Reduce your Google Ads CPA: 2026 guide

-14 to -22% CPA in 30 days: that's the median impact of applying the 10 levers below across the mature accounts we audit. Each lever alone brings 2 to 5%, but systematic stacking produces the composite drop — without raising budget, without changing creative.

If you've been running Google Ads for more than 6 months, you've probably noticed one thing: your CPA has plateaued. You've optimized bids, added negatives, polished your ads — and yet cost per acquisition won't move. Worse, it's creeping up month after month.

Good news: it's almost never a market problem. On the SteerAds 2025-2026 sample, 80 to 92% of accounts (depending on size / sector) have at least 4 of the 10 levers in this guide misconfigured. And the average savings observed when all are fixed: -14 to -22% on CPA in 30 days (variable by maturity), without raising budget.

This guide is the complete roadmap. We start with a quick diagnostic (3 questions), then run through the 10 levers in order of impact/effort ratio. Budget 1 to 2 person-days to apply everything — and 2 to 4 weeks to see the impact in the data.

What counts as a "good" CPA in 2026?

A "good" CPA is a CPA below your gross margin per acquisition. Not a CPA that's low in absolute terms.

Example: if your average order value is $80 and your gross margin 35%, you can spend up to $28 on acquisition ($80 × 0.35) and stay profitable on the first purchase. With a SaaS subscription model and an average LTV of $600, that same "good CPA" can climb to $200 or even $300.

Here are median CPAs observed by sector in the US market (source: our aggregated 90-day data, stratified by vertical):

Takeaway: the gap between the median and the top 10% is 2 to 2.5×. This guide shows you how to close in on the top 10%.

How do you diagnose your CPA in 3 questions?

Before tackling the 10 levers, answer these 3 questions honestly. They'll tell you where to look first.

1. Are you tracking the right conversions, with the right value?

If the answer is "I'm not sure," that's probably problem #1. Broken tracking skews everything else: Smart Bidding optimizes toward a nonexistent goal, you're reallocating budgets on bad signals. If in doubt, start with lever 3 (Smart Bidding) by checking tracking first. See also our guide ROAS, CPA, CPC: the key metrics.

2. Does your account structure segment intent?

If you have ad groups with 30+ keywords, the answer is no. Result: your ad can't be relevant for everyone, your Quality Score plateaus, your CTR drops, your CPC climbs. Direction: lever 1 (match types) + lever 4 (Quality Score).

3. Does your landing page match your ad?

An ad promising "24h delivery" and a landing displaying "3 to 5 days" = high bounce rate, broken conversion, CPA explodes. Direction: lever 8 (landing pages).

Lever 1 · Activate the right match type

Typical CPA impact: -8% to -15% · Effort: 1 to 2 hours

Since 2022, Google has been pushing broad match aggressively in new campaigns. It's convenient for Google (more impressions = more revenue) but catastrophic for you if your negatives don't keep pace: broad will match completely off-topic queries.

The rule: always start in phrase match, switch to broad only after 3 months of solid data and with a robust negative list. Exact match is useful for branded keywords, less so for generic keywords in 2026 (Google increasingly ignores the distinction).

Concrete action: audit the 20 keywords that spend the most. For each broad keyword with a conversion rate under 1%, switch to phrase + negative the useless queries detected.

Lever 2 · Shared negative list across campaigns

Typical CPA impact: -10% to -20% · Effort: 30 min to 2h

This is the highest-ROI lever. Most accounts have negatives at the ad group or campaign level, but no shared list applied across the whole account. Consequence: you add the same negative 15 times, and when you forget one, everything slips through.

4-step method:

1. Export the Search Terms from the last 90 days (Campaigns → Keywords → Search Terms).
2. Filter on queries that cost > 0 and converted 0. Sort by decreasing cost.
3. For each row, decide: relevant (do nothing) or non-relevant (add to the list).
4. Create a Shared negative list (Tools → Shared library → Negative keyword lists), paste your negatives, apply it to all search campaigns.

Recommended starting base: 100 generic negatives (free, download, torrent, jobs, career, internship, review, forum, comment, reddit, quora, wiki, definition, meaning, what is, tutorial, youtube, crack, amateur, kids, baby…) then you refine.

Lever 3 · Smart Bidding (Target CPA / Maximize Conversions)

Typical CPA impact: -12% to -25% · Effort: 15 min

If you're still on Manual CPC on an account over $1,000/month, you're leaving at least 15% on the table. Google's automated strategies use 70+ real-time signals (device, time, user history, intent…) you can't match manually.

Prerequisite: at least 30 conversions in the last 30 days on the campaign (so the algo has enough signal). Below that, start on Maximize Conversions without a CPA target.

Recommended path:

1. Week 1-2: Maximize Conversions (no target) to stabilize conversion volume.
2. Week 3+: switch to Target CPA at your average CPA + 10% (to leave the algo headroom).
3. Month 2: reduce the target by 10% every 2 weeks as long as volume holds.

Lever 4 · Quality Score (the 3 components)

Typical CPA impact: -15% to -50% · Effort: 4-8h

Moving an average QS from 5/10 to 8/10 cuts your CPC by roughly 30%, so your CPA by nearly as much. The 3 components measured by Google: Expected CTR · Ad Relevance · Landing Page Experience. For each: "Above Average / Average / Below Average." Our detailed Quality Score guide breaks down each component and gives specific tactics to push a QS from 5 to 8.

Actions by component:

• Low Expected CTR → the exact keyword must appear in at least one headline. Add 2 numbers ("Save 28%," "24h delivery"), 1 CTA ("Quote in 2 min").
• Low Ad Relevance → your 20 keywords can't all be served by the same ad. Split the ad group into 3-4 thematic groups of 5-10 keywords each.
• Low Landing Page → see lever 8. The 3 key factors: speed (< 2.5s LCP), no mobile pop-ups, coherence with the ad.

Lever 5 · Remarketing and similar audiences

Typical CPA impact: -10% to -18% · Effort: 2-3h

Search remarketing (in Observation mode, not Targeting) lets the algo automatically over-bid on users who've already visited your site. CPA on these audiences is typically 2 to 3 times lower than on cold traffic.

Minimum audiences to create:

1. All visitors in the last 30 days
2. Visitors who viewed a product page (or pricing page)
3. Added to cart without purchase (e-commerce)
4. Existing customers (upload via Customer Match) → exclude from acquisition campaigns
5. Audience similar to existing customers (Google generates it automatically)

Lever 6 · Creative: 3-5 variants per ad group, optimized rotation

Typical CPA impact: -5% to -12% · Effort: 1-2h per ad group

One ad per ad group = no testing = CTR that stalls. Target: 3-5 Responsive Search Ads per ad group, each with 15 different headlines and 4 descriptions. Google tests combinations and keeps the best.

Variation framework:

• Ad 1: price / promo angle
• Ad 2: benefit angle (time saving, ROI…)
• Ad 3: social proof angle (# of customers, rating)
• Ad 4: urgency / scarcity angle
• Ad 5: specific feature angle

Lever 7 · Extensions (sitelinks, callouts, structured snippets)

Typical CPA impact: -6% to -10% · Effort: 1h

Extensions increase the clickable surface of your ad by 10-20% at no extra cost. Result: higher CTR → higher Quality Score → lower CPC → lower CPA. The 4 extensions to always activate:

• Sitelinks — 6 to 8 links to your strategic pages (categories, pricing, reviews, support).
• Callouts — 6 to 10 short benefits ("24h delivery," "Free returns," "Support 7 days a week").
• Structured Snippets — lists of services/products ("Services: SEO, SEA, Social Ads").
• Price extensions — if applicable (products or subscriptions).

Lever 8 · Landing pages (speed, message-match, 1 CTA)

Typical CPA impact: -15% to -30% · Effort: 2-5d

Often the most effort-heavy lever but with the biggest impact. Three non-negotiable rules:

1. Speed: LCP < 2.5s on mobile. Check on pagespeed.web.dev. Grab the easy wins: image compression (WebP), lazy loading, self-hosted fonts.
2. Message-match: the landing H1 must echo the exact keywords from the ad. The user who clicks "Motorcycle insurance quote" must see "Motorcycle insurance quote" at the top of the page, not "Our insurance offers."
3. 1 single primary CTA per landing. No 3 different buttons. No distracting navigation. No aggressive chatbot.

Lever 9 · Dayparting and geographic segmentation

Typical CPA impact: -5% to -12% · Effort: 1-2h

Not every hour of the week is equal. In most B2B sectors, conversions concentrate in business hours (9am-6pm, Monday-Friday). Yet your bids run 24/7.

Action: in Google Ads → Campaigns → Settings → Ad schedule, analyze performance by time slot. If night CPA is 2× daytime CPA, apply a bid modifier of -50% on those slots (or -100% = pause). Same logic for geo: some regions convert poorly (rural, certain states) — adjust.

Lever 10 · Placement exclusions (Display/YouTube)

Typical CPA impact: -20% to -40% on Display · Effort: 1h

If you run a Display or Performance Max campaign, you're likely serving on sites burning your budget: mobile game sites, kids' apps, duplicate-content sites, MFA (Made For Ads).

Action: export the Where ads showed report (last 30 days), sort by cost, exclude anything that spent more than $50 without converting. Also add as systematic exclusions: mobile app sites (adsenseformobileapps.com), "kids," "games," "sensitive content" categories.

Which CPA lever should you prioritize by effort and impact?

Only if Display/Performance Max campaigns are active. Heavy lift = schedule over 2-4 weeks.

How does SteerAds automate these 10 levers?

Of these 10 levers, SteerAds detects 8 automatically after OAuth connection (2 minutes) and proposes one-click fixes. Only levers 6 (creative) and 8 (landing pages) require human intervention — but even for those, the AI generates variants you can validate in 30 seconds.

Start with the free audit (3 minutes, no credit card) to see which of these 10 levers are misconfigured on your own account — then activate automatic optimization if the identified priorities match your goals.

Average result observed across the accounts we track: -14 to -22% CPA in 30 days (depending on account maturity) without raising budget, and 5 hours per week saved vs manual management. To go further once your CPA is under control, our guide to the 10 Google Ads errors covers the structural technical debt that complements this list of optimization levers.

Sources

Official sources consulted for this guide:

• support.google.com
• support.google.com
• support.google.com

FAQ