How to get better SEO work from your agency with Paddy Moogan

Most agency relationships start well.

There is energy. There is a kickoff. There are new faces, fresh audits, quick findings, a plan, and usually a few early ideas that make everyone feel like the relationship was the right call.

Then real life turns up.

The meetings become familiar. The same people join the same calls. Reports land on the same day each month. The agency is still doing the work, but somehow the relationship feels less sharp than it did at the beginning.

That was the thread running through my conversation with Paddy Moogan.

Paddy and I go back a long way. We worked together at Distilled, where he eventually ran the UK business. He later co-founded and ran his own agency for ten years, exited in 2024, and now works with agency founders as a coach, board member, and advisor and supports new and emerging agency leaders.

So he has seen this from almost every angle: employee, agency leader, founder, seller, advisor. I have my own scar tissue from building and running Distilled, and these days at SearchPilot we spend most of our time with in-house SEO, product, and growth teams at large ecommerce and enterprise sites.

That combination made this a fun one. A bit nostalgic in places, yes. But mostly practical.

Because the question is not "are agencies good or bad?" That is too simple. The better question is: where do agencies still create real value, and what does the in-house team need to do to get that value?

Key takeaways

• Do not default to hiring an agency. Start by asking what should stay in-house and where outside experience would genuinely change the outcome.
• Agencies are most useful when they give you decision confidence: pattern recognition, outside context, and backup on calls where the stakes are high.
• Some SEO work should stay close to the business: product priorities, trading goals, engineering constraints, internal politics, customer knowledge, and final accountability.
• Agency relationships often drift around month five or six. The early work is done, the meetings settle, and both sides can get too comfortable.
• Treating an agency like a task machine usually gets you task completion. Involving them in the right conversations gets you better thinking.
• AI will change the shape of agency work, especially junior delivery tasks, but it also gives newer people a different way to stand out if they know how to use it well.
• The best long running relationships are not the easiest ones. They usually include high expectations, awkward conversations, and enough trust to keep pushing.

Do not default to an agency

Paddy started with a point that might sound strange from someone who has spent most of his career in agencies.

Brands should not default to using agencies.

I liked that because it is honest. Agencies can be useful, sometimes incredibly useful, but they should not be the automatic answer to every SEO problem. If I were running an in-house SEO team, I would want to keep as much ownership inside the company as possible.

That does not mean doing everything yourself. It means being clear about what you are actually buying.

Too often, someone joins a brand and brings in an agency because that is how they have worked before. New head of SEO, new CMO, new VP, new agency. I have benefited from that pattern, and I have also been on the wrong end of it. It can be rational. It can also be lazy.

The first question should be: what do we need that we cannot get from inside the building?

If the answer is "more hands", fine. Sometimes capacity is the problem. But if the answer is "we need judgement from someone who has seen this 20 times before", that is a different kind of agency relationship.

And usually a better one.

The agency's best value: decision confidence

Paddy used a phrase I liked: decision confidence.

That is where agencies can still earn their place.

An in-house team has depth. They know the business, the politics, the product priorities, the trading pressures, the customer base, and the things that will never show up in a web crawl. That context is hard to beat.

But the tradeoff is that you are usually working on one site, or one portfolio of sites. You may not have seen the same problem across twenty other companies in the last two years.

That is where an agency can help.

A good agency can say, "We have seen this before. Here is what happened when another retailer tried it. Here is what went wrong in the migration nobody expected. Here is the bit of the plan that looks safe, but probably is not."

Site migrations are a simple example. The process is not mysterious. Most experienced SEOs know the checklist. But migrations go wrong for odd reasons. The value of someone who has been through twenty of them is not that they know the checklist better. It is that they remember the weird failure modes.

That matters even more when the decision is expensive, political, or scary.

A CMO making a big call can feel pretty lonely. A Head of SEO pushing for a major change across engineering and product can feel exposed. A good agency sits at that table and gives the in-house team confidence to act.

That is much more useful than another 80 page audit that nobody reads.

Consulting versus doing

This brought us back to a distinction Paddy and I have both lived with for years: consulting versus agency delivery.

I have never cared much for the purity test of whether a company is "really" a consultancy or "really" an agency. But there is a useful distinction underneath it.

Consulting is judgement, context, advice, frameworks, experience. Agency delivery is doing the work: producing the audit, writing the copy, creating the brief, building the report, managing the backlog.

Both can matter. But they are not the same thing.

Paddy's view was that the "doing" side is more exposed than it used to be, partly because of AI and partly because in-house teams have got stronger. If an agency's value is mainly "we will complete this task for you", that work is more at risk.

The work that is harder to replace is the experienced judgement. The person who can sit with the client, understand the stakes, read the room, and help make a better decision.

That is not junior work. It is not mechanical work. It comes from seeing patterns over time.

Which leads to the awkward question we kept circling back to: where do those experienced people come from if the junior work disappears?

The junior talent problem

A lot of great in-house SEOs started in agencies.

That makes sense. Agency life is a firehose. You see many sites, many industries, many CMS setups, many stakeholder groups, many weird technical problems. A year in agency life can feel like three or four years somewhere else.

But if AI starts hollowing out the junior tasks, where do tomorrow's strong mid-level people come from?

Paddy has been talking to agency operators about this exact problem. His view is not doom laden, but it is realistic. Junior people are under more pressure than people with years of experience. The basic work is easier to automate, or at least partially automate.

But there is another side to it.

The next generation coming into the workplace may be much more AI native. Some of them already are. They will not use AI as a bolt-on. They will use it as part of how they think, research, write, analyze, build, and communicate.

That can be a real advantage.

If a junior candidate can sit in an interview, open Claude or ChatGPT, and show how they reason through a problem, that tells you something. Not because the answer is perfect, but because you can see the thinking.

Paddy mentioned his own Distilled interview years ago, where he was asked to map out a site architecture on a whiteboard. The point was not the exact architecture. It was watching how he thought.

The modern version may be: share your screen, use the AI tool, and show me how you solve the problem.

That is a different skill. But it is still a skill.

AI needs management

One of the funnier parts of the conversation was how similar AI prompting is to management.

People who would give a human colleague a hopelessly vague one-line brief will happily write five pages of detailed context for an AI model. They will explain the goal, provide examples, give constraints, correct the output, refine it, and go back and forth patiently.

Which is, of course, what they should have been doing with people all along.

Paddy made the point that good communication skills may become even more useful in an AI-heavy workplace. You need to explain what you want. You need to break down the problem. You need to judge the output.

That is the line someone had used with Paddy: AI can help you cook an amazing meal, or it can burn your house down.

I think that is about right.

For agencies and in-house teams, the lesson is not "replace juniors with AI." It is "hire and develop people who know how to use these tools without losing their judgement."

What should stay in-house

Some work should not be fully outsourced.

The closer a task is to business context, the more careful I would be about handing it away.

An agency can help you think. It can bring examples. It can challenge your plan. It can do research. It can provide backup. But the in-house team owns the tradeoffs.

Product priorities live inside the company. Engineering constraints live inside the company. The politics of getting something shipped live inside the company. Customer understanding often lives inside the company too, even if it is not always well distributed.

That is why the best agency relationships are not a handoff. They are a collaboration.

The agency should not be off in a corner producing things. It should be close enough to understand what matters, but not so close that it forgets its job is to bring outside perspective.

That is a balance. It is also where a lot of relationships go wrong.

The account that complains is not always the one at risk

This is something I have thought about a lot on the SaaS side at SearchPilot as well.

The account that complains is often less at risk than the account that has gone quiet.

Complaints mean there is still energy in the relationship. The client still cares enough to tell you what is wrong. That does not make the conversation fun. It does make it useful.

The real danger is apathy. The relationship feels polite, quiet, smooth, and dead.

Paddy agreed. If a client raises a concern, that is a sign they still care. Both sides need to show that they care at different moments.

That is where the intangible parts of agency work start to matter more. The scope of work and deliverables may get you through the first year. After that, the relationship itself matters more.

Small moments count. Sending an article because it made you think of the client. Asking a better question in a meeting. Remembering a business priority from three months ago. Bringing an idea before being asked.

Those things sound soft, but they are often the difference between being treated as a partner and being treated as a supplier.

Do not turn your agency into a task machine

If you treat an agency like a task machine, do not be surprised when you get task completion.

Some agencies are built for that. They are happy to receive requests, complete them, and move on. There is nothing wrong with that if that is what you bought.

But if you hired an agency for strategic input, do not accidentally push them into the service provider trap.

Bring them into the right conversations. Not every meeting. Not every Slack channel. Not every internal debate. But enough that they understand where the business is heading and where their work fits.

Agencies can feel when they are being pushed to the side. They know when they are not part of the decision making anymore. Once that happens, it is easy for them to disengage. Not always deliberately, but naturally.

The work gets smaller. The ideas get safer. The relationship gets more transactional.

That is bad for everyone.

The agency business model shapes the relationship

One reason agency relationships drift is that the agency model itself creates pressure.

Paddy sees this constantly now that he works with agency founders. Agencies at different stages often have the same problems, just at different times. They need enough work, enough people to do the work, and a high enough standard across that work.

That never really goes away.

Agencies have multiple clients. They have margin pressure. They need to train juniors. They need senior people on the right work. Over time, it is normal for an account to have more junior people involved than it did at the start.

That is not always a bad sign. It is how the model works.

But in-house teams should understand it. If you want senior thinking, you need to create the conditions for it. Use senior agency people where they matter most: decisions, reviews, planning, risk, difficult tradeoffs. Do not burn their time on meetings where nothing is decided.

And be honest about where the agency is best. If you ask them to do work outside their sweet spot, they may still say yes. Agencies are not always great at saying no, especially when retention and revenue are involved.

So ask: is this really what you are best at?

That question can save both sides trouble.

The SearchPilot comparison

One interesting tangent in our conversation was how different SearchPilot feels from the agency world, even though our roots are in consulting.

At SearchPilot, we do provide a lot of hands-on support. Our professional services team is comfortable with enterprise contracts, custom paperwork, InfoSec, legal reviews, and the relationship work that comes with large customers. That comfort came from the agency world.

But the business is different because the service is anchored in software.

In an agency, the thing you provide can stretch in many directions. Clients ask for more. You are good at one thing, so they ask you to help with another. It is hard to say no.

SearchPilot has more focus. The product does what it does. We can add features and improve the service around it, but the center of gravity is clearer: controlled SEO experiments that isolate impact and show what worked. That connects closely to the positioning we use internally: SearchPilot helps teams move from guessing to knowing by making SEO testable through controlled experiments.

That focus makes some things easier. It also reminded me of how hard focus is in an agency.

If I could go back and advise my agency-founder self, I would probably tell him to be braver about focus earlier.

I am not sure he would have listened.

Put search in control mode with SearchPilot

There is a natural link between this conversation and the way we think about search at SearchPilot.

A good agency relationship should help an in-house team make better decisions. But the strongest decisions still need evidence. Opinion helps. Experience helps. Pattern recognition helps. At some point, though, you need to know whether the change worked.

That is where testing changes the conversation.

Search is often the biggest channel and the least understood. SearchPilot makes SEO and GEO testable, so teams can move from guessing to knowing. We run controlled experiments across category pages, product detail pages, navigation, content, internal linking, and other high value site sections, then show the impact with timelines and confidence.

For in-house teams working with agencies, that can also improve the relationship. The agency can bring ideas. The internal team can bring context. SearchPilot can help prove which ideas deserve rollout.

That is a healthier way to work. Less theatre. Less recycled reporting. More learning.

Stop trying to predict the future. Experiment to discover it. If you want tailored test ideas for your top PLPs and PDPs, schedule a demo and we will share a starter list and a clear path from validation to velocity to control.