You Don't Pay Too Much. You Pay Wrong.

Avinash Kaushik spent 16 years at Google. Now, as Chief Strategy Officer at Human Made Machine, he recently made a claim that rattled a lot of agencies: SEO and digital marketing budgets should be renegotiated 25% to 75% lower. A piece on Search Engine Journal lays out the full argument, backed by two decades of industry experience. At first glance, it reads like a death sentence for anyone selling digital marketing services.

But Kaushik isn't saying "you pay too much." He's saying "you're paying for things AI already does." That distinction matters.

At difrnt., we see this debate from both sides: as an agency delivering marketing services, and as a team that uses AI daily in our internal processes. Our takeaway isn't that prices need to drop. It's that the structure of marketing contracts needs to change fundamentally. And it's a conversation every client should be initiating, regardless of budget size.

What AI already handles and why it matters for your contract

The numbers in Kaushik's argument are hard to dispute. His estimate: 80% of campaign setup work is now automated by the platforms themselves. Keyword research, account structure, bid management, pacing: all handled programmatically. And 60% of reporting costs disappear when AI tools can interpret data without human translation.

This isn't futuristic talk. Google Ads already uses integrated AI models for bidding, audience targeting, and budget optimization. GA4 flags anomalies on its own. A technical SEO audit that used to take two weeks now generates in a few hours with specialized tools. And when bidding is automated, the human role shifts from execution to strategic direction.

Kaushik also makes a point that few openly acknowledge: manual intervention in AI-driven campaigns can actually sabotage the algorithms. When an agency "rescues" a campaign during its learning period, it resets the optimization timeline. Sometimes the most valuable thing a specialist does is not touch anything and let the data work.

The real question isn't whether AI does this work. It's why some agencies still bill as though they're doing it manually.

From volume to value: restructuring the contract

Kaushik proposes a clear three-pillar restructure. 40-50% of the budget as a base retainer covering governance, data infrastructure, and strategic direction. 30-40% as project fees for strategic analysis, creative work, and authority building. And 15-25% as performance bonuses tied directly to revenue or real business impact, not platform vanity metrics.

As Head of Strategy at difrnt., I see daily where value gets lost in traditional contracts. A client paying for 4 articles per month, 2 reports, and a quarterly audit receives activity. But do they receive direction? Do they get a strategy connecting SEO to real business objectives? Do they get someone who tells them "don't do that" when the data shows it isn't working?

Most of the time, no. A contract built around deliverables rewards volume. It doesn't reward the decision to do less, but better. As Kaushik puts it: "You can pay for the past, or you can pay for the present." Simple framing for a problem we see in nearly every account audit we run.

This doesn't mean your agency is incompetent. It means the contract structure creates the wrong incentives. It's a design problem, not an intent problem. And any client can fix it by asking "what am I actually paying for?" before asking "am I paying too much?".

Three checks before any renegotiation

The first step isn't cutting the budget. It's auditing the current contract and sorting every line item into three categories: what AI already handles, what requires human judgment, and what falls under data governance.

Data ownership. This is the first checkpoint. Do you have direct access to GA4, Search Console, advertising accounts, and tracking pixels? If not, any renegotiation is a risk. In markets like Romania, where many businesses started their digital journey with freelancers who created accounts under personal emails, ownership isn't a technical detail. It's a precondition for any meaningful discussion about pricing or performance.

Output clarity. If your agency can't explain how SEO contributes to your sales pipeline, you're not paying for strategy. You're paying for reports. And data without clear decisions doesn't generate results, no matter how many dashboards you receive each month. Ask yourself honestly: "What business decision did I make based on the last report?" If the answer is "none," that's where the real problem sits.

Adaptability. The agencies that stay relevant aren't the ones resisting change. They're the ones who've already integrated AI into their processes and are using the freed-up time for judgment work, not for producing more of the same deliverables. Ask your agency what processes they've automated in the past year. If the answer is vague or defensive, you're probably paying for manual execution that no longer makes economic sense.

What stays fundamentally human

Kaushik is right that a large portion of execution has become commoditized. But four things remain impossible to automate, no matter how advanced AI gets.

Business context. Why Q3 is critical for your company, why a competitor just changed their pricing strategy, why the founder won't accept a particular brand positioning. An AI model doesn't sit in board meetings. Internal context is built through months of close collaboration, not a few-line prompt.

Stakeholder relationships. Alignment between marketing, sales, and management is deeply human work. An SEO campaign without buy-in from the sales team generates traffic, not customers. No automated bidding algorithm solves an internal priority conflict.

Gray-zone decisions. When CTR goes up but revenue goes down, you need judgment, not an automated report. When two strategies show comparable data but different risk profiles, you need experience and the willingness to recommend the less comfortable option. The real value of a strategic partner shows in exactly these moments.

And long-term brand building. In a world where AI cites brands (not just websites), who writes and how the brand positions itself matters more than ever. Authority is built through consistency, original perspectives, and sustained presence, not through volume of generic content produced by the kilogram.

Kaushik estimates 15-25% growth in demand for exactly this type of strategic work. So this isn't about spending less on marketing. It's about redirecting the budget toward activities that produce real impact.

The industry is changing, and Kaushik names the direction correctly. But the answer isn't "pay less." It's "pay for what matters." If your marketing contract looks the same as it did in 2020, the budget isn't the problem. The structure is.