They Already Had an Agency, SEO Content and Automation. It Still Wasn't Working.

They had an agency. SEO work. Automated content. It still wasn't working. Here's what Rocket Room taught me about the difference between marketing activity and an actual operating system.

One of the most dangerous positions a business can be in is not having no marketing.

It is having a lot of marketing activity and assuming that means the underlying system is working.

When we started working with Rocket Room, they were not starting from zero.

They had already invested in external marketing support. SEO work had been done. Content was being published. There was even an automated system writing and pushing articles into the website's CMS.

From the outside, it looked like marketing was happening.

But once we got underneath the surface, the picture was much less clear.

There was legacy content mixed with newer work. There were technical inconsistencies across the site. Tracking was incomplete. The website had been through different versions, tools and publishing workflows. It was difficult to separate what was useful from what was simply adding more noise.

The obvious response would have been to add another content machine.

Publish more articles. Target more keywords. Create more reports. Recommend more things for the team to do.

That would have been the wrong move.

Rocket Room did not need more marketing activity.

It needed someone to understand the whole system, decide what actually mattered and take responsibility for fixing it.

More output does not mean more progress

Automated content systems are easy to build now.

You can connect an LLM to a keyword list, generate articles and send them directly into a CMS. An agency or internal marketer can produce dozens of pieces of content every month without a particularly large team.

The production problem is disappearing.

But producing more content does not automatically create a stronger search foundation.

If the technical structure is inconsistent, more content can make the website harder to manage. If the positioning is unclear, more pages repeat the confusion. If tracking is incomplete, nobody can tell whether the work is improving traffic, engagement or enquiries.

You end up with a machine that produces activity without a reliable way to determine whether the activity is useful.

That was the important distinction at Rocket Room.

The business did not lack people doing things. It lacked a single operating layer connecting diagnosis, implementation and measurement.

We had to understand the system before adding to it

The first stage of the Rocket Room work was not content production.

It was assessment.

We audited the existing website for traditional search visibility and how clearly AI systems could understand it. We looked at the content structure, metadata, schema, performance and analytics setup.

We also had to understand how the website was actually operating.

Which site was live? Which CMS was still publishing? Which content had been created manually? Which articles were coming from the legacy automation? Which technical changes had genuinely reached production?

These sound like basic questions, but this is what happens when responsibility is spread across an agency, a developer, a CMS, automation tools and the internal team.

Everyone can be completing their individual task while nobody has a reliable picture of the whole system.

Chad's role was to create that picture, identify the highest-value problems and then execute directly against the website.

The work included improving schema and metadata, strengthening the technical foundation, refining content for traditional and AI search, implementing GA4 properly and configuring channel tracking.

Once the changes were live, the work shifted into weekly monitoring and optimisation rather than another one-off audit.

That continuity mattered.

The same system that identified the problem was still present when the results arrived.

What changed over the next three months

Between June and September 2026, Rocket Room's sessions increased from 1,278 to 3,155.

That is a 147% increase.

New users rose from 777 to 2,082, an increase of 168%. Engagement increased from 51.8% to 58%.

Organic search became the largest channel, accounting for 1,512 sessions and 48% of total September traffic.

The most interesting result, however, came from ChatGPT.

ChatGPT sessions grew from 14 in July to 100 in September. That is still a relatively small channel compared with Google, but the quality of those visitors was unusually strong.

In September, visitors arriving through ChatGPT stayed for an average of 3 minutes and 52 seconds. Paid social visitors stayed for 1 minute and 4 seconds.

ChatGPT visitors stayed 3.6 times longer.

They also had a 65% engagement rate and a 35% bounce rate, the lowest bounce rate of any channel with at least 100 sessions that month.

This is why I think AI visibility needs to be treated as more than another impression metric.

Someone arriving through an AI recommendation has often already explained their problem, compared their options and received a degree of qualification before visiting the website.

The volume may still be smaller, but the intent can be considerably stronger.

The attribution needs to be honest

It would be very easy to turn these numbers into a claim that Chad generated every additional visitor.

That would not be honest.

Rocket Room's growth came from several channels. Paid social also grew significantly during the period. GA4 tracking only began in April, so we do not have a year-over-year comparison. Booking completion and revenue attribution were not yet implemented.

We can see when visitors start an enquiry form, but we cannot yet say how many became completed bookings or how much revenue each channel generated.

What we can say is that the website's technical and measurement foundation improved during a period of sustained growth.

We can also say that organic search more than doubled, ChatGPT traffic grew by more than 600%, and AI-referred visitors displayed significantly stronger engagement than paid social traffic.

Those results justify continuing the work and improving the attribution layer.

They do not justify pretending that one tool caused every positive movement across the business.

That distinction matters to me because marketing is full of people claiming responsibility for outcomes they cannot actually prove.

If AI marketing is going to replace parts of the agency model, it needs to be more accountable, not less.

The agency wasn't necessarily the problem

I do not think the useful conclusion is that Rocket Room's previous agency was incompetent.

Agencies can produce good work. Automated content systems can produce useful content. Developers can implement strong technical changes.

The problem is the structure around them.

An agency usually owns a defined scope. A content system owns production. A developer owns implementation. Analytics records what happened after the fact.

The client is still left coordinating the handoffs and deciding what matters.

That works when the client has a strong internal marketing function.

It works much less effectively when the founder or operator is already managing the rest of the business and hired outside support precisely because they did not want to become the marketing manager.

Rocket Room showed me that the next generation of marketing software cannot simply produce more recommendations.

It needs to take responsibility for the operating loop:

• Understand the business.
• Inspect what already exists.
• Determine the real constraint.
• Choose the highest-value intervention.
• Execute the work.
• Measure what changed.
• Remember why the decision was made.
• Then decide what happens next.

The value is not just intelligence at each individual step. The value is continuity across all of them.

The real problem was the missing operating system

The strongest lesson from Rocket Room was simple:

The problem wasn't a lack of marketing activity. It was a lack of an operating system deciding what should actually be done.

There was an agency.

There was SEO activity.

There was automated content.

There were developers and marketing tools.

But none of those things automatically created ownership of the full distribution system.

This is the gap we are trying to close with Chad.

Not by producing another dashboard or another monthly report. Not by filling the website with AI-generated articles. And not by giving the founder a list of fifty things they should now manage themselves.

The goal is to give the business one marketing teammate that can understand the existing system, identify what matters, execute the work and show the evidence afterward.

Rocket Room is still an early case study. The next phase is completing booking and revenue attribution so we can connect visibility improvements to actual commercial outcomes.

But the first three months already showed something important.

More content was not the answer.

More activity was not the answer.

The answer was creating a system that could decide, act and learn.

Sources

• Alexander Shoucair, Chad — Rocket Room case study data, June–September 2026. Primary source.