
What Most Get Wrong
What Most Google Shopping Operators (and Most Agencies) Get Wrong
Trusting PMax ROAS as the performance signal
PMax consistently reports the highest ROAS of any campaign type (often 5x–12x) because it bundles branded search, retargeting, and prospecting into one number. When you isolate incremental non-brand performance, that number typically drops to 2x–4x. Agencies that optimize to platform ROAS instead of MER are optimizing a metric that flatters the account, not the business.
Letting PMax run brand traffic without isolation
PMax optimizes toward the highest probability of conversion. Branded queries are easy wins: low CPC, near-certain conversion. Without explicit brand exclusion lists and a dedicated Standard Shopping brand campaign, PMax quietly shifts budget toward branded traffic, inflates reported ROAS, and leaves you with no clear read on whether you can profitably acquire new customers. A controlled Haus.io experiment found that excluding brand terms from PMax drove 24% more incremental revenue.
Ignoring feed health until something breaks
Price and availability mismatches between feed and landing page are the number one cause of product disapprovals. Incorrect or fabricated GTINs don't just disapprove individual products; they can trigger account-level warnings. Most operators only open GMC Diagnostics when a campaign suddenly drops. By then, the erosion has been running for weeks.
Letting new and niche SKUs become zombies
PMax allocates spend toward products with conversion history. A new SKU or niche item has none, so the algorithm avoids it, so it never builds history: a self-reinforcing cycle that leaves 20–40% of a catalog invisible. The fix is routing zombie SKUs into a Standard Shopping campaign on Manual CPC or Maximize Clicks, where Smart Bidding's data requirements don't apply. Most accounts never do this.
Running PMax without campaign-level negative keywords
Campaign-level negative keywords in PMax only became available in January 2025. Operators who haven't updated their structure since then are still spending on search terms they would never consciously bid on. This is not a minor inefficiency; it is uncontrolled spend on queries with no purchase intent, and it inflates impression volume while suppressing conversion rate.
Why Now
The Window Is Open, But It Closes When Your Competitors Rebuild Their Structure
Google's October 2024 change to PMax prioritization was the biggest structural shift in Shopping in years. The old playbook (dump everything into PMax, set a tROAS, and let Google decide) no longer works the same way. Ad Rank now governs which campaign serves, which means a disciplined Standard Shopping setup can outcompete a lazy PMax campaign for the same query on the same product.
Most brands are still running the old playbook. Most agencies are still selling it. The operators who rebuild their hybrid structure now (Standard Shopping on hero SKUs, PMax on discovery, brand fully isolated, feed titles rebuilt around the first 70 characters, zombie SKUs resurrected into Manual CPC campaigns) will own the auction positions that their slower competitors vacate.
The AI layer compounds this. Feed optimization at scale (testing title structures across hundreds of SKUs, flagging GTIN coverage gaps, identifying which custom label segments are bleeding margin) is work that used to require a full-time feed manager. Applied correctly, AI runs this as a continuous background process, not a quarterly project. That means the gap between a well-run Shopping account and a neglected one widens faster than it ever has.
The brands that move in the next 90 days will be running structurally superior accounts before the next peak season. The ones that wait will spend Q4 trying to catch up inside a learning phase.
The Mechanism
Where AI Creates Real Edge in a Google Shopping Account
Real productivity, not AI theater. Here's where it actually moves a number for google shopping brands.
Merchant Center Feed Optimization
What AI does: AI audits every product title against the 70-character display threshold, flags missing or mismatched GTINs against Google's validation database, identifies price-and-availability discrepancies before they trigger disapprovals, and generates title variants front-loaded with brand, product type, and differentiating attribute, across hundreds or thousands of SKUs simultaneously.
The result: Feed health maintained continuously, not repaired quarterly. GTIN coverage improvements that Google's own data associates with a 20% average click increase. Title structures that capture high-intent query traffic competitors' feeds miss.
Why it matters here: In Google Shopping there is no keyword field. Your feed title is your bid. A manually managed feed drifts. An AI-monitored feed compounds.
Analytics & Attribution
What AI does: AI reconciles platform-reported ROAS against Shopify revenue and total ad spend to produce a true MER figure weekly. It flags the gap between PMax-reported ROAS and blended MER, identifies which campaigns are inflating the account-level number through branded traffic cannibalization, and surfaces conversion volume per campaign to determine whether Smart Bidding thresholds are met before any structural split is made.
The result: A weekly MER reconciliation that tells you whether the business is working, not just whether the ad account looks good. Attribution inflation quantified, not guessed at.
Why it matters here: A 7x PMax ROAS and a 2.6x MER coexist in the same account all the time. The gap is the business problem. You can't fix what you can't see.
Digital Ads (Campaign Structure & Bid Management)
What AI does: AI monitors tROAS performance against conversion volume thresholds per campaign, flags when a campaign falls below the ~60-monthly-conversion floor where Smart Bidding becomes unreliable, identifies zombie SKUs by impression-share-to-conversion ratio, and surfaces brand query share in PMax Search Term Insights to verify exclusion lists are holding.
The result: Structural decisions (when to split Standard Shopping from PMax, when to move a SKU to Manual CPC, when to tighten or loosen tROAS) made on live data, not monthly reviews.
Why it matters here: PMax's learning phase burns 40–60 days of budget when you make structural changes blind. AI-informed decisions reduce the number of expensive resets.
Creative (Asset Group Optimization)
What AI does: AI generates and tests multiple headline, description, and image asset combinations per PMax asset group, monitors asset-level performance ratings inside Google Ads, and flags auto-generated assets that may create compliance risk: copy Google writes on your behalf that you may not control or even know exists until a policy violation surfaces.
The result: Asset groups that improve over time rather than stagnate at launch. Auto-generated asset risk caught before it becomes a suspension event.
Why it matters here: PMax auto-generated assets have caused compliance violations operators discovered only after account suspensions. Monitoring this is not optional for a brand running significant spend.
Conversion Optimization
What AI does: AI audits landing pages against feed-level product data to catch price and availability mismatches before GMC flags them as disapprovals. It also reviews product landing page conversion rate by SKU segment (identifying whether low Shopping conversion rate is a bid problem, a feed problem, or a page problem) and surfaces which product groups are pulling AOV down inside PMax.
The result: Disapprovals caught pre-auction rather than post-drop. Conversion rate issues diagnosed at the SKU level, not the campaign level.
Why it matters here: PMax tends to optimize toward SKUs that convert easily, which often means lower AOV. Standard Shopping holds AOV better, but only if the landing pages it sends traffic to convert. Knowing which is which changes the structural decision.
Ready to see what this looks like for your google shopping brands business?
No obligation. A senior strategist will show you exactly where the wins are.
The Strategy
How a Google Shopping Account Should Actually Be Built in 2025
The foundation is the feed. Before any campaign structure conversation, the Merchant Center feed needs to be audited: GTIN coverage verified against Google's database, title structures rebuilt to front-load brand, product type, and the differentiating attribute in the first 70 characters, custom labels assigned by margin tier and velocity, and a Content API or high-frequency fetch schedule confirmed so price and availability stay current. A structurally correct campaign running on a drifting feed is still a broken account.
Campaign structure follows conversion volume. For accounts generating 500+ monthly conversions, a hybrid setup is the right architecture: Standard Shopping on top-revenue SKUs with tight product group segmentation and explicit tROAS targets, PMax handling long-tail discovery and new audience expansion. The split is typically 70/30 or 80/20 PMax-to-Standard by budget, but the Standard Shopping campaigns carry the strategic weight: margin protection, brand isolation, zombie SKU resurrection.
Brand isolation is non-negotiable. PMax gets brand exclusion lists. A dedicated Standard Shopping campaign runs brand terms at lower priority, capturing branded demand without contaminating the non-brand performance signal. Search Term Insights in PMax are audited weekly to verify brand query share on non-brand campaigns is near zero. This is the only way to know whether you can profitably acquire new customers, not just serve existing ones.
Zombie SKUs get their own Standard Shopping campaign on Manual CPC or Maximize Clicks. No Smart Bidding: these products lack the conversion history Smart Bidding requires to function. The goal is impression volume and initial conversion data, not ROAS. Once a SKU has enough history, it graduates to a Smart Bidding campaign. Without this process, 20–40% of a catalog never gets a fair test.
The governing metric is MER: total Shopify revenue divided by total marketing spend across every channel. Shopping ROAS and PMax ROAS are daily tactical signals. MER is the weekly business check. When the two diverge significantly, the gap is the diagnostic: attribution inflation, brand cannibalization, or spend on channels that aren't being measured.
The one number that governs this
North star: Blended ROAS / MER (total revenue ÷ total marketing spend). Daily levers: Shopping campaign tROAS and PMax asset group performance. The gap between platform ROAS and MER is the number that tells you whether the account is working or just looking good.
How We Help
What We'd Actually Do With Your Google Shopping Account
We'd start where the leverage is highest and work outward. For most Google Shopping accounts, that means the feed and attribution before any campaign restructuring, because rebuilding campaign structure on top of a drifting feed or inflated attribution numbers resets the clock without fixing the problem. Here is the sequenced engagement:
Merchant Center Feed Audit & Rebuild
GTIN coverage verified, title structures rebuilt around the 70-character display threshold, custom labels assigned by margin tier and clearance status, Content API or fetch frequency confirmed. This runs before any campaign changes.
Attribution & MER Reconciliation
We establish your true MER baseline (Shopify revenue divided by total marketing spend) and quantify the gap between platform-reported ROAS and business-level performance. Attribution inflation is measured, not assumed.
Campaign Structure Rebuild (PMax + Standard Shopping Hybrid)
Standard Shopping campaigns built around top-revenue SKUs with tight product group segmentation and explicit tROAS targets. PMax scoped to long-tail discovery. Budget split determined by your account's conversion volume, not a template.
Brand Isolation & Exclusion List Management
Brand exclusion lists enforced in PMax. Dedicated Standard Shopping brand campaign built with campaign priority controls. Weekly Search Term Insights audit to verify brand query share on non-brand campaigns.
Zombie SKU Resurrection
New and niche SKUs routed into a Standard Shopping campaign on Manual CPC or Maximize Clicks. Conversion history built before any Smart Bidding is applied. Graduation criteria defined so SKUs move into the main structure when they're ready.
AI-Assisted Feed Monitoring & Asset Group Optimization
Continuous GMC Diagnostics monitoring for disapprovals, price mismatches, and GTIN errors. PMax asset group performance tracked and rotated. Auto-generated asset review to catch compliance risk before it becomes a suspension.
Paid Media Management (Google Ads)
Weekly tROAS adjustment cadence, conversion volume monitoring per campaign, campaign-level negative keyword maintenance post-January 2025 update, and monthly MER reconciliation against Shopify revenue.
Who's Behind This
Who we are, and what makes us different
Sagum is a performance marketing agency founded in January 2017 in St. George, Utah. We've spent 8+ years growing real brands and being judged on KPIs, not vanity metrics.
We deliberately limit how many clients we take so each one gets senior attention. We treat your numbers like our own, we never run generic playbooks, and your strategy is built for your business, because shouldn't your brand's marketing be custom to your brand?
Sagum.ai is our AI arm: the same proven operators now build AI into the work wherever it creates real edge, not as theater, but as leverage applied with discipline.
- 8+ years growing brands on performance KPIs, not vanity metrics
- Limited client roster, with senior attention on every account
- An extension of your team; your success is tied to ours
- Custom strategy per brand, never a generic playbook
- AI built in where it moves a number; judgment over hype
“Sagum is a performance marketing agency that's spent 8+ years growing brands by treating their numbers like our own. We take on few clients, never run generic playbooks, and now build AI into the work wherever it creates real edge, not hype. Your strategy is built for your business, and our success is tied to yours.”
“After six years, Sagum is our most important partner: trusted, communicative, and caring about our business as if it's their own.”
Proof
Broke a 2-year ROAS plateau with +115% ROAS at the same spend
House of Jade
Challenge
House of Jade had been running paid search and Shopping campaigns for two years without moving their ROAS meaningfully. Spend was stable, platform numbers looked acceptable, but the account had plateaued: the kind of plateau where nothing is visibly broken but growth has stopped.
What we did
We rebuilt the campaign structure around the products and audiences actually driving contribution, fixed the attribution layer so optimization decisions were based on real performance signals, and restructured bidding to stop optimizing for revenue on low-margin SKUs.
Result
The result was a 115% ROAS improvement at the same spend level, and their biggest, most profitable Q4 on record. The spend didn't change. The structure and the signals did.
Your MER and Your Platform ROAS Shouldn't Be This Far Apart
No obligation. We'll review your feed health, campaign structure, and MER gap in one session, and tell you exactly where the leverage is, whether we work together or not.
Sagum · January 2017 · St. George, Utah · 8+ years