A mature e-commerce merchant allocates 55 to 65% of their Google Ads budget to Shopping, 25-30% to brand+generic Search, 10-15% to complementary PMax — this split isn't universal but remains the median observed on top quartile ROAS accounts.
"Should I put everything into Shopping or keep some Search?" It's a question 4 out of 5 e-commerce merchants ask their agency on the first call. The answer depends on concrete factors — account maturity, catalog structure, customer search behavior, inter-channel cannibalization — and rarely on a dogmatic preference for one or the other. On our 2025 sector benchmark, a starting US e-commerce merchant captures on average 66 to 78% of their clicks via Shopping, 14 to 22% via generic Search, and 8 to 12% via brand Search (medians, by vertical). This split isn't universal, and above all, it isn't optimal in every context.
This guide breaks the Shopping vs Search arbitrage into 8 actionable sections: fundamental intent differences, relative strengths of each format, budget split by maturity tier, inter-channel cannibalization, landing page impact, when to run both, and the most costly allocation mistakes. Numbers come from our aggregated data over 90 days, across all e-commerce sectors.
How do Shopping and Search capture the same intent differently?
Let's start by clearing up a common misunderstanding: Shopping and Search don't capture two different intents. They often capture the same purchase intent, but at different moments in the journey and with different signals. A user searching "running shoe" has the same overall intent as another typing "Asics Gel Nimbus 25 blue size 9" — buy athletic shoes. The difference comes down to the maturity of their search and the mode of representation that helps them decide.
Search is a textual response to a written query. The user types a phrase, Google displays 4 to 10 text ads with headline + description + URL. The arbitrage happens on semantic relevance between query and ad. The user essentially clicks on text — the product page comes after.
Shopping is a visual response. Google displays a grid of product cards with photo + short title + price + brand + merchant. The user immediately sees what they're about to buy before clicking. Click arbitrage happens on visual appeal, displayed price, recognized brand, aggregated reviews. The click is worth much more in pre-click information.
This fundamental difference — text vs image — creates two very different optimization worlds. Search optimizes semantics (keywords, match types, negatives, Quality Score). Shopping optimizes the feed (titles, images, GTIN, price, stock). An account that performs in Search isn't automatically a Shopping performer, and vice versa — the levers are largely disjoint. To go deeper on Shopping setup, our complete Shopping guide covers the entire feed + Merchant Center side.
What does Shopping do better than Search?
Shopping has four structural advantages that Search can't match in product e-commerce. Understanding them allows correct budget arbitrage instead of allocating by habit.
1. Intent qualified by the visual. The user sees price + product + brand before the click. This pre-filtering eliminates non-intentional exploration clicks. Someone clicking on a $139 product card isn't surprised by the price — they validated it pre-click. Observed result in most cases: Shopping average cart +4 to +12% vs generic Search, and systematically higher post-click CVR.
2. Automatic scaling via the feed. One feed = 1,000, 10,000, 100,000 indexed products without manual intervention. Each SKU becomes a potential ad. In Search, each keyword must be defined, bid, monitored — costly linear scaling. In Shopping, adding a product to the feed is enough to make it servable. This asymmetry explains why Shopping dominates large catalogs.
3. Eligibility for PMax and Free Listings. The Shopping feed also powers Performance Max (dominant component), Free Listings (organic Shopping results, free), YouTube Shopping, Maps products, Google Images. One dollar invested in Shopping feed quality pays off on five different inventories. Search doesn't have this multi-surface leverage effect.
4. Lower but more qualified CTR. On the surface, Shopping shows a median CTR of 1.0 to 1.6% versus 2.7 to 3.5% for generic Search depending on vertical. But Shopping CTR is already filtered by the visual (price, product), so the click that arrives is more qualified. Higher post-click CVR, higher average cart — net win for product e-commerce.
The 2026 Google Ads e-commerce guide details the complete Shopping + PMax + Search + YouTube ecosystem for a mature e-commerce account — recommended complementary reading before finalizing a budget arbitrage.
What does Search do better than Shopping?
Don't get us wrong: Search retains structural advantages that Shopping can't replicate, even in 2026. Abandoning them under the pretext that "Shopping converts better" is a mistake we see in about 1 audited account in 4.
1. Granular query control. Search offers match types (broad, phrase, exact), negative keywords at every level (campaign, group, shared list), dayparting, precise geo-bidding, device modifiers. None of this exists at the same level in Shopping. If your business requires precision — excluding certain days, regions, queries — Search remains essential.
2. Better performance in descriptive long-tail. A query like "women's hiking shoe waterproof size 8 red" is hard to address in Shopping — too precise for a feed to contain all attributes in a balanced way. In Search, a dedicated ad group with phrase match keywords perfectly absorbs this type of query with a high CTR.
3. Brand protection and competitor terms. Search lets you bid on your own brand (defense, median ROAS between 6 and 12 depending on vertical, well above Shopping) and on competitor brands (attack). Neither game is possible in Shopping, which only responds on product catalog. An e-commerce merchant who doesn't defend their brand in Search lets competitors buy back their customers.
4. Only option for services and non-products. Insurance, banking, SaaS, coaching, law firms: no physical product to index, Shopping unusable. Search (plus Display and YouTube) remains the only relevant Google Ads channel. This section is an important reminder: the Shopping vs Search arbitrage only applies to product e-commerce. Official documentation on Google Ads Shopping and Google Ads Support.
What's the typical budget split by account maturity?
There's no universal split — each market, catalog, seasonality imposes its adjustments. But in our sector panel, a pattern returns with remarkable regularity by account maturity (measured in monthly spend). This pattern serves as a calibratable starting point, not a fixed truth.
Beginner tier (< $5k/month). Priority is accumulating conversion signals so Smart Bidding Strategies have something to learn from. Shopping (or PMax Shopping inventory) carries 80% of the budget — fast scaling via the feed, higher CVR in e-commerce, fewer manual micro-decisions. The remaining 20% goes to a simple Search campaign: brand + a few highly qualified generic terms (product name + key attribute).
Intermediate tier ($5k — $30k/month). The account has accumulated enough data to segment. Shopping drops to 65%, generic Search rises to 25% (you can now open profitable descriptive long-tail groups), brand Search comes out as a dedicated campaign (10%) to protect the brand and cleanly measure isolated brand ROAS.
Advanced tier (> $30k/month). The split diversifies: Shopping/PMax 50%, segmented generic Search 30% (multiple campaigns by thematic cluster), brand Search 10%, YouTube and Display expansions 10%. This tier introduces audience signals, dynamic remarketing, and segmentation by margin or seasonality in Shopping.
Shopping ↔ Search cannibalization: who eats whom?
Inter-channel cannibalization is the main cause of budget allocation error. If Shopping eats ~14% of clicks that would otherwise go through brand Search, increasing the brand Search budget only produces ~86% incremental conversions — the rest would have converted anyway. Without clean measurement, allocation decisions become blind.
Shopping eats generic Search. When a user types "men's running shoe" and sees both Shopping and Search ads, the visual click wins in 54 to 66% of cases depending on vertical. This is the most frequent and massive cannibalization. Performance-wise, it's often beneficial: Shopping converts better. Measurement-wise, it artificially inflates Shopping ROAS and depresses generic Search ROAS.
PMax eats brand Search without brand exclusions. Performance Max serves on all surfaces, including brand Search, and without explicit brand exclusion list setup, it can intercept your brand's name queries at reduced CPC — which gives the illusion of exceptional PMax ROAS. Result: brand Search statistically collapses because it's deprived of its most profitable queries. Activating PMax brand exclusions is mandatory to maintain measurement integrity.
Median cannibalization from Shopping to brand Search: 10 to 18% depending on vertical. The cleanest measurement method remains the 14-day holdout test: via Google Ads experiments, cut Shopping on 50% of a random audience for 14 days and measure the variation in Search conversions on the control vs test segment. The absolute difference = real cannibalization. Faster but less precise method: the Search Terms Report crossed with the Path to Conversion report to identify queries passing through both channels.
judging brand Search "unprofitable" because its volume is low — when in fact PMax absorbed 80% of its queries via the absence of brand exclusions. Before cutting a brand Search campaign due to insufficient ROAS, systematically check the brand exclusion lists of all active PMax campaigns and re-measure over 14 days.
Landing pages and UX: impact on allocation
The choice of landing page should align with the source channel. In most cases, a channel-landing mismatch costs up to -14 to -26% CVR depending on vertical — one of the most underexploited levers of Shopping vs Search allocation.
Shopping → Product Detail Page (PDP). The user has already seen the product and price in the ad. They expect to land directly on the product page, not on a category to re-explore. Sending a Shopping click to a category page = immediate frustration + high bounce + degraded CVR. 100% of performing Shopping accounts we audit send to the exact PDP corresponding to the clicked SKU.
Generic Search → Category page or thematic LP. The user on generic Search is in exploration mode. "Men's running shoe" doesn't expect a specific PDP — they want a choice of products to compare. Well-filtered category page, with facets (brand, price, size, reviews) and a few best-sellers highlighted. More powerful alternative: a dedicated thematic landing page ("The best men's running shoes 2026") with editorial tone and curated selection.
Brand Search → Homepage or brand landing. The user typed your brand — they already know you. Homepage or valorizing brand landing, with clear USP and best-sellers visible above-the-fold.
Mobile first: 75% of US e-commerce traffic in 2026. A Shopping click sent to a slow or poorly structured mobile PDP is over-penalized. Why? Because Shopping brings already decided users; if they turn back, it's due to UX, not product. Google captures this via post-click signals and lowers the feed's Quality Score. If your mobile UX is mediocre, Shopping suffers more than Search in relative terms. Mobile performance documentation on Think with Google.
When should you run both simultaneously?
Short answer: almost always in product e-commerce. The two formats are complementary, not substitutes. The rare cases where one alone is enough:
- Very small catalog (< 20 active SKUs): Shopping lacks diversity to be dominant. Search takes over with dedicated ad groups per product.
- Private brand with no name search volume: no brand Search to defend, you can focus on Shopping + generic Search.
- Non-product service (insurance, SaaS, coaching, banking): Shopping unusable, 100% Search.
- Very constrained budget (< $1k/month): concentrate on standard Shopping or PMax to avoid dilution. Add Search only when volume allows.
In all other cases, the question isn't "Shopping or Search" but "in what proportion". Hybrid is the norm, not the exception. The right pattern systematically includes cross-channel query sculpting via shared negatives, to route each query to the channel where it performs best.
Cross-channel query sculpting. The principle: negatives don't just exclude irrelevant queries, they also route a query to the right channel. Example: exclude descriptive long-tail queries from Shopping via negatives so they fall on Search (where they perform better), and exclude brand queries from generic Search via negatives so they fall on brand Search (where ROAS is 3x higher). This type of orchestration requires an architectural audit — our Smart Bidding guide and our complete 2026 Performance Max guide cover the associated mechanics in detail.
Which allocation mistakes cost 20% ROAS?
Five allocation mistakes recur in the majority of accounts we audit. Each individually costs 4 to 8% of account ROAS; cumulatively, they explain the typical 15 to 25% gap between an "average" account and a well-orchestrated account.
1. Everything in Shopping with no brand Search. Without a brand campaign, your competitors (or your own misconfigured resellers) buy back your name queries at low CPC. Measured result: median loss of 4 to 9% of total revenue depending on vertical compared to an identical account with brand Search activated. Fix: 1 dedicated brand Search campaign, minimal budget (3 to 8% of spend), exact + phrase match types, cross-channel negatives.
2. Everything in Search with no Shopping. You leave 40% of visual intent to your competitors. In product e-commerce with catalog > 20 SKUs, it's structurally suboptimal — Shopping's higher post-click CVR translates directly into lost revenue. Fix: launch standard Shopping or PMax by D+30 of the account at the latest.
3. Fixed account-level budget with no dynamic arbitrage. Shopping peaks in Q4 (Black Friday, Christmas), brand Search relatively stable year-round. A budget locked at 50/50 leaves Shopping under-fed in Q4 and Search over-fed in Q1-Q2. Fix: budget envelopes by channel reviewed every 4 weeks, with a dynamic rule "if marginal ROAS > target +20% for 2 weeks → +10% budget."
4. Identical target ROAS for Shopping and Search. Shopping converts better → tolerates a lower ROAS for the same absolute profit. Forcing tROAS 5 everywhere suboptimizes Shopping (which could scale at tROAS 4 with more volume) and Search (where tROAS 6+ is sustainable on brand). Fix: differentiated tROAS by channel and by campaign — brand 7-10, generic Search 4-5, Shopping 3-4 depending on margin, PMax 4-5.
5. No brand vs generic split in Search. Mixing in a single Search campaign makes brand ROAS measurement impossible and dilutes Quality Score. Fix: 2 distinct Search campaigns, separate budgets, separate tROAS, cross-channel negatives to avoid contamination.
If you suspect one of these mistakes in your own account, a free SteerAds audit automatically diagnoses the 5 points in under 48 hours — budget split, measured cannibalization, PMax brand exclusions, differentiated tROAS, isolated brand Search. The tool then recalculates real ROAS by channel and alerts on budget allocation drifts every week.
To go deeper, an essential external resource: Search Engine Land regularly publishes studies on Shopping vs Search allocation in the Google Ads ecosystem, with case studies by e-commerce sector.
Sources
Official sources consulted for this guide:
FAQ
Should you always launch brand Search in addition to Shopping?
Yes, in most US e-commerce configurations (around 92 to 96% depending on vertical). Without a dedicated brand Search campaign, your competitors (and even your own resellers) buy back your brand queries and intercept some of your highest-converting customers. On the SteerAds 2025-2026 sample, the median brand Search ROAS sits between 6 and 12 — well above the account average. Required budget is usually low (3 to 8% of total spend), while the opportunity cost of not doing it is very high. Only exception: a purely private brand with no measurable name search.
My Shopping is inventory-limited: should I shift to Search?
Not automatically. Before redistributing to Search, check three things: (1) inventory script — if fewer than 20 active SKUs, Shopping does lack diversity and Search takes over; (2) Merchant Center approval — 24 to 36% of accounts we audit have 10 to 40% of SKUs in 'disapproved' that skew the reading; (3) implicit exclusions (low_roas custom_label, zero stock) that mechanically reduce the eligible pool. If Shopping remains inventory-constrained after cleanup, then yes: shift 15 to 25 points of budget to descriptive long-tail generic Search.
How do you measure Shopping-to-Search cannibalization?
The most reliable method remains the 14-day holdout test: cut Shopping on 50% of your audience (via Google Ads experiments) and measure the variation in Search conversions on the same segment. In our sector panel, the median cannibalization from Shopping to brand Search sits between 10 and 18% — meaning a notable share of brand Search conversions would be captured by Shopping if Shopping were cut. Faster but less precise method: export the Search Terms Report and identify queries that passed through both channels for the same user via the Path to Conversion report.
Does Performance Max replace the Shopping plus Search split?
No, and confusing them is an architectural mistake. PMax contains Shopping inventory (it's even its dominant component in e-commerce) but doesn't exploit manual Search granularity — no query sculpting, no negatives at the same level, no fine match-type control. The right architecture in 2026 often combines three bricks: standard Shopping (Priority Low or Medium) for long-tail sculpting, PMax for discovery and remarketing, branded Search and generic Search for precise queries. The three don't replace each other, they complement each other on different intents.