Your cost per acquisition is high. Before you cut budget or blame the algorithm, here are the four areas that actually move the needle: search term waste, bidding strategy, audience signals and landing page conversion rate.
By Jack Goldsmith, Founder & Performance Marketer, Social Surge · 18 August 2026
How do you reduce CPA on Google Ads for an e-commerce store?
The fastest lever is usually bidding strategy: switch from Maximise Clicks to a conversion-based strategy (Target ROAS or Maximise Conversion Value) once you have 30-50 conversions per month. Beyond bidding, the three areas that consistently cut CPA are negative keyword hygiene, audience signal refinement on Performance Max, and improving your post-click conversion rate. Most high-CPA problems trace back to one of these three causes.
Reducing CPA on Google Ads for e-commerce comes down to three variables: where your budget is being consumed (wasted clicks on low-intent searches), how your bidding algorithm is optimising, and what happens after the click. Fix the right lever first and CPA can fall 20-40% without cutting total spend.
The most common CPA problem in the accounts we manage is not a bidding issue. It is budget leaking to the wrong searches. Before touching any settings, pull your search term report for the last 60-90 days and segment by conversion rate. You will almost certainly find two things: a handful of high-intent search terms driving most of your conversions, and a long tail of loosely matched, zero-conversion terms eating 30-50% of your budget. The fix for the second group is a robust negative keyword list, not a budget cut.
Work through your search term report in batches of 100-200 rows. Flag any term where there are no conversions after spending 2-3x your target CPA, or where the intent is clearly wrong: research queries, brand searches for other companies, informational terms where you are selling a product rather than explaining one. Add brand-specific negatives at campaign level and generic intent-mismatch terms (free, cheap, diy, second-hand, review, forum) at account level. In the accounts we manage, a first negative keyword pass typically recovers 15-25% of wasted spend within 30 days without any drop in conversion volume.
If you are running Maximise Clicks or Manual CPC and wondering why CPA is high, the answer is straightforward: neither strategy is trying to minimise your CPA. They optimise for traffic volume, not purchase probability. Switching to Target ROAS or Maximise Conversion Value tells the algorithm to prioritise quality over quantity, but it requires a sufficient data threshold before it works reliably.
Google's Smart Bidding documentation recommends at least 30-50 conversions in the past 30 days before switching to a conversion-based strategy. Below that threshold the algorithm has too little signal and tends to over-bid or under-bid erratically. If you are under the threshold, focus on conversion volume first: use Maximise Conversions with a sensible budget cap, build the data set to 30-50 conversions per month, then migrate to tROAS once you have cleared it.
When you do set a tROAS target, base it on your actual margin rather than a benchmark or a round number. A tROAS of 600% sounds strong but may throttle delivery to only the safest auctions and leave profitable volume on the table. In our experience, starting 20-30% above your historical ROAS and tightening gradually as volume grows is more reliable than jumping straight to an aspirational target.
If you are running Performance Max campaigns (which covers most e-commerce advertisers today), your CPA levers are different from standard Search. PMax does not give you keyword-level control. Instead, you shape the algorithm through audience signals and asset quality.
Audience signals are not targeting in the traditional sense. They tell the algorithm where to start learning, not where to cap delivery. The signals that consistently lower CPA in the PMax accounts we manage are: your own customer lists (particularly high-LTV purchasers), website visitors who completed a purchase, and in-market audiences relevant to your product category. Layering these signals on top of your product feed gives the algorithm a faster route to profitable customers.
The single most common CPA problem in PMax accounts is vague or absent audience signals combined with an over-aggressive tROAS target set before the learning phase is complete. We covered the full breakdown in our Performance Max troubleshooting guide.
The formula for CPA is straightforward: cost per click divided by conversion rate. If your conversion rate is 0.8% and your average CPC is £0.80, your CPA is £100. Double the conversion rate to 1.6% and CPA falls to £50 without touching the ads at all. Most e-commerce store owners optimise their campaigns relentlessly and ignore the landing page entirely.
Before making any campaign changes, compare conversion rate across your key product and category pages using GA4 or your analytics platform. Pages below 1% conversion rate on paid traffic are typically the highest-leverage place to focus your attention. Common culprits: slow mobile load times, a price mismatch between the ad copy and the landing page, no clear delivery or returns information above the fold, and checkout flows with unnecessary friction (forced account creation is a reliable conversion killer).
In the accounts we manage, landing page improvements targeting these specific friction points have reduced CPA by 20-35% with no change to bids or budgets. It is not glamorous work, but the arithmetic is undeniable. A free PPC audit from us always includes a post-click review alongside the campaign analysis.
For Shopping and Performance Max campaigns, your product feed functions as your targeting layer. Google uses your titles, descriptions, product type taxonomy and custom labels to decide which searches trigger your ads. A feed with vague titles such as "Men's Jacket, Blue" matches broadly and poorly. A feed with specific titles such as "Men's Waterproof Cycling Jacket, Windproof, Navy, Size XXL" matches higher-intent searches and typically converts at a lower CPA.
The feed improvements that move CPA most reliably are: adding your primary search term to the product title within the first 70 characters; using Google's product category taxonomy correctly (it directly affects auction eligibility); adding GTINs where available (improves eligibility and quality signals); and using custom labels to separate high-margin from low-margin products so you can apply differentiated tROAS targets by profitability tier. If your Merchant Center has active product disapprovals, these reduce your eligible inventory and push more spend toward fewer products, which distorts your CPA data. We covered feed diagnostics and Merchant Center fixes in depth in our Merchant Center troubleshooting guide.
Sometimes CPA is high because the product, price point or competitive context makes paid acquisition genuinely expensive relative to your margin. No bid adjustment fixes a product that nobody searches for at your price. The diagnostic check: compare your organic conversion rate for the same products. If organic is also low, the problem is upstream of advertising and no campaign optimisation will solve it sustainably.
The other non-platform explanation is attribution. Stretched conversion windows, particularly 90-day click-through windows on products with a 2-3 day purchase cycle, inflate apparent ROAS and deflate apparent CPA by claiming credit for conversions that email, organic or direct traffic actually drove. Check your conversion window settings in Google Ads and cross-reference with GA4 assisted conversion reports. We covered this in detail in our conversion windows guide.
If you have worked through all four of the areas above and CPA is still not moving, a structured account review usually surfaces the issue quickly. Our Google Ads management service includes a full wasted-spend and bidding audit as part of onboarding, and the free PPC audit covers the same ground with no obligation.
There is no universal benchmark because CPA depends on your product margin, average order value and repeat purchase rate. A useful starting point: your CPA should sit below your gross profit per order. For a product sold at £120 with 40% margin (£48 gross profit), a sustainable CPA target is typically £20-£35. In the accounts we manage, CPA ranges from under £10 for low-AOV consumables to £80-£120 for premium cycling and outdoor gear.
For most e-commerce stores, Target ROAS (or Maximise Conversion Value with a tROAS constraint) outperforms Target CPA because it captures order value variation. If you sell products at different price points, tROAS rewards the algorithm for finding higher-value orders; Target CPA treats a £30 sale the same as a £300 one and gives the algorithm no reason to prioritise either.
Quick wins such as negative keyword work and tROAS target adjustment can show results within 2-4 weeks. Structural changes including bidding strategy migration, campaign restructure and feed overhauls typically take 6-10 weeks to stabilise because Smart Bidding needs time to exit the learning phase after major changes. Avoid making multiple large changes simultaneously, as each one resets the learning phase and prolongs the process.
Yes, and it is usually the better approach. Cutting spend often reduces conversion volume, which starves the Smart Bidding algorithm of signal and pushes CPA back up. The more reliable route is improving post-click conversion rate and eliminating wasted spend through negative keyword hygiene, while maintaining or growing total conversion volume so the algorithm keeps learning.
Book a free PPC audit: full account review, wasted-spend analysis and a clear growth plan, delivered in plain English.
Book Your Free Audit