What the agency fee structure actually reflects
Three models account for almost all PPC agency pricing:
- Percentage of ad spend: most common model, 10 to 20 percent range, with 15 percent as the industry median. Aligns agency revenue with client spend growth, but does not align with campaign performance or the actual work involved.
- Flat retainer: $1,500 to $10,000 per month for US-based agencies; broadly SGD 600 to 8,000 for Singapore-based specialists, depending on scope and account complexity. Predictable for the client; uncorrelated from results.
- Hybrid: base retainer plus a performance-based component tied to conversion volume, ROAS, or new customer acquisition cost. The most honest structure -- aligns at least part of agency economics with client outcomes.
AI automation has reduced the execution cost of PPC management by an estimated 20 to 30 percent over the past two years -- smart bidding, automated extensions, responsive search ad generation, and now AI Max are absorbing tasks that once required hands-on work. Most agencies have not passed those savings to clients. The standard fee has held at 15 percent of spend.
The resulting question is not whether the fee is "fair" -- it is whether the remaining human work justifies the fee. That requires knowing what the AI handles and what it does not.
leapbuzz's search advertising service prices around strategy, architecture, and measurement -- not around routine campaign management tasks the platforms now handle. The engagement model reflects what AI cannot do, not what PPC management used to cost when it required manual bid adjustment.
Google AI Max for Search: what it automates and what it cannot
Google AI Max for Search is a campaign-level toggle, confirmed live on April 30, 2026. It is not a new campaign type -- it extends AI control over three dimensions of existing Search campaigns:
- Query matching: AI Max expands match beyond current broad match to include semantically adjacent queries that share the user's intent without containing the keywords. Practically, this means your ads may serve on queries your keyword list would never trigger.
- Ad copy generation: AI Max generates ad headline and description combinations from your uploaded asset library, selecting the combination predicted to perform best for each query context.
- Landing page selection: the system can select landing pages beyond your specified final URL if it predicts a different page on your domain would convert better for the query.
What still requires human expertise
AI Max does not resolve -- and introduces -- several governance requirements that remain human decisions:
- Cross-campaign conflict detection: if AI Max on a generic campaign starts serving on queries that your brand campaign targets, you get internal auction conflict and inflated CPCs. Identifying and managing that conflict requires human oversight across the account structure.
- Negative keyword governance: AI Max's expanded matching surfaces queries the keyword list would not include, many of which are irrelevant. Active negative keyword management becomes more important, not less.
- Brand governance: AI Max may generate ad copy that does not reflect your current brand positioning, approved claims, or legal sign-off requirements. Content review cadence increases, not decreases.
- Margin modelling: the AI optimises toward whichever conversion event you define. Defining the right conversion hierarchy -- and weighting by contribution margin rather than revenue -- is a strategic decision the platform cannot make.
- CRM signal quality: AI Max's performance scales with the quality of conversion data you feed it. Offline conversions, phone call attribution, and CRM match rates are human-managed inputs that determine the ceiling of AI performance.
The orchestrator shift. The model for PPC expertise has moved from "tinkerer" (adjusting bids, testing manual keyword variants) to "orchestrator" (designing the conversion signal architecture, governing the AI's inputs and exclusions, measuring incrementality). A partner still working in the tinkerer model post-AI Max is delivering diminishing returns.
CPC benchmarks by industry, 2026
The table below draws from WordStream's 2026 analysis of 13,474 US Google Ads accounts (April 2025 to March 2026). These are US median values; see the five-market section for international adjustments.
| Industry | Avg CPC (USD) | YoY trend | Note |
|---|---|---|---|
| Legal | $9.87 | Rising | Highest-CPC category; personal injury and mass tort terms exceed $50/click |
| Home Improvement | $8.33 | Stable | High local competition; brand terms significantly cheaper than generic |
| Dental | $8.00 | Rising | Implant and cosmetic dentistry terms can exceed $15/click |
| B2B SaaS (non-branded) | $5.34 | +29% YoY (steepest rise) | Category saturation and AI tool proliferation driving CPC inflation fastest here |
| All industries average | $5.42 | Moderate rise | Blended across all categories in the WordStream dataset |
| Finance (broad) | $3.39 | Stable | Lower than expected due to broad category definition; high-intent mortgage/insurance terms are materially higher |
| Ecommerce | ~$1.16 | Declining | PMax has shifted spend toward Shopping placements with lower per-click cost |
B2B SaaS non-branded CPCs rising 29 percent year-on-year is the most significant trend in the dataset. The cause is category proliferation: every AI tool, productivity app, and workflow platform is bidding on the same generic terms ("project management software," "CRM platform," "AI writing tool"). Branded term CPCs in SaaS have stayed flat or declined as AI-assisted brand campaign management has improved quality scores. The budget implication: generic SaaS terms are becoming less viable for prospecting, and branded search defense is more cost-efficient than ever.
What a PPC management contract should and should not cover
Scope creep in PPC engagements almost always flows in one direction: clients expect deliverables the contract does not include. The list below is the industry standard scope for a full-service PPC management engagement.
What should be in scope
- Campaign strategy and structure design (initial and ongoing refinement)
- Keyword research and negative keyword management (ongoing governance)
- Bid strategy selection, monitoring, and adjustment recommendations
- Ad copy creation, testing, and rotation management
- Audience list management (remarketing, customer match, in-market audiences)
- Conversion tracking setup, auditing, and quality verification
- Budget pacing and allocation recommendations across campaigns
- Performance reporting at the agreed cadence
What is typically out of scope (should be explicitly stated)
- Landing page design and CRO work (separate engagement)
- Creative asset production: video, photography, display banners
- CRM integration and offline conversion import setup
- Organic search strategy or SEO work
- Social media advertising (Meta, LinkedIn, TikTok) unless bundled
- Marketing attribution modelling and incrementality testing
A partner that lists these exclusions explicitly in the contract is being transparent. A partner whose contract is vague about exclusions will face scope disputes that consume the account manager's time and erode the relationship. Ask to see the full scope exclusions list before signing.
Agency vs in-house vs AI consultancy: the real comparison
Internal PPC management commitment grew from 44 to 71 percent of companies between Q4 2023 and Q4 2024, per the State of PPC 2025 report. The shift is AI-driven: automated bidding, responsive search ads, and now AI Max have removed enough routine management work that smaller accounts can be run competently in-house with less specialist time than they required three years ago.
| Model | Strengths | Weaknesses | Best fit |
|---|---|---|---|
| In-house | Deep product knowledge, fast sales feedback loops, no fee markup on spend, full data access | Limited cross-vertical pattern recognition, harder to scale team expertise, senior talent hard to retain | Accounts over $50K/month where internal expertise can be justified; companies with strong analytics infrastructure |
| Traditional agency | Cross-vertical pattern recognition, platform relationships and beta access, team depth to handle surges | Margin markup on spend, account manager turnover, templated approaches across clients, slow internal sign-off | Companies without internal search expertise; accounts with complex multi-channel requirements across many platforms |
| AI-native consultancy | Strategy-only focus (not billing routine management hours); conversion architecture, incrementality, budget arbitration -- the high-leverage decisions | Not structured for full execution outsourcing; requires client to have some internal capability or a lower-cost execution partner | Companies with in-house execution capability that need a senior strategic layer; accounts where the question is 'is this working' not 'can you run it' |
leapbuzz positions as the AI-native consultancy model: we do not bill hours on Smart Bidding oversight or routine ad copy rotation. Our search advertising engagement focuses on the conversion signal architecture, incrementality testing design, and cross-channel budget arbitration that AI cannot make. That is a different purchase than full-service agency management, and it makes most sense alongside an in-house executor or a lower-cost execution partner who handles the routine management layer.
Five-market PPC cost context
| Market | CPC vs US | Google market share | Key consideration |
|---|---|---|---|
| USA | Baseline | ~88% | Highest-competition PPC market; AI tool proliferation driving B2B SaaS CPCs up 29% YoY; Microsoft Ads meaningful for B2B (Bing audience) |
| Canada | ~29% below US nationally; major metros approach US levels | ~88% | Similar platform mix to US; Quebec requires French-language ad copy (OQLF); privacy law PIPEDA governs remarketing lists; Amazon Canada marketplace relevant for ecommerce PPC |
| Australia | ~5% below US median | ~93% | Bing/Microsoft Ads has a more meaningful market share than in SG; ACCC governs comparison advertising and pricing claims in ad copy; state-level price variations affect dynamic ad copy |
| Singapore | High for financial/property categories; lower overall volume | ~97% | Financial services CPCs among APAC's highest (MAS-regulated sectors); Bing market share negligible; small total market volume means budget efficiency is more important than scale |
| Malaysia | Significantly lower than US across most categories | ~95% | Lower CPCs mean paid search is more accessible for mid-size businesses; Google remains dominant; local language ad copy (Bahasa Malaysia) often outperforms English for consumer categories |
A note on Singapore and Malaysia CPC data: published benchmark data with the granularity and sample size of the WordStream US dataset does not exist for these markets in verifiable form. The relative cost direction above reflects practitioner consensus and directional data; absolute CPC benchmarks should be drawn from your own account data rather than third-party sources for these markets.
Seven questions to ask before signing a PPC management contract
These questions have two functions: they identify competent partners and they frame the commercial relationship before misaligned expectations cause issues.
- What is your methodology for setting the initial bid strategy, and when do you override Google's algorithm recommendation? A capable answer describes specific scenarios (new conversion data, competitor entry, margin changes) where they deviate from default smart bidding. An answer that defers to "what Google recommends" is not an answer.
- How do you structure campaigns when branded and non-branded terms are in the same account? Brand and generic should be segmented. If they are not, brand's higher quality scores inflate the generic campaign's average, masking underperformance. The answer should describe how brand budget is protected and how brand contribution is measured separately.
- What is your negative keyword governance process and how often is it reviewed? With AI Max expanding match, negative keyword management becomes more critical. A weekly or bi-weekly search term review cadence is the minimum standard. Monthly is insufficient.
- How do you prove the account's performance is incremental to what would have happened with no management? Most partners cannot answer this. The ones who can describe holdout testing methodology, geo-based incrementality design, or platform lift studies. If the answer is "we compare to previous period performance," that is not an incrementality methodology.
- What deliverables are explicitly excluded from your contract? Exclusions that are not written down will become disputes. Landing page, creative, CRM, and attribution are the common gaps.
- Who specifically will manage this account and what is their Google Ads experience? Senior pitchers who hand off to junior executors is the standard agency model. Know who is in the account day-to-day before signing.
- What reporting cadence and metrics will you commit to in writing? Commitments are reporting frequency, specific metrics included (including quality metrics like cost-per-conversion and conversion rate, not just impressions and clicks), and access to the live account so you can verify independently.
leapbuzz publishes its engagement scope and answers all seven of these questions in the first discovery call, before any commercial conversation. Our search advertising service and our performance marketing service both start with the question of what AI handles versus what strategic judgment delivers -- because the answer determines what you should be paying for.
