Paid Social for E-Commerce: What Works, What Wastes

The short answer

The short answer Paid social for e-commerce is a creative and catalogue business. Facebook traffic campaigns average $0.70 per click at a 1.71% click-through rate, so the cost is in the conversion rate, not the click. Win it with catalogue-driven campaigns, high creative volume, margin-based targets and holdout testing rather than last-click reports. The numbers that frame the.

The short answer

Paid social for e-commerce is a creative and catalogue business. Facebook traffic campaigns average $0.70 per click at a 1.71% click-through rate, so the cost is in the conversion rate, not the click. Win it with catalogue-driven campaigns, high creative volume, margin-based targets and holdout testing rather than last-click reports.

The numbers that frame the decision

The LocaliQ Facebook advertising benchmarks put all-industry traffic campaigns at $0.70 average cost per click and a 1.71% click-through rate, with lead generation campaigns at $1.92 per click and a 7.72% conversion rate. Clicks are cheap. That is precisely why undisciplined accounts lose money here: the platform will happily sell you volume at $0.70 forever.

IRP Commerce market data for July 2026 recorded an average store conversion rate of 2.22%, up 16.21% year over year, an average order value of £122.26, and mobile at 62.40% of sales. In that same last-click dataset, Facebook paid accounted for 0.34% of sales against Google paid search at 60.28%.

Treat that 0.34% carefully. It is a last-click figure, and social is systematically undercounted by last-click because it creates demand that later resolves through a branded search. Reading it as “Facebook does not work” is the wrong conclusion. Reading it as “you cannot manage social on a last-click report” is the right one.

The catalogue does the heavy lifting

Product-feed campaigns beat static creative in almost every store account, because the system can match a person to a product rather than to a message. Meta reported that Advantage+ shopping campaigns drove 12% lower cost per purchase conversion than business-as-usual ads across a study of 15 A/B tests, and that the format can automate up to 150 creative combinations at once (Meta on Advantage+ shopping campaigns).

The prerequisite is the same feed discipline that governs your Google surfaces. Accurate availability, correct pricing, clean titles and populated identifiers. If your catalogue is stale, automation amplifies the error at scale. Fix the feed once and both channels improve, alongside the free placements Google offers across Search, the Shopping tab, Images, Lens, YouTube and Maps when attributes, shipping and returns are complete (Google free product listings).

Creative volume is the real budget line

Input	Common practice	Better practice
Creative refresh	One set per quarter	New concepts every two to three weeks
Format mix	Single static image	Vertical video, user content, catalogue, static
Testing unit	Headline variants	Distinct concepts, then variants within winners
Conversion value	Order revenue	Gross profit per order
Audience	Interest stacks	Broad plus catalogue, purchasers excluded
Measurement	Platform-reported ROAS	Geographic holdout twice a year

With mobile at 62.40% of sales, creative has to work at arm’s length on a small screen with the sound off. Product visible in the first second, the offer in text on the frame, and a format that fills the vertical viewport. Anything repurposed from a desktop banner is starting from behind.

Set targets on margin

A 3:1 return on ad spend is comfortable at 65% gross margin and loss-making at 25%. Pass gross profit rather than order revenue as the conversion value so the bidding system optimises toward contribution rather than turnover. Then net returns off before judging a product group, because a category with a 30% return rate is not performing the way the dashboard says.

Compare across channels on the same basis. The LocaliQ search advertising benchmarks for the period ending 1 June 2026 put shopping, collectibles and gifts at $4.14 per click on search with a 4.01% conversion rate. Social clicks cost a fraction of that and convert at a fraction of the rate. Neither is inherently better; the allocation question is which one grows contribution at the margin, and that is answered by holdout tests, not by dashboards.

What to stop paying for

Stop stacking narrow interest audiences, which raises costs without improving quality now that the delivery system finds buyers more efficiently on broad. Stop prospecting to recent purchasers at full acquisition cost when email or SMS would have worked. Stop running one static image for months. Stop advertising out-of-stock or low-margin SKUs. Stop scaling budget on a good week, because the learning reset usually costs more than the extra spend returns.

Stop making allocation decisions on platform-reported ROAS alone. Run a geographic holdout at least twice a year, measure total store revenue rather than attributed revenue, and size the channel from that.

What to build instead

A prospecting campaign on broad targeting with the catalogue attached and purchasers excluded. A retargeting layer segmented by product-page viewers, cart abandoners and past purchasers, each with different messaging and different value targets. A creative pipeline producing several genuinely distinct concepts a month rather than variants of one. And a landing experience built for a phone, because at a 2.22% baseline conversion rate a small improvement there outperforms most bidding work. That is where conversion work compounds, alongside the feed and organic work that reduces how much of your catalogue you rent.

The retargeting layer people build backwards

Most store accounts run one retargeting campaign against a 30-day site-visitor audience and call it done. That audience blends three very different people: someone who bounced off the home page, someone who viewed a product twice, and someone who abandoned a full cart. They deserve different messages and different value targets, and merging them means the cheapest impressions go to the least interested group.

Split it into at least three segments with descending audience size and ascending bid: product-page viewers, add-to-cart abandoners, and checkout abandoners. Exclude purchasers from all three for a window matched to your genuine repurchase cycle, which for consumables might be 30 days and for furniture might be two years. Then run a separate reactivation campaign against past purchasers with a different offer entirely. That structure typically costs nothing extra and stops the account paying prospecting prices to reach people who were already going to buy.

The bottom line

Cheap clicks make social easy to overspend on. Attach the catalogue, produce real creative volume, optimise to gross profit, exclude existing customers from prospecting, and validate with holdout tests rather than attribution reports.

Tack Media is a Meta Partner, Shopify Partner and Google Partner and has run e-commerce accounts since 2009. For a creative and measurement review, book at calendly.com/tack-media-agency/talk-to-an-expert or call 310-620-1141.