Ecommerce PPC management: the playbook, including PMax vs Search

What ecommerce PPC management involves

Ecommerce PPC management is the weekly work of turning a product catalog and an ad budget into profitable orders on Google Ads: structuring campaigns, feeding Merchant Center clean product data, making sure the conversion Google optimizes for is a real purchase with its real value, splitting budget between Search and Performance Max, and reporting on revenue rather than clicks. In the accounts we inherit, a stall usually traces to tracking or structure long before it traces to bids.

Google Ads for ecommerce asks more of your data than lead generation does. A lead generation account can live on keywords and a form. A store cannot. Every product needs a price, image, identifier and landing page that match, and every sale must reach Google with its value. Get those right and bidding has something true to work with. If those are wrong, automation scales the mistake.

The job runs as five loops at different speeds:

• Structure, quarterly: which campaigns exist and how they avoid bidding against each other.
• Measurement, weekly: purchase events, transaction IDs, values and primary actions.
• Feed, continuously: titles, GTINs, variants and custom labels.
• Budget and bids, by season: targets, pacing and where extra budget goes.
• Reporting, monthly: revenue, ROAS and contribution, tied to what finance sees.

If you are still deciding whether to run this in-house or bring in help, our guide on when to hire a PPC expert covers the signs and the skills to vet.

Account structure that scales

A scalable ecommerce Google Ads account has four parts with separate jobs: a brand Search campaign, a non-brand Search campaign built on exact and phrase keywords for your proven categories, a Performance Max campaign driven by the Merchant Center feed with your brand excluded, and a small test cell. Eligible exact match Search keywords are prioritized over Performance Max for identical queries, so the structure itself decides who serves what.

The structure works because of how Google decides which campaign enters the auction. If a search query is identical to an eligible exact match keyword in a Search campaign, that Search campaign is prioritized over any phrase or broad keyword and over Performance Max.⁴ Phrase and broad keywords share priority with an identical Performance Max search theme, then relevance and Ad Rank decide. So the Search campaign owns the queries you have proven, and Performance Max works on everything else.

Two caveats matter. First, the rule applies only to eligible keywords. A budget-limited Search campaign, low search volume status or disapproved ads can hand a keyword to Performance Max.⁴ Second, brand. Brand exclusions block Performance Max from brand searches including misspellings, and Google recommends them when brand runs in a separate campaign.⁵ Without them, Performance Max can claim cheap brand conversions and look better than it is.

Keep asset groups tied to real product groupings, and use product filters so the items in an asset group match its landing page, as Google advises.³ Group by category first, and by margin tier once the feed carries custom labels.

Performance Max vs Search: which to use, and when

Use both, with different jobs. Search is the precision engine for queries you already know convert, and its eligible exact match keywords are prioritized over Performance Max for identical queries. Performance Max is the reach engine: one campaign that serves across YouTube, Display, Search, Discover, Gmail and Maps from your feed and assets. On Astrid & Miyu, Search returned 10.33x on 28% of spend and Performance Max 4.66x on 62%.

The question usually comes from an account where one engine is eating the other. The fix is rarely to switch one off. Give each engine its job and fence the overlap with exact match keywords in Search, plus brand exclusions and negative keywords, which Google lists as Performance Max controls.³

Read the ROAS gap correctly. Search sits closest to purchase intent, so its higher ROAS does not mean Performance Max is failing. At Astrid & Miyu, Performance Max still produced £4.1M of the £8.49M total, and its spend tripled for the record November.¹ The mistake is judging Performance Max against Search. Judge each against its own break-even (see the reporting section).

When to lean on Search

• You have categories with steady, known queries and a clear landing page for each.
• You need tight control of copy and spend per query, say on a hero product or promotion.
• Your conversion history is thin. Target ROAS in Search and Shopping campaigns needs at least 15 conversions in the past 30 days.⁶

When to lean on Performance Max

• Your feed is clean and your catalog is wide, so the feed does the targeting.
• Your purchase tracking and values are trustworthy, because it optimizes toward whatever you say counts.
• You want reach beyond Search, with brand exclusions in place.

Fixing conversion tracking (the IKEA example)

Fix tracking before anything else, because Smart Bidding and Performance Max optimize toward whatever conversion you mark as primary. IKEA UAE's inherited Performance Max campaign counted 475K plus app opens at AED 0.04 each and reported a 28.1x ROAS that was not real purchase revenue. Resetting the conversion to actual purchases, averaging AED 776 each, produced a verified 6.12x.

The IKEA account looked excellent on a dashboard. The 2024 setup was optimized for app installs, so the algorithm found hundreds of thousands of cheap app opens and reported a return that could not prove incremental sales.² A high ROAS on the wrong event is worse than a modest one on the right event, because it tells the bidding system to keep buying the wrong thing.

What we changed, in order:

• Conversion definition. App opens stopped counting. Actual purchases became the conversion, with their value.
• Bidding. Maximize Conversion Value replaced Target CPA plus Target Impression Share, so the system optimized for revenue rather than volume.
• Scale, in steps. Monthly impressions grew from 9M to 21M over 11 months while ROAS held at 6.12x through each step.²

A tracking checklist for any store

• Purchase is the only primary action. Google uses primary conversion actions for bidding; secondary actions are for observation only and do not drive bids unless placed in a custom goal.⁸ Add to cart, app opens and newsletter sign-ups belong in secondary.
• Every order carries a unique transaction ID. If a customer reloads the confirmation page, the tag can fire twice for one order. With a transaction ID, Google Ads recognizes the second as a duplicate.⁹ A hardcoded or static ID causes significant undercounting, because Google Ads may process only its first instance.
• GA4 gets the same ID. The GA4 purchase event's transaction_id parameter exists to avoid duplicate purchase events.¹¹ Use the same ID in both so order counts reconcile.
• Values are real. Google Ads predicts future conversions and their values from the conversion values you report, so a wrong value trains the bidding.⁶ Send order value in the store's currency, and decide once whether it includes tax and shipping.
• Leave the learning phase alone. Google describes a standard 7 to 14 day learning phase for Smart Bidding and warns that changing budgets, targets or conversion goals during it resets the window.⁸ Batch your changes.

A quick audit: compare last month's purchase count in your store, GA4 and Google Ads. If the gap is bigger than attribution windows and returns explain, fix tracking before you touch a bid.

Feed optimization in Merchant Center

The Merchant Center feed is the targeting for Shopping ads and a core input to Performance Max, so feed quality sets the ceiling on both. The essentials are titles that match the landing page and name the variant, correct GTINs (Google warns that products submitted without an available GTIN may have limited visibility), and custom labels that let you split bidding by margin, season or clearance.

Performance Max combines your data feeds and assets with Google AI to decide where to serve.³ Without keywords, titles and descriptions do much of a keyword list's job, which is why feed work often moves results faster than bid work.

In the accounts we audit, custom labels are usually the most underused lever in ecommerce PPC campaign management. Google describes them as labels to help organize bidding and reporting in Shopping campaigns, and allows up to five per product.¹⁰ Once margin tier is a label, you can build Shopping product groups by margin and give each its own target. In our practice we apply the same labels to Performance Max asset groups through product filters. Without labels, a low-priced accessory and a high-margin hero product compete for budget under one target.

Fix disapprovals and missing identifiers before rewriting titles, because an ineligible product does not serve at all, and keep feed and site in sync on price and availability. If the product pages themselves need work, start with speed and conversion fixes on the store.

Budget allocation

Allocate budget by job and by season, not by channel habit. Keep Search funded so it never goes budget-limited on proven queries, give Performance Max the scaling budget, cap test cells, and move money toward peak weeks in advance. Astrid & Miyu tripled Performance Max spend for the record November gifting peak, and the program delivered £2.08M that month at 5.40x ROAS.

Three rules from that account generalize.

• Fund Search to its ceiling on proven queries. If the Search campaign is budget-limited, Performance Max may occasionally serve on those exact match terms.³⁴
• Sequence budget to the season. At Astrid & Miyu, a September dip was answered by pulling October budget forward, which recovered ROAS to 6.29x and set up November, when Performance Max spend tripled and the program delivered a record £2.08M at 5.40x ROAS.¹ The account had plateaued around £600K a month before the rebuild.
• Cut what does not earn its place. Demand Gen was returning 1.05x when we inherited the account.¹ Shopping and Demand Gen took the remaining 10% of spend across the period (our arithmetic), in a supporting role.

Know how Google spends a daily budget before you plan a peak. A campaign can spend up to twice its average daily budget on a given day, but over a month it will not spend more than 30.4 times that average.⁷ Raise budgets a few days before a peak, and do not stack target changes inside the learning phase.⁸ For a new Performance Max campaign, Google's budget panel offers Low, Medium and High suggestions, and the lowest is the minimum it recommends.³

Reporting that ties to revenue

Report on purchase revenue, ROAS and the margin left after ad spend, and split every number by engine and by brand versus non-brand. A blended ROAS that looks healthy can hide a brand campaign carrying a weak prospecting one. Set a break-even ROAS from your gross margin, then judge Search and Performance Max each against it rather than against each other.

Start with the break-even line. At a 40% gross margin, each dollar of revenue leaves 40 cents to cover ad spend, so break-even ROAS is 1 divided by 0.40, or 2.5x (illustrative arithmetic). Below that line, the order loses money. Each margin tier gets its own line, another reason to carry margin in a custom label.

In Google's own example, $5 of sales per $1 of ad spend is a 500% target ROAS.⁶ Maximize Conversion Value spends the whole budget for the most value; Target ROAS prioritizes a specific return. Move to the second once you know your line and have the volume.

What a monthly report should show

• Revenue, spend and ROAS by engine, with brand and non-brand on separate lines.
• Performance Max by channel. Google's channel performance report shows how a Performance Max campaign delivers across Google's channels and inventory, with drill-downs into ad formats including ads with product feeds.¹²
• Orders reconciled against the store. Purchases in Google Ads, GA4 and your ecommerce platform, with the gap explained.
• Average order value and conversion rate, to separate traffic quality from basket size.
• Contribution after ad spend. Revenue times margin, minus spend. This is the number finance recognizes.

Watch for the IKEA pattern: ROAS climbing while store revenue does not. In our experience, that is usually a conversion definition problem.

Case results: Astrid & Miyu and IKEA

Astrid & Miyu, a fashion and apparel brand, had a four-channel program stuck near £600K a month; restructuring it into a Search and Performance Max two-engine system produced £8.49M at 5.95x blended ROAS. IKEA UAE had a purchase tracking problem; resetting conversions to real purchases produced AED 34.9M of 2025 revenue at a verified 6.12x ROAS.

The Astrid & Miyu lesson is structural. Re-sorted so Search captured proven demand and Performance Max scaled reach, the program delivered a record November without a peak-season ROAS collapse. The full breakdown is in the Astrid & Miyu case study.

The IKEA lesson is about measurement. The headline ROAS fell from 28.1x to 6.12x and the account got better, because the new number was real. Once the account optimized toward purchases, monthly impressions grew from 9M to 21M over 11 months and monthly revenue roughly tripled between February and November 2025. The details are in the IKEA UAE case study.

If your account shows either pattern, a plateau across too many campaign types or a ROAS that does not match the bank, start with the matching fix. If you want a second pair of eyes, our PPC management team can run that audit on your account in the same order this playbook follows: tracking, structure, then feed.

Kritika Saxena

Head of Growth and Strategy, Resourcifi

Kritika Saxena is Head of Growth and Strategy at Resourcifi. She owns growth, marketing and AI search strategy for the company and its clients, and she reviews the marketing guides we publish for accuracy.

Resourcifi on LinkedIn →

Sarah Kamal

Reviewer. Head of Client Relations, Resourcifi

Sarah Kamal is Head of Client Relations at Resourcifi. She owns the client experience from kickoff through delivery and reviews the response commitments and engagement details we publish.

Resourcifi on LinkedIn →

Keep reading

Related guides worth your time

← →

Paid & growth When to hire a PPC expert: signs, skills, and what to expect Know when to hire PPC management services: the signals, skills to vet, and what a strong engagement delivers. Clutch 4.9-... Read guide → AI in marketing AI for Marketing How to use AI in marketing as an operating model: five workloads, brand-voice guardrails, AIEO visibility, and personaliz... Read guide → AI in marketing AI Use Cases in Marketing AI marketing use cases that actually ship in 2026: the production use cases by function, real adoption numbers, and the h... Read guide → AI in marketing What are AI search optimization tools? A comparison of trackers, prices and limits What are AI search optimization tools? See how they track brand visibility in ChatGPT, Perplexity and AI Overviews, with... Read guide → Mobile & apps 25 mobile app ideas worth building, with costs and revenue models 25 mobile app ideas across AI, health, fintech and commerce, with the core features, a sourced build cost range and the m... Read guide → Agency & white-label Agency AI Pricing What AI development cost depends on: ranges by phase (POC $75k to $150k), the six pricing models, scoping with a paid POC... Read guide → Agents & RAG Agentic RAG: When to Use It and How to Build It Agentic RAG explained: how it differs from naive and advanced RAG, the key patterns like corrective RAG and self-RAG, the... Read guide → Agents & RAG AI Agent for Fintech: Risk, Compliance, Ops, Customer AI agents in finance: fraud, AML, KYC and servicing use cases, how to build with money-movement guardrails and human appr... Read guide → Agents & RAG AI Agent for Healthcare: Use Cases, Governance & Implementation AI agents in healthcare: the use cases that pay off first, how to build one HIPAA-safe on FHIR with clinician review, and... Read guide →