Why the SEO agency model broke (and what to hire instead)

The agency SEO model runs on playbooks, and AI search has been invalidating playbooks faster than agencies can rewrite them. That is the whole argument. If your agency’s results have gone soft over the last two years and nobody on the account can explain why, this is usually what happened.

Now, “agencies bad, consultants good” is an extremely convenient thing for a consultant to say, so let me make the narrow version of the claim and then tell you where it falls apart. I’ll also tell you when you should hire the agency instead of me, because that situation is real and I’d rather you get it right.

The traffic assumption underneath every SEO retainer

Almost every SEO program sold between roughly 2015 and 2023 rested on one assumption: if you rank, you get the click. Rankings were the deliverable because rankings converted reliably into sessions. That assumption is now broken, and there is decent data on how broken.

The Pew Research Center tracked the browsing behavior of 900 U.S. adults and found that when an AI summary appeared, users clicked a traditional result in 8% of searches, against 15% when no summary appeared. Roughly half. Links inside the AI summary itself got clicked in about 1% of visits. Users were also more likely to end the session entirely after seeing a summary (26%, versus 16% without one).

SparkToro’s 2026 analysis of Similarweb clickstream data found that 68% of U.S. Google searches ended without a click between January and April 2026. The share of searches producing at least one click dropped 9.51 percentage points between 2024 and 2026.

So position three is not what position three used to be worth. Everybody in the industry knows this now. The interesting question is why some shops adapted in a quarter and others are still shipping the 2021 deliverable list.

What you are actually buying from an agency

An agency sells you a system. That’s not an insult, it’s the business model, and it’s why agencies can serve forty clients at once when I can serve a handful.

The system looks like this:

• A documented playbook that encodes what worked
• A service catalog built from that playbook (X posts, Y links, Z audits per month)
• A staffing pyramid where senior people sell and scope, and less experienced people execute against the playbook
• Training and QA that keep execution consistent across accounts

Every piece of that is a genuine strength when the fundamentals hold still. The playbook is how an agency delivers the same quality on your account and on the account of the client who signed last Tuesday. It’s how they hire a 26-year-old in March and have them productive by May.

But notice what the playbook is in accounting terms. It’s the agency’s core asset, and its value depends entirely on how long the underlying rules stay true. When search fundamentals shift every five years, a playbook is a fantastic investment. When they shift every eight months, it’s a liability that the whole org is structurally motivated to defend.

Why the playbook is the thing that fails

Three things happen inside an agency when the ground moves, and they compound.

1. The playbook is the product, so admitting it’s stale is expensive. Rewriting it means retraining everyone, rebuilding the service catalog, and renegotiating what clients were promised. There is enormous quiet pressure to decide the old approach still basically works.

2. Execution is separated from observation. The person running your account is following a process, not questioning it. That’s the design. It also means the people closest to your data are in the worst position to say “this stopped working.”

3. And honestly, the incentives point the wrong way. Retainers reward consistent delivery of the agreed scope. Nobody gets promoted for walking into a QBR and saying half of what we sold you last year is now low-value work. I have watched capable, well-meaning people at good agencies fail to make that argument internally, not because they were wrong, but because the org had no place to put it.

None of that requires anyone to be lazy or dishonest. It’s just what happens when you productize expertise in a category that stops sitting still.

Where agencies genuinely win

I said I’d tell you when to hire the agency. Here:

• You need volume. Forty location pages, a 300-SKU catalog migration, content at a pace no single person produces. Agencies have capacity. I don’t.
• You need many specialisms at once. Paid, SEO, creative, dev, all coordinated, all on one invoice.
• You need institutional continuity. If your program has to survive whoever quits, an agency with documentation and a bench is genuinely lower-risk than one person. (If I get hit by a bus, your program stops. That’s a real thing to weigh.)
• Procurement is in the room. Enterprise vendor forms, insurance minimums, MSAs, security review. Agencies are built for that and I’m mostly not.

If two or more of those describe you, hire the agency. Just make them show you what they changed in the last year.

Where a consultant wins right now

The consultant advantage isn’t talent. There are brilliant people at agencies. The advantage is a shorter distance between noticing something and acting on it.

When I see AI Overviews eat a query cluster for a client, there’s no playbook to amend, no team to retrain, no service catalog that becomes wrong. I read the change, test it on my own properties, and adjust the same month. That’s not discipline on my part. It’s just that a shop of one has nothing to protect.

The other piece is that you get the senior person on the actual work. At a typical agency, the person in your pitch is not the person in your account. That gap is survivable when the playbook is good. It is much less survivable when the right move requires judgment that isn’t written down anywhere yet.

Agency vs consultant, honestly

	Agency	Independent consultant
Capacity	High — parallel execution across a team	Limited by one person’s hours
Speed of adaptation	Slow; playbook and training must change first	Fast; nothing to retrain
Who does your work	Usually a junior or mid-level executor	The person you hired
Continuity risk	Low — bench and documentation	High — single point of failure
Breadth	Many channels under one roof	Deep in a narrow lane
Typical Portland cost	$130–200/hr, $1,000–3,000/mo programs	Comparable rate, no account layer

The cost line surprises people. Independent doesn’t reliably mean cheaper. My rate is $130 an hour, which sits at the low end of the Portland agency range, but plenty of good consultants bill above what an agency charges. What changes is the ratio of what you pay to who touches the work.

Five questions to ask either one

Use these on agencies. Use them on me too.

1. What did you change about your approach in the last twelve months, and what triggered it?
2. Who specifically will do the work after I sign, and what’s their experience?
3. What percentage of my target queries currently trigger an AI Overview, and what’s your plan for those?
4. What will you tell me to stop doing?
5. What can’t SEO fix for my business?

Question one is the whole thing. A shop that can’t name something concrete it changed since last year has either not been paying attention or has decided the change doesn’t apply to them. Question four is the tiebreaker, because anyone selling you a system will find it very hard to recommend less of it.

What I actually think

SEO didn’t stop working. Organic search is still one of the most profitable channels a business can build, and I’d argue the gap between GEO and SEO is narrower than the people selling GEO as a separate product want you to believe. What changed is that the click is no longer the automatic reward for the ranking, which means the work has to aim at being cited and trusted rather than merely ranked.

That’s a judgment-heavy job right now. Judgment-heavy work is exactly what productized delivery is worst at.

If you want to talk about what that looks like for your business, get in touch. If you’d rather see how I structure the work first, the Portland SEO services page lays out the engagement models, and the SEO consultant page covers the advisory version where your team executes.

---

Sources

• Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results (July 22, 2025) — pewresearch.org
• SparkToro, In 2026, Less Than One Third of Google Searches Still Send a Click — sparktoro.com