Every DTC brand I audit has a Meta account they treat seriously. They have a media buyer or an agency, they run creative testing, they know their cost per purchase. Then I look at their Google account, and it is almost always the same thing: one Performance Max campaign, launched twelve months ago, never touched. Budget set to fifty pounds a day. ROAS sitting at 1.8x. Founder wondering why Google does not work for their brand.
Google Shopping is not broken. The setup is. Most DTC brands are leaving a meaningful revenue channel completely uncaptured because they treat it as a secondary channel that runs itself. It does not run itself. It needs the same structural thinking you would apply to Meta, with different levers.
This post covers the exact structure I use when rebuilding a Google Shopping account for a DTC brand - the feed, the campaign architecture, the bidding logic, and the Performance Max setup that works across drinks, beauty, and wellness at the scale most founder-led brands operate at.
Why Google Shopping matters more than most founders think
The argument for Google Shopping is simple. Meta and TikTok are interruption platforms. You are buying your way into someone's feed and hoping they are in the right mindset. Google Search and Shopping are intent platforms. A person typing "magnesium glycinate supplement UK" into Google has self-selected as someone who wants to buy that product. You are not creating demand, you are capturing it.
For DTC brands selling health, beauty, wellness, and functional food and drink products, the search volume is significant. Tens of thousands of product-specific searches happen every month across the UK and EU. Brands that are not on Google Shopping are simply ceding that intent to competitors and to Amazon.
There is also an attribution dimension. Google Shopping often closes purchases that were first touched by a Meta or TikTok ad. A customer sees your brand on Instagram on Tuesday, searches "buy [product name]" on Thursday, and converts through Google Shopping. A last-click model attributes the whole sale to Google. A first-touch model attributes it to Meta. The truth is that both contributed. If you are only looking at last-click ROAS by channel, you are likely undervaluing both simultaneously.
Run a post-purchase survey question asking how customers heard about you. You will almost always find Google is a more important consideration touchpoint than your last-click data suggests. That should inform how much budget it gets.
The product feed: the lever almost nobody optimises
Google matches your products to search queries using your Merchant Center feed. The quality of that feed determines which searches trigger your ads. Most DTC brands connect Shopify to Merchant Center via the Google channel app and leave the feed exactly as it pulls from the product catalogue. That is usually a mistake.
Product titles: your most important feed element
Google uses your product title as the primary signal for query matching. The default Shopify product name is almost never optimised for this. A product named 'Glow Serum 30ml' will match almost nothing. A product titled 'Vitamin C Brightening Serum - 30ml - Ascorbic Acid 15% - Brightens Dark Spots - Vegan' will match dozens of high-intent searches.
The formula for DTC product titles is: primary keyword + key product attribute + size or quantity + secondary benefit + format. You want the most important keyword at the start of the title because Google truncates in Shopping ads and the first 25-30 characters carry more weight. Use DataFeedWatch, Channable, or a custom supplemental feed in Merchant Center to override titles without touching your Shopify product names.
Product descriptions: used for matching, not display
Descriptions are not shown in Shopping ads, but they contribute to query matching. Use the first 500 characters to include all relevant search terms your customers would use: ingredient names, benefits, use cases, certifications (vegan, cruelty-free, organic), and product format. Write naturally, not as a keyword dump, but be comprehensive.
A good description for a protein powder would include: the protein source (whey, pea, blend), the flavour, the primary benefit (muscle recovery, meal replacement), dietary certifications, and key differentiators like hormone-free or grass-fed. Each of these surfaces your product to a different high-intent search.
Product images: the conversion lever in the feed
Your main product image is what appears in Shopping ads. Google has strict requirements (white or light background, no text, no watermarks for Shopping listings), but within those rules there is significant room for differentiation. Images that clearly show the product size, the form factor, and the label drive higher click-through rates than generic pack shots.
For beverages, show the product poured into a glass alongside the bottle. For supplements, show the capsules or powder alongside the pack. For skincare, show texture or application alongside the bottle. Additional images in the feed (image_link, additional_image_link attributes) can show lifestyle context. These appear in Shopping ads in some formats and improve performance.
Custom labels: how you control bidding by product
Custom labels are attributes you add to your feed that have no meaning to Google but allow you to segment your products in campaign structure. Use them to group products by margin tier, hero vs. secondary SKU, seasonal relevance, or new vs. established product. Label your highest-margin hero SKUs as 'tier-1'. Label seasonal products as 'bfcm-priority'. Then in Google Ads, you can create separate campaigns or asset groups for each tier and bid differently.
This is the mechanism that lets you put more budget behind your best-margin products and less behind slow movers. Without custom labels, you are bidding the same way on every product in your catalogue and optimising toward average, not maximum, profitability.
Campaign architecture: Standard Shopping plus Performance Max
The single-PMax-campaign approach is the default for many brands because it is what Google recommends and what most agencies set up by default. It is rarely the best structure for a DTC brand with a small-to-mid product catalogue.
The structure I use looks like this. Two campaigns: a branded Standard Shopping campaign that captures searches containing your brand name, and a non-branded Performance Max campaign that covers everything else. Here is why each part matters.
Campaign
Purpose
Bid strategy
Branded Search
Capture your own brand name searches. Protect branded intent from competitors.
Target CPA or Max Conversions with tCPA
Non-brand PMax
All non-branded intent. Shopping, Search, YouTube, Display coverage across your catalogue.
Target ROAS starting at 3x, adjusted by category
Hero SKU Standard Shopping (optional)
Granular bidding control for your 2–3 highest-margin products at high purchase intent.
Manual CPC or Enhanced CPC for data gathering
The branded campaign runs separately because PMax will happily spend budget on branded searches (someone typing your name) and claim the ROAS credit, which inflates PMax performance metrics while not actually driving incremental purchases. Exclude your brand name as a negative keyword at the campaign level for your PMax campaign.
The hero SKU Standard Shopping campaign is optional but valuable for brands with two or three products that drive the majority of revenue. Standard Shopping gives you impression share data, which tells you what percentage of relevant searches you are actually showing up for. If your hero SKU has 40% impression share, you know there is significant volume you are not capturing. PMax does not surface this granularly.
Performance Max asset groups: where most brands leave money on the table
Performance Max asset groups are where you provide Google with the creative raw material it uses across Shopping, Display, YouTube, Gmail, and Search Discover. Most brands fill this out once with generic assets and never revisit it. The algorithm serves whatever it has. If what it has is weak, the campaign underperforms.
Create at least two asset groups: one for each distinct product category or use case in your catalogue. A beauty brand with a skincare range and a haircare range should have separate asset groups with relevant images, copy, and audience signals for each. The relevance signal this sends to Google meaningfully improves placement quality.
For each asset group, upload at least 15 product and lifestyle images, 5 square images, 5 portrait images, at least 3 video assets (YouTube videos, even simple ones), 5 headlines (15 characters for short, 30 for long), and 5 descriptions. Google will test combinations and tell you which perform. The more variants you give it, the more data it can generate.
Audience signals are the other critical piece. Connect your Klaviyo or Shopify customer list to Google via Customer Match. Upload your email subscribers as a signal. Add your Shopify store visitors segmented by purchase intent depth. Add in-market audiences for your category. These signals tell Google where to look first when optimising. Without them, it starts cold.
Bidding: why starting too high kills your account
Target ROAS bidding in Google tells the algorithm: for every pound I spend, generate this many pounds of revenue. Set the target too high and the algorithm restricts spend so aggressively it cannot find enough auctions to enter - your budget goes unspent and you generate no data. Set it too low and you spend money on conversions that do not cover your margin. The right starting point depends on your contribution margin.
Contribution margin
Break-even ROAS
Recommended starting tROAS
Note
35%
2.9x
3.5–4x
Tight margin - limit to high-intent Shopping only
45%
2.2x
3–3.5x
Workable for most DTC brands at this range
55%
1.8x
2.5–3x
More room to scale volume; test lower tROAS
65%
1.5x
2.5x
Can be aggressive on volume - monitor blended margin
The algorithm needs approximately 30–50 conversions per month to optimise reliably. If you are a smaller brand with 15–20 purchases from Google per month, the bidding algorithm does not have enough signal. In this case, run the account on Maximise Conversions for the first 4–6 weeks to build data volume, then introduce a tROAS target once you have consistent weekly conversion data.
Do not change your tROAS more than once per week. Every change restarts the learning period. Google's smart bidding learns from recent auction data - too many adjustments and the account lives in perpetual learning mode, which means erratic spend patterns and unstable results. Make a change, wait two weeks, evaluate, adjust if needed.
The five mistakes that kill Google Shopping performance for DTC brands
No brand exclusion in PMax
Running Performance Max without excluding your brand name means Google serves your ads on branded searches (people who already know and want your brand) and takes the ROAS credit. This inflates your PMax numbers and cannibalises traffic you would have captured cheaply through organic search or a branded search campaign. Add brand name negative keywords at the campaign level as a standard step.
Feed pulled directly from Shopify without optimisation
The default Shopify-to-Merchant-Center integration pulls product titles, descriptions, and images as they exist in your store. Store product names are rarely optimised for search intent. If your product is named 'The Daily Glow' in Shopify, Google will match it to almost nothing relevant. Use a supplemental feed or a tool like DataFeedWatch to override titles with keyword-rich versions without changing your storefront.
Setting tROAS too high too early
Starting a new campaign at 5x or 6x tROAS is almost always a mistake. Google cannot find enough eligible auctions to spend the budget, the campaign starves for data, and the founder concludes Google does not work. Start at 3x, let the algorithm build conversion history over 4–6 weeks, then optimise. The platform rewards patience with progressively better signals.
No conversion tracking audit before spending
Google's Smart Bidding optimises toward whatever conversion action you have tagged. If your Shopify purchase event fires on any order confirmation page (including failed payment attempts), double-counts across GA4 and the Google tag, or is set to 'every conversion' rather than 'one per click', you are feeding corrupted data to the bidding algorithm. Audit your conversion setup before spending serious budget. Every misfire sends a false positive that degrades bidding quality.
Evaluating Google in isolation by last-click ROAS
Google Shopping often closes purchases that were initiated on Meta, TikTok, or email. If your attribution is last-click, Google gets all the credit. But if you then reduce Google budget because 'Meta is performing better this week', you will likely see conversion rates drop across all channels because you have removed the close-intent channel from the path. Use a post-purchase survey and data-driven attribution in GA4 to understand the actual contribution each channel makes.
What good Google Shopping performance looks like at DTC scale
For context on realistic expectations: a well-structured Google Shopping account for a DTC brand spending £2,000–£8,000 per month should deliver a blended ROAS (Shopping and PMax combined, excluding branded) of 3–5x within three months of proper setup. Brands with strong organic search presence, high-quality feeds, and good audience signal data can achieve 5–8x on Shopping-specific placements because the intent quality is higher.
The trajectory matters. Month one is typically lower as the algorithm learns: expect 1.5–2.5x while the bidding stabilises and Google builds conversion history. Month two improves as Smart Bidding gets more signal. By month three, a well-fed account with proper structure and audience signals should be performing predictably at your target ROAS or above.
The brands that fail are the ones who evaluate performance at week two, conclude it is not working, make five simultaneous changes, restart the learning period, evaluate again at week four, change the structure again, and repeat the cycle indefinitely. Google's algorithm is genuinely good when you let it learn. Most founders do not have the patience for it because Meta gives feedback faster. Google's feedback loop is slower, but the intent quality it captures is meaningfully higher.
Get your Google Shopping account audited
The free scorecard takes three minutes and covers paid media channel efficiency alongside email, conversion rate, and unit economics. It will show you where your biggest constraint is right now and whether Google Shopping is the growth lever you should be prioritising.
If you want a full review of your Google Merchant Center feed, campaign architecture, and bidding strategy, the Brand Growth Audit covers your complete paid media stack with a prioritised action plan. Three days, Loom walkthrough, written report.
Frequently asked questions
Should DTC brands use Performance Max or Standard Shopping?
Most DTC brands should run both, not one or the other. Standard Shopping gives you granular control over product-level bidding and clean impression share data. Performance Max adds coverage across YouTube, Display, Gmail, and Search Discover, and it layers in audience signals to find higher-intent buyers. The right structure is a brand campaign in Standard Shopping plus a non-brand Performance Max campaign targeting all other intent signals.
What is a good ROAS target for Google Shopping for a DTC brand?
The right target ROAS depends on your contribution margin, not an industry average. If your contribution margin is 45%, your break-even ROAS on media spend alone is 2.2x. Most DTC brands should target 3–4x tROAS on Google Shopping to leave room for fixed costs and generate meaningful profit. Setting tROAS too high (6x+) typically restricts spend so tightly that the algorithm cannot enter enough auctions.
How important is the Google Merchant Center product feed for Shopping performance?
The product feed is the single biggest lever in Google Shopping performance. Google matches search queries to your products through your feed titles, descriptions, and attributes. Brands that invest in feed optimisation typically see 30–50% improvement in click-through rate and conversion rate without changing a single bid.
What audience signals should DTC brands use in Performance Max?
The most effective audience signals for DTC brands in PMax are: customer match lists from Klaviyo or Shopify, website visitors segmented by purchase intent depth, similar audiences built from your purchaser list, and in-market audiences matching your category. Feed these as signals, not as targeting restrictions. PMax uses them to bias where it looks, not to limit who it can reach.
How do you prevent Performance Max from cannibalising your brand search campaigns?
Add brand terms as negative keywords at the campaign level for your PMax campaign. Google allows brand exclusions in Performance Max through campaign settings. Without this, PMax will serve on branded queries and claim the conversion credit, which inflates your PMax ROAS while cannibalising traffic that would have converted organically or through a lower-cost branded search campaign.
Why do most DTC brands see poor results from Google Shopping early on?
The three most common reasons are: launching with too high a tROAS target before Google has enough conversion data, poor product feed quality meaning Google cannot match ads to relevant searches, and running PMax with no audience signals or asset groups. Most brands also make the mistake of evaluating Google Shopping in isolation by last-click attribution, which understates its true contribution across the customer journey.
About the author
Caner Veli founded and exited Liquiproof, scaling from zero to 3,000+ retailers globally in under 6 years. He now runs Purposeful Profits, a focused growth consultancy for founder-led DTC and CPG brands. 12 named sprint clients. 518% average growth. 27x highest ROAS. Read more about Caner →