Rising CPCs in 2026: why paid search costs jumped 12% and what to do

Average Google Ads search CPC hit $2.96 in Q1 2026 — up 12% year-over-year, the steepest jump since 2022. Legal services took a 34% hit. Insurance and finance both crossed $7. Even verticals that historically rode the inflation curve more gently — local services, home repair, ecommerce — saw 6-15% increases. WordStream’s Q4 forecast projects another 8-10% climb through the rest of the year.

The natural reaction is to assume Google is “extracting more rent” and there’s nothing to do. That’s partially true and mostly wrong. The accounts I see actually paying that 12% are the ones running on platform defaults. The accounts running tight feedback loops, strong first-party signal, and smart campaign segmentation are paying 0-4% more YoY on the same volume. The gap between disciplined accounts and default accounts just doubled in 12 months.

What’s actually driving the increase

Four structural factors, not all equal in weight:

1. AI Overviews and zero-click results compressed organic supply

When Google’s AI Overviews answer a query at the top of the page, fewer users scroll to organic results. Click-through rates on positions 1-10 dropped 18-34% across most categories. The clicks that remain are concentrated at the top of the page — and that top of the page is increasingly ads. Demand for those ad slots went up. Supply stayed the same. Price went up. (Full breakdown in the AI Overviews zero-click playbook.)

2. Performance Max ate budget that used to go to cheaper inventory

Performance Max campaigns won’t tell you where your clicks come from at granular level. Many of those clicks are coming from search inventory that used to be auctionable separately as cheap long-tail queries. Now they’re bundled into PMax auctions with higher floor prices. The net effect: auction prices on adjacent terms went up because PMax is bidding into them aggressively.

3. AI-generated competitors flooded auction depth

Tools that auto-generate keyword lists, ad copy, and landing pages dropped the cost of entering paid search to near zero. Vertical-specific competitors that didn’t exist 18 months ago are now in the auction. They’re often inefficient bidders — which doesn’t lower prices, it raises them, because they bid above value while they figure things out.

4. Smart Bidding’s value-based bidding pulled budget toward high-LTV users

Smart Bidding optimizing for conversion value (vs. raw conversions) means it’ll happily pay more for a click that’s likely to convert at high value. That’s good for the advertiser whose LTV signal is correct. It’s also another mechanism pushing average CPCs up across the auction.

What’s working: the defensive playbook

The accounts holding effective CPA flat or improving while market CPCs rise share five practices. None of them are exotic. All of them are missing from most accounts.

Practice 1: Strong first-party data piped back to the bidder

This is the single biggest lever. First-party data activation — Customer Match with LTV segmentation, Enhanced Conversions, offline conversion imports — gives the bidder better signal about who actually converts at value. With cleaner signal, the bidder pays more efficiently per conversion even as auction-level CPCs rise.

Accounts running strong first-party signal show effective CPA flat or down 5-10% YoY despite the 12% CPC increase. Accounts on default signal show effective CPA up 10-18%. Same auction, different outcome.

Practice 2: Segmenting campaigns by margin tier

Not every conversion is worth the same. A consultation request that closes at 35% with a $12K deal value is worth a much higher CPL than a top-of-funnel newsletter signup that converts at $80 LTV. Mixing them in a single campaign means Smart Bidding optimizes toward the cheaper conversion type and ignores the high-margin one.

The fix: separate campaigns by conversion tier, each with its own tROAS or tCPA target calibrated to actual margin. The work takes 1-2 days. The CPA improvement is typically 15-25%.

Practice 3: Negative keyword discipline at the SKAG / query level

PMax and Smart Bidding have made keyword targeting feel like a relic. It isn’t. Negative keywords are now more important than positive ones because the platforms are inferring intent from broad signals, and they get it wrong often.

Pull your search terms report weekly. Negative out any query that doesn’t match buyer intent. For PMax specifically, use the brand exclusion list and the negative keyword list (added in 2024 and still under-used). Accounts that do this religiously typically see 8-15% lower CPA within 30 days.

Practice 4: Audience-layered bidding instead of broad targeting

Bidding higher on Customer Match audiences, in-market segments matched to your offering, and remarketing lists — and bidding lower on broad untargeted impressions — is one of the oldest tactics in paid search. It still works. The 2026 update: layer multiple audience signals together and use observation bidding adjustments on Performance Max audience signals to feed the bidder strong starting hypotheses.

Practice 5: Real attribution feeding the bid

If your conversion events are wrong, your bidder is wrong. The most common attribution failure in 2026: counting form submissions as conversions when only 8% of forms close to a deal. The bidder optimizes toward form fills and ignores actual revenue. Fix: pipe closed-deal data back via offline conversion imports. The bidder relearns within 2-4 weeks and starts optimizing toward revenue.

Marketing attribution stack 2026 covers the full setup.

The campaign-mix shift that’s working in 2026

For SMB-to-mid-market accounts feeling the squeeze on Google search costs, the campaign mix that’s delivering best blended ROAS in 2026 looks roughly like this:

• 40-50% Performance Max with strong first-party signal feed
• 15-25% branded search defense
• 10-20% tightly targeted non-brand search on highest-intent keywords with audience layering
• 10-15% Demand Gen for top-of-funnel
• 5-10% YouTube Ads for brand reinforcement
• 0-10% Discovery / display retargeting

Compared to the 2024 mix (which was heavier on non-brand search), the shift moves budget toward channels where the auction inflation is less severe and where strong first-party signals give a bigger relative advantage.

What’s not working

For the sake of completeness, the moves that consistently fail in 2026:

• Lowering tROAS targets to “stay in the auction” — pulls in unqualified traffic, conversion rates drop, blended CPA gets worse not better
• Adding “exact match” keywords expecting tight matching — exact match in 2026 is functionally a phrase match with AI synonym expansion, and the auction pricing has caught up
• Trying to outcompete on creative alone — strong ads help, but they can’t overcome a broken signal feed
• Pausing campaigns and waiting for CPCs to drop — they won’t, and you lose Smart Bidding’s learning history
• Switching agencies without changing the data plumbing — most CPC problems aren’t agency-skill problems; they’re tracking and attribution problems

A 60-day “stop bleeding” project

For an account where CPCs jumped sharply YoY and CPA is following:

• Days 1-10: Audit conversion tracking end-to-end. Confirm CAPI / Enhanced Conversions are firing with full match keys. Confirm offline conversion imports are running for any post-click revenue.
• Days 11-25: Segment campaigns by margin tier. Set bid targets calibrated to actual margin per conversion type. Layer audiences.
• Days 26-40: Build out LTV-segmented Customer Match lists. Push to Google and Meta. Observe Smart Bidding recalibration over 2-3 weeks.
• Days 41-60: Negative keyword sweep weekly. PMax brand exclusion review. Creative refresh on the campaigns where ad quality scores are mid (under 7).

Expected outcome by day 60: effective CPA flat to down 8-15% even as auction CPCs continue rising. The auction got more expensive; the account got more efficient at the same time.

FAQ: rising CPCs in 2026

Are CPCs going to keep rising in 2027? Almost certainly, yes — the structural drivers (AI Overviews compressing organic clicks, PMax bundling, AI-generated competitor entry) aren’t reversing. The differential between disciplined and default accounts will keep widening.

Is it time to shift budget away from Google search entirely? For most accounts, no — search intent is still the highest-converting traffic at scale. The fix is running search more efficiently, not abandoning it. Diversifying into LinkedIn, Reddit, Pinterest, or TikTok makes sense as an addition, not a replacement.

What’s a “normal” CPC inflation rate to budget for in 2026? Plan for 10-15% blended CPC inflation YoY. If your account is running clean signal, your effective CPA should be flat or improving — the gap is what your data discipline buys you.

Is Smart Bidding making CPCs worse? It’s making auction-level CPCs higher because it bids more on high-LTV users. But it’s making effective CPA better when the signal feeding it is correct. The problem isn’t Smart Bidding; it’s bad signal flowing into Smart Bidding.

Can negative keywords still meaningfully reduce CPCs? Yes — especially on Performance Max and broad match search campaigns. A clean negative keyword list typically reduces wasted spend 8-15% in mature accounts.

The honest 2026 framing

Rising CPCs in 2026 aren’t a temporary spike. They’re the new operating environment, and they’re going to keep rising. The advertisers who win in this environment aren’t the ones with the biggest budgets — they’re the ones with the cleanest signal, the tightest campaign segmentation, and the most disciplined attribution.

The defensive playbook above doesn’t require new tools or new talent. It requires roughly two months of disciplined work to fix the tracking, segment the campaigns, and feed real first-party signal to the bidder. The accounts that do that work are running 20-40% cheaper effective CPA than their competitors right now. That’s not a future advantage. It’s available today.