E-Commerce Paid Acquisition · Shopping, Meta & Channel Economics

The economics of paid e-commerce acquisition are determined by three variables: traffic cost (what you pay per click or per thousand impressions), conversion rate (what percentage of paid traffic purchases), and average order value (what they spend when they do). Together these determine your effective CPA — what you pay per acquired customer — which must be measured against your gross margin to determine whether the channel is profitable.

The most important calculation: the break-even ROAS. If your gross margin is 40%, you need at least £2.50 in revenue for every £1 of ad spend to break even on the first purchase (1 / 0.4 = 2.5× ROAS). At a 3× ROAS you have 13p of gross profit per £1 of ad spend after platform fees and COGS. This minimum ROAS target must be established before any campaign goes live — it is the floor below which the channel is destroying gross margin, not creating it.

Google Shopping (now called Performance Max with a Shopping component, or Standard Shopping) shows product listings — with image, title, price, and retailer — directly in search results for product queries. Shopping ads are the dominant format for product-specific queries because they display the product visually with price before the user clicks, enabling higher commercial intent traffic than text ads.

Shopping campaigns are product feed-based — unlike text ad campaigns where the advertiser writes ad copy, Shopping ads are generated automatically from the product data feed submitted to Google Merchant Center. This means feed quality is the primary lever for Shopping ad performance: clear, accurate, keyword-rich product titles; correct product categorisation (Google Product Taxonomy); complete attribute data (size, colour, material, GTIN); and accurate pricing and availability.

Campaign structure for Google Shopping: The most common effective structure is segmenting campaigns by product category and by performance tier — separating bestsellers (which justify higher bids) from long-tail products (which need lower bids to be profitable). Single Product Ad Groups (SPAGs) — one product per ad group — enable the most granular bid control but require significant management overhead at large catalogue scale. Priority settings in Shopping campaigns allow budget and bid hierarchy to be managed across multiple campaigns targeting the same products.

Negative keyword management: Shopping campaigns require systematic negative keyword lists to prevent ads from showing for irrelevant queries. Unlike search campaigns, Shopping campaigns do not have keyword targeting — they match based on the product feed, so negative keywords are the primary tool for excluding irrelevant traffic (competitor brand searches, informational queries, out-of-stock products).

Performance Max (PMax) is Google's automated, cross-channel campaign type that replaced Smart Shopping campaigns in 2022. It uses machine learning to serve ads across Google Search, Shopping, Display, YouTube, Gmail, and Discover — using a product feed, creative assets, and audience signals to automatically determine where to serve which ad format to whom.

PMax provides less transparency and control than Standard Shopping — the advertiser cannot see individual search term reports, product-level performance breakdowns are limited, and the algorithm determines channel allocation automatically. Documented advertiser feedback has consistently noted that PMax campaigns tend to prioritise branded search terms (where conversion is easy because the user already wants to buy) over prospecting, which can inflate ROAS metrics while limiting true new customer acquisition.

Best practice for PMax in e-commerce: use audience signals (customer match lists, website visitor lists) to guide the algorithm toward relevant prospects rather than relying entirely on Google's own signals; create asset groups segmented by product category so creative and product feed data are tightly matched; and supplement PMax with Standard Shopping campaigns for top-performing products where you want more control over bid strategy and placement.

Meta's Dynamic Product Ads (DPA) automatically generate ads from a product catalogue, showing users the specific products they browsed on the advertiser's website (retargeting) or products algorithmically selected as relevant to their interests (prospecting). DPAs are the most automated form of Meta e-commerce advertising — the advertiser provides the product feed and creative templates, and Meta handles product selection, audience matching, and delivery optimisation.

The DPA product catalogue requires: product ID, title, description, image URL, landing page URL, price, availability, brand, and product category. Higher-quality feed data (better titles, higher-quality images, correct category assignment) directly improves ad relevance and conversion rates — the same feed quality principles that apply to Google Shopping apply to Meta DPA.

Audience types for Meta DPA: (1) retargeting — product viewers, add-to-cart abandoners, and purchase abandoners from your site (highest intent, highest ROAS); (2) upsell/cross-sell — existing customers shown complementary products; (3) broad audience prospecting — Meta's algorithm selects relevant products and audiences based on catalogue and pixel data. The retargeting bucket typically delivers the highest reported ROAS; broad prospecting is needed for acquisition at scale but at lower efficiency.

E-commerce retargeting reaches users who visited the site, viewed products, or added items to cart but did not purchase. It is consistently the highest-ROAS programmatic/social campaign type because it targets people who already demonstrated purchase intent — they know the brand, they saw the product, they did not complete the purchase for whatever reason.

Retargeting audience segments by priority:

1. Cart abandoners (highest priority, highest bid): Added items to cart within the last 3–7 days. The gap between their intent and their conversion is the smallest — the most common reasons are distraction, price uncertainty, or checkout friction. Dynamic ads showing the specific products in their cart with a free shipping or discount incentive are the highest-converting retargeting format.
2. Product page viewers (high priority): Viewed a specific product page but did not add to cart. Dynamic ads showing the viewed product are appropriate; the ad should address potential barriers (highlight reviews, returns policy, delivery time).
3. Category page viewers (medium priority): Browsed a category but did not view specific products. Show bestselling products from the category they browsed.
4. Site visitors (lower priority, frequency cap carefully): Visited the site but did not engage deeply. Broad retargeting with frequency caps (no more than 7 impressions per week) to maintain awareness without becoming intrusive.

ROAS targets must be set at a level that ensures campaigns are profitable after accounting for COGS, not just revenue. The formula: minimum target ROAS = 1 / gross margin percentage. At 40% gross margin, the minimum break-even ROAS is 2.5×. At 60% gross margin, it is 1.67×. At 25% gross margin (common in competitive retail categories), it is 4×.

Product-level ROAS targets are more accurate than blended campaign ROAS targets. A campaign selling both a £10 impulse item (20% margin) and a £200 high-margin product (60% margin) needs a different target ROAS for each product to be profitable. Applying a single 3× ROAS target to both means the low-margin product is unprofitable and the high-margin product is being under-invested in.

The LTV adjustment: if customers acquired through a paid channel have strong repeat purchase behaviour, the first-purchase ROAS target can be relaxed — because the second and third purchases recover the economics. The LTV-adjusted ROAS target: break-even spend as a fraction of total expected customer gross profit, not just first purchase gross profit. This is the basis for brands with high LTV (Gymshark, ASOS) running acquisition campaigns at ROAS levels that appear loss-making on first purchase but are profitable on a 12-month LTV basis.

The product feed is the data file that drives Google Shopping, Meta DPA, Pinterest Shopping, and programmatic retargeting. Feed quality directly determines ad relevance, click-through rate, and conversion rate — it is the foundation of e-commerce paid advertising performance.

Product title optimisation: Google Shopping matches products to search queries based largely on product titles. Include: brand, product name, key product attribute (colour, size, material), product type, and key feature in that order. "Nike Air Zoom Pegasus 40 Women's Running Shoe Black UK7" is more likely to match relevant queries than "Nike Pegasus Women's."

Product categorisation: Use the most specific relevant Google Product Taxonomy category available. Correct categorisation improves ad relevance and can affect cost — some categories have different benchmark CPCs.

GTIN (Global Trade Item Number): Including the correct GTIN (EAN, UPC, or ISBN) for branded products significantly improves Shopping ad performance — Google uses GTINs to match products to shopping queries and to aggregate product reviews. Missing or incorrect GTINs result in lower ad impression share.

Feed hygiene: Out-of-stock products in the feed lead to ad traffic going to unavailable products. Automated feed updates (synced with inventory in real time or at minimum daily) prevent wasted ad spend on items that cannot be purchased.

Allocating paid acquisition budget across channels requires understanding the contribution each channel makes to acquisition at each stage of the funnel. A practical framework:

• High-intent capture (35–50% of paid budget): Google Shopping and branded paid search — capturing users who are actively searching for the product or brand. Highest conversion rate, lowest funnel. Prioritise this before any other paid channel.
• Retargeting (15–25%): Meta DPA and display retargeting for cart abandoners and product viewers. High ROAS because the audience is warm. Essential for recovering the purchase intent generated by organic and upper-funnel paid activity.
• Prospecting — paid social (25–40%): Meta broad audience campaigns and TikTok for new customer acquisition. Lower ROAS than retargeting but necessary for new customer growth. Invest proportionally to the brand's growth ambition.
• Upper funnel — awareness (0–15%): Video, display, or influencer for brands at the scale where brand awareness is a meaningful growth lever. Not appropriate for early-stage brands where every pound should go toward measurable acquisition.

Each paid channel reports its own performance in its own attribution model — and they all count the same conversion differently. A customer who clicked a Google Shopping ad, then saw a Meta retargeting ad, then clicked through to purchase will appear as a Google Shopping conversion in Google Ads, a Meta conversion in Meta Ads Manager, and a direct or organic conversion in GA4 (if the final session was direct). The sum of all three platform-reported conversions significantly exceeds the actual number of conversions.

The practical approach: use GA4 with data-driven attribution as the single source of truth for channel-level performance. Accept that GA4 will report lower conversions per channel than each platform's self-reported numbers. Use platform-reported numbers for campaign-level optimisation signals within each platform (Meta's algorithm needs conversion signals to optimise delivery); use GA4 for cross-channel budget allocation decisions.

The central challenge in scaling paid acquisition is that incremental spend in a channel typically faces diminishing returns — the first £10,000/month in Meta reaches the most efficient audience; the next £10,000 reaches a progressively less efficient audience. This is the ROAS curve: ROAS declines as spend scales within a channel because the algorithm must reach progressively less-relevant audiences to spend the budget.

Strategies for scaling without ROAS collapse: expanding to new channels rather than scaling within a single channel; expanding to new geographies (domestic campaign efficiency metrics often do not transfer directly to new markets, but geographic expansion provides new audiences); improving conversion rate (which increases the revenue generated per visit, raising the ROAS at any given spend level); and improving LTV through retention (which increases the threshold ROAS at which new customer acquisition is profitable).

Sources & Further Reading