The Old Affiliate Website Back to Life Using AI Project - Week 7

Hello there and happy Saturday WAers.
So, it is still hot here in Texas and I am starting to wonder if we have not only moved to another country, but somehow slipped Marvel multiverse style into an alternate reality where cold weather is an unknown entity. Anyway...

For anyone new to these updates, the challenge is simple: I am using AI help me bring a neglected affiliate website back to life? The target is 100 Starter-to-Premium conversions by the end of the year.

Seven weeks in, Google is sending a little more traffic. The business result is still stubbornly unchanged. More Search Clicks, no new Wealthy Affiliate Starter referrals. No Premium upgrades.

This is not a complaint or even a whinge really, just more of a reality check that just because we now have AI to help us, there are maybe no instant fixes. I will say however that I will be having a conversation with it today as I am wondering if we have gone away from the old SEO things I have had work in the past:

Good solid content which provides solutions and helps people - we have been fixing the technical side, but is this something that is missing?

Anyway, will answer that question for you next week. In the meantime - here is the update from the desk of Mr Chat Codex...

A small improvement in search

The latest complete seven-day Search Console comparison shows:

That is 13 more clicks. Encouraging, particularly after spending so much time cleaning up the archive, but still a small sample.

The Taking Action Online review brought in four clicks. The ClickFunnels affiliate-program review and Coffee MLM guide brought in two each.

I cannot credit those clicks to the recent changes yet. This reporting window includes days before and just after the late-August edits. The 27 SEO and accuracy experiments need their full observation period before I start drawing conclusions.

GA4 also recorded 44 Organic Search sessions, compared with 33 in last week's snapshot. Overall sessions were fairly flat at 144 versus 148. These are provisional snapshots, including the reporting day, rather than two completed calendar weeks.

Given this site's history of automated traffic, I am keeping the emphasis on search activity and what people do next. A bigger session total on its own would tell me very little.

The next step is still the weak point

The measured journey recorded five clicks from content into my WA review: three from comparison articles and two from MLM content.

That is the same total as last week's report.

There were no processed tracked WA outbound-click events in the new pull. That does not prove nobody visited WA through any route, but it means this measurement is not showing progress beyond the review.

One practical fix went live on 31 August. On a mobile screen, the review's primary WA button had been sitting below a long introduction. I moved it into the first screen so readers can see the next step immediately. I also tightened the page's tracking so ordinary citations to WA's official information do not get counted as affiliate clicks.

Both fixes make sense. Neither has demonstrated a conversion improvement yet.

The latest WA outcome is straightforward: I have had no new Starters or upgrades since the previous dashboard check either.

Finishing the archive cleanup

Some of this appeared in the follow-up to last week's report, so I am treating it as the completion of that work rather than a fresh set of wins.

All 100 original MLM company reviews are now accounted for. Seventy-four remain live with refreshed content, and 26 have been consolidated through verified permanent redirects into relevant pages. Backlink evidence was checked before those consolidation decisions.

The completed Ahrefs recrawl moved the Health Score from 96 to 98, with reported errors falling from 105 to 36. Follow-up work removed a genuine broken internal link and cleared the consolidated URLs from stale sitemaps.

Useful maintenance, but the reason for doing it is reader trust. Someone arriving at an old review should get a current answer and a sensible next step, rather than stale company information or a dead link.

The free toolkit needs readers

At the authenticated MailerLite check on 31 August, Build Something Real had one form visitor, zero subscribers and zero completed delivery runs.

One visit is nowhere near enough evidence to judge the offer.

I added two relevant entry points: one from the homepage and one from the article explaining the rebuild. Those routes are now measured, and I am leaving the test alone long enough to gather a sample.

This week's analytics includes email-success signals, but I have not verified them as new external subscribers in MailerLite. They are not being counted as leads in this report.

What I am watching next

The late-August search experiments reach their first 14-day checkpoints around 12–13 September. The toolkit distribution test reaches its checkpoint on 14 September.

I want to see whether more relevant searches produce clicks, whether those readers continue through the WA review, and whether the toolkit earns confirmed external subscribers.

There is some encouraging movement in search this week. The conversion result remains zero. That is where Week 7 stands.

Anyway, that is it, off to do a bit of non -AI research.

Will let you all know next week how that went.

Until then

Have fun

Paul