Cost Per Acquisition (CPA) is the single most useful metric in Google Ads. It cuts straight through the noise of clicks, impressions and CTR and tells you the only thing that matters: how much it cost to get a customer.
And it's the metric Ben gets asked about most often - usually phrased as some variation of "our CPA is too high, can you fix it?"
The honest answer is: almost always, yes. Across 250+ Google Ads accounts Ben has personally managed, the average CPA reduction in the first 90 days of a Revenue Takeover sits at 32-48%. That's not because Ben is doing anything magic - it's because most accounts have 4 or 5 obvious leaks, all of which compound to inflate CPA.
This guide walks through the exact diagnostic framework, the 14 levers Ben actually pulls, and the realistic timeline for seeing CPA come down without tanking conversion volume.
What CPA actually means (and what counts as 'high')
CPA = Total Ad Spend ÷ Total Conversions.
So if you spend £5,000 in a month and generate 50 conversions, your CPA is £100. Simple maths - but the metric hides several traps:
- What's a "conversion"? A purchase is a conversion. So is a "Contact us" form submit. So is a newsletter signup. If you're tracking soft conversions alongside hard ones, your CPA looks artificially low and you'll optimise toward the wrong outcomes.
- What about offline conversions? If 60% of your leads close offline (phone, sales team) and you're not feeding that data back into Google, your reported CPA is meaningless.
- What's the customer worth? A £200 CPA is brilliant if your average customer LTV is £8,000. It's catastrophic if your AOV is £180.
The right question isn't "is my CPA high?" It's "is my CPA profitable, given my LTV and margin?" A useful target: keep CPA below 20-30% of customer lifetime value for a sustainable, scalable account.
The 4-question diagnostic Ben runs first
Before changing anything in an account, Ben runs through these four questions. They almost always reveal where the CPA inflation is actually coming from - and the fix follows from the diagnosis.
1. Is the CPA high because the cost is high, or because the conversions are low?
This is the fundamental split. CPA = Cost ÷ Conversions, so it can be inflated from either side:
- Cost-side problem: CPCs are too high, you're paying for irrelevant clicks, or impression share is being wasted on the wrong queries.
- Conversion-side problem: CTR is fine and CPCs are reasonable, but conversion rate is in the toilet - usually a landing page, offer, or audience-fit issue.
The fix is completely different depending on which side is broken. Roughly two-thirds of high-CPA accounts Ben audits are conversion-side problems, not cost-side. People assume "high CPA" means "I'm paying too much per click" - usually it means "the clicks aren't converting."
2. Is conversion tracking actually accurate?
An astonishing percentage of accounts - Ben estimates around 40% of accounts he audits - have broken, misconfigured, or duplicate conversion tracking. Common issues:
- Tags firing on every page load instead of the conversion event
- Both GA4 and Google Ads tracking the same event = double-counting
- Critical conversions (phone calls, offline sales) not tracked at all
- Wrong attribution model distorting credit
If tracking is broken, every other tactic in this guide is built on sand. Fix tracking first.
3. Are the smart bidding signals strong enough?
Smart bidding (Maximise Conversions, Target CPA, Target ROAS) needs at least 30 conversions per 30 days at the campaign level to work. Below that threshold, smart bidding can't optimise reliably and CPAs swing wildly.
If your campaign has 8 conversions a month and you're on Target CPA, the algorithm is essentially guessing. The fix isn't to "tighten the target" - it's to either consolidate campaigns or move back to Maximise Clicks until volume is sufficient.
4. What's the search-term-to-conversion ratio?
Pull a 90-day search terms report. How many unique search terms triggered ads? How many of them converted at least once?
If you're paying for clicks across 4,000 unique queries and only 60 of them have ever converted, you have a targeting problem - you're casting too wide a net. Tightening targeting (negatives, match types, audiences) is one of the fastest CPA wins available.
CPA benchmarks by industry (2026)
These benchmarks are aggregated from WordStream's 2025 industry report, Google auction insights, and Ben's own data across 250+ accounts. CPAs have generally risen 10-15% year-on-year as CPCs have crept up and competition for first-party data has intensified.
- Legal services: £95-£280 average CPA
- Insurance: £80-£175 average CPA
- Finance & accounting: £65-£190 average CPA
- Home services (plumbing, HVAC): £45-£120 average CPA
- Dental & medical: £55-£140 average CPA
- B2B SaaS (free trial signup): £45-£140 average CPA
- B2B SaaS (demo request): £180-£450 average CPA
- E-commerce (general): £18-£55 average CPA
- Travel & hospitality: £25-£75 average CPA
- Education / training: £55-£160 average CPA
- Real estate (buyer leads): £75-£220 average CPA
If your CPA is significantly above the top of these ranges, you almost certainly have a fixable account problem. If you're already inside the range and want to push lower, you're into the realm of compound optimisation - small wins across many levers.
14 proven tactics to reduce CPA
Here are the 14 levers Ben pulls, in roughly the order he'd pull them in a typical account audit. Most accounts won't need all 14 - hitting 5 or 6 of these properly will usually cut CPA by 25-40%.
1. Fix and modernise conversion tracking
The single highest-leverage tactic. Move to GA4 + Google Ads with proper enhanced conversions. Implement offline conversion imports if you have a sales team. Set conversion values at the action level so smart bidding can prioritise high-value leads.
Accounts that implement enhanced conversions properly typically see 5-12% lower CPAs within 30 days, simply because Google's algorithms are now optimising against accurate signal.
2. Build aggressive negative keyword lists
Pull a 90-day search terms report and add as negatives every query that:
- Generated clicks but zero conversions across 100+ impressions
- Contains job-hunting terms ("jobs", "careers", "salary")
- Contains DIY or research terms ("how to", "tutorial", "free")
- Includes competitor brand variants you don't want to bid on
- Is geographically irrelevant despite location targeting
A typical 90-day negative keyword sweep removes 15-30% of wasted spend from a poorly maintained account.
3. Tighten match types
Broad match without smart bidding and conversion data is a money pit. Move to phrase or exact match for high-CPA campaigns until you have 30+ conversions/month, then experiment with broad on top with strict negative lists.
4. Improve Quality Score
Quality Score directly affects how much you pay per click. Going from a Quality Score of 5 to 8 typically reduces CPC by 30-40%, which flows straight through to CPA. Quality Score levers:
- Ad relevance - include the exact keyword in headlines
- Expected CTR - test multiple ad variants per ad group, prune the losers
- Landing page experience - keyword-aligned pages, fast load, mobile-optimised
5. Rebuild landing pages for conversion intent
Most landing pages are built for SEO or branding - not paid conversion. A purpose-built landing page typically converts 2-5x better than a homepage or generic services page.
The basics that move the needle:
- Headline matches the ad copy and search intent exactly
- Single, obvious primary CTA above the fold
- Trust signals (reviews, logos, certifications) within the first scroll
- Form length minimised - every extra field reduces submission rates 4-7%
- Page load under 2.5 seconds on mobile
One of Ben's e-commerce clients cut CPA 41% simply by replacing a generic category page with a campaign-specific landing page that mirrored the ad copy.
6. Use Maximise Conversions or Target CPA properly
Once you have 30+ conversions/month, smart bidding outperforms manual or Maximise Clicks in roughly 78% of head-to-head tests Ben has run. The mistakes to avoid:
- Setting an aggressive Target CPA below the campaign's historical average - throttles the algorithm
- Switching strategies every two weeks - resets the learning phase
- Running smart bidding with no conversion value set - algorithm has nothing to prioritise
7. Layer in audience signals
Adding observation audiences (in-market, custom intent, customer lists, website visitors) gives smart bidding richer signal. Customer Match lists are particularly powerful: feeding Google a list of past customers lets it optimise toward look-alike audiences with much higher conversion probability.
Accounts using Customer Match see CPAs 8-15% lower on average than identical campaigns without it.
8. Dayparting and ad scheduling
Pull the Day & Hour report. If conversions cluster in business hours but spend is even across 24/7, you're paying for unproductive clicks at 2am. Schedule ads to align with when prospects actually convert - typical CPA reduction: 10-18%.
9. Geographic optimisation
Pull the Geographic report. Locations that generate clicks but no conversions should be excluded or have bid adjustments applied. Critical setting: change location targeting from "Presence or interest" to "Presence only" to avoid paying for clicks from people just researching your area.
10. Device bid adjustments
Mobile, desktop, and tablet rarely convert at the same rate. For most B2B and high-AOV e-commerce, desktop converts 1.5-3x better than mobile. Apply bid adjustments accordingly. (For local services and emergency intent searches, mobile usually wins - the inverse adjustment applies.)
11. Refine ad copy with single-purpose ad groups
Single Keyword Ad Groups (SKAGs) are out of fashion, but the principle still holds: tightly themed ad groups with copy that matches keyword intent precisely outperform messy "kitchen sink" ad groups. Ben targets 5-15 closely related keywords per ad group with 3-4 RSA variants each.
12. Cull underperforming ads, audiences and placements
The 80/20 rule applies brutally to Google Ads. Roughly 80% of conversions come from 20% of keywords, the same goes for ads, placements (PMax/Display), and audiences. Pause the bottom performers monthly. Reallocate budget to the winners.
13. Implement value-based bidding
If your business has variable customer values (some leads worth £500, some worth £50,000), generic CPA optimisation undervalues big wins. Move to Maximise Conversion Value or Target ROAS with value-tagged conversions. The algorithm starts chasing high-value conversions specifically.
Clients moving from Target CPA to Target ROAS with value tagging typically see 20-35% improvement in revenue per pound spent, even if headline CPA stays similar.
14. Run a CRO sprint on the funnel, not just the ad
The biggest lever no one talks about. If you double your landing page conversion rate, you halve your CPA. Run quarterly CRO sprints on:
- Above-the-fold value proposition
- Form length and friction
- Trust signals and social proof placement
- Mobile UX (over 60% of paid traffic is mobile)
- Checkout / lead form completion friction
Realistic timeline for CPA reduction
One of the most damaging myths in PPC is that you can cut CPA in a week. You can't - and any agency claiming you can is either lying or about to tank your conversion volume to make the metric look good.
Realistic expectations from Ben's 250+ takeovers:
- Days 1-7: Audit and fix tracking. No CPA movement yet, but you're now measuring accurately.
- Days 8-30: Negative keyword sweeps, location/device/dayparting fixes. Expect 10-20% CPA reduction.
- Days 31-60: Landing page rebuilds, smart bidding stabilises after learning phase. Expect further 10-25% reduction.
- Days 61-90: Value-based bidding, audience layering, copy testing wins compound. Expect final 5-15% reduction.
The cumulative result across the 90 days: typically a 32-48% CPA reduction, with conversion volume holding steady or increasing.
5 mistakes that quietly inflate CPA
- Tracking soft conversions alongside hard conversions in the same column. Newsletter signups and purchases shouldn't both count as "conversions" if the algorithm is optimising blindly across both. Assign different conversion actions and only let smart bidding optimise toward the ones that matter.
- Lowering Target CPA without giving it time to settle. Aggressive target reductions throttle the campaign. Step CPA targets down by no more than 10-15% per fortnight.
- Pausing campaigns too early. A campaign with 12 conversions in week 1 isn't proven to work - it's a small sample. Give campaigns at least 30 days before drawing conclusions.
- Optimising at the wrong level. CPA varies massively by ad group, keyword, audience, device, and time. Looking only at the campaign-level CPA hides the leaks. Drill down weekly.
- Ignoring brand campaigns in CPA reporting. Branded search nearly always has the lowest CPA in the account, and it can mask high-CPA non-brand campaigns when blended together. Always segment.
How Ben tackles CPA in the 90-Day Revenue Takeover
Every account that comes through the 90-Day Revenue Takeover follows the same CPA reduction pattern:
- Phase 1 (Days 1-7): Full audit, conversion tracking rebuild, baseline CPA established by campaign and keyword theme.
- Phase 2 (Days 8-30): Account rebuilt from scratch with proper structure, negative keyword foundation, refined match types, and aligned landing page recommendations.
- Phase 3 (Days 31-60): Relaunch with smart bidding properly configured, audience signals layered in, daily monitoring and optimisation cycles.
- Phase 4 (Days 61-90): Scale - winning campaigns get more budget, value-based bidding rolled out, compound optimisation across all 14 levers above.
The result Ben targets in every takeover: a measurably lower CPA and higher conversion volume by day 90. Both metrics moving in the right direction is what separates real account work from cosmetic optimisation.
The bottom line
If your Google Ads CPA is too high, the answer is rarely "lower your bids." Bids are a single lever in a system of fourteen. The accounts with sustainably low CPAs are the ones where tracking is bulletproof, structure is tight, smart bidding is properly fed, landing pages convert, and the bottom 20% gets pruned monthly.
Most accounts have at least 4-5 of these levers misconfigured at any given time. Fix them one at a time, in order, and CPA reduction follows naturally.
If you want Ben to personally identify exactly where your CPA is leaking - and what it should look like in 90 days - request your free 90-Day Growth Plan.
