Growing TheAverageGuy.tv with AI: What the Data Is Teaching Me

August 2026 — Establishing the Baseline

I have been using AI to take a much closer look at TheAverageGuy.tv — not simply to create more content, but to understand what more than a decade of publishing has already produced.

August became the point where that work started turning into measurable findings.

Experiment started: August 2026
Latest update: August 2026
Current phase: Baseline established; optimization experiments underway

What I Learned First

The First Number That Changed My Thinking

A 75% drop in clicks looks terrible. If I had stopped there, the response would have been obvious: get more clicks.

But the rest of the data did not behave the same way. Shipped-product value declined only about 4%, and earnings declined roughly 12%.

That does not mean fewer clicks are better. It means the story is more complicated than traffic volume alone, and it pushed better page-level attribution higher on the list.

Affiliate performance, June to August

Affiliate clicks: approximately −75%
Shipped-product value: approximately −4%
Affiliate earnings: approximately −12%

Search Intent Told a Similar Story

Broader informational searches in the sample produced a click-through rate of about 0.57%. Comparison-oriented searches produced about 1.09%.

People searching for BLUETTI versus EcoFlow are not simply looking for a definition. They are closer to making a decision.

That matters because The Average Guy is strongest when I can bring real experience into that decision instead of simply repeating a specification sheet.

Search CTR by intent

Informational queries: approximately 0.57%
Comparison queries: approximately 1.09%
Relative difference: about 1.9× higher CTR for comparison searches

The Archive Became Part of the Growth Strategy

On the Home Gadget Geeks side, we identified 517 episodes, 226 guests and 88 recurring topics.

I have always known the archive was large. Knowing that and being able to use it are two different things.

An old review may still rank. A podcast discussion from years ago may contain an explanation that is still useful. Several pieces around the same subject may belong together in a larger Know IT Guyde.

AI has been useful because it lets me look across more of that history at once. It does not replace what I know about the site. It helps me find the parts I no longer remember well enough to use.

One Thing Changed More Than the Numbers

The biggest shift may be how I think about the next hour of work.

For years, publishing something new almost automatically felt productive. Now a new article has to compete with another possibility: would that hour be better spent improving something that already has evidence behind it?

Sometimes the answer will still be new content. Sometimes it may be updating an old page. Sometimes the right answer may be leaving a successful page alone.

AI makes it incredibly easy to generate activity. Activity is not the same thing as improvement.

What I Am Doing Because of August

September moves the experiment from understanding into testing. I am putting more attention on existing pages that already have search demand, improving attribution, connecting related content more deliberately and continuing the Know IT Guydes as a way to organize larger technical subjects.

When we make an important change, I want the baseline recorded before we touch it. That gives future months something August did not have: results.

What I Am Not Claiming Yet

I cannot say AI has increased traffic. I cannot say the SEO changes that started in September worked. I cannot say comparison content will always outperform informational content. And I cannot yet explain why affiliate clicks declined so much faster than downstream activity.

Those are open questions. That is why this page will continue.

August 2026 Baseline

Content: 1,455 WordPress posts scanned; 542-item working catalog; 517 HGG episodes; 226 guests; 88 topics.

Affiliate footprint: 274 pages containing affiliate links; 873 affiliate links; June-to-August clicks approximately −75%; shipped-product value approximately −4%; affiliate earnings approximately −12%.

Search intent: broad informational CTR approximately 0.57%; comparison-query CTR approximately 1.09%; comparison-query CTR about 1.9× higher.

Existing-content test case: BLUETTI vs. EcoFlow — 8,849 impressions; 75 clicks; approximately 0.85% CTR; average Google position around 4.6.

August status: Baseline established.

Why I Am Keeping This Running Record

A lot of useful work disappears into conversations. I can spend hours working something out with ChatGPT, OpenClaw or another tool, make a decision and then come back months later without all of that context immediately available. The same is true of my own memory.

So I want the important parts preserved somewhere more durable.

This page is partly for readers, partly for me, and partly for whatever AI system I am using six months from now.

If something works later, I want to be able to trace it back to the baseline. If something fails, I want that failure preserved too.

That makes this more than a monthly report. It becomes the history of the experiment.

August is where that history starts.