#51 | My AI Content Audit: 80% Had to Go

TL;DR: Running a blog or content site that’s been quietly declining? An AI-powered audit system I used after a year of neglect—what to keep, fix, or delete.

👋 Hello content creators,

Four months.

That’s how long I did almost no work on my website projects after my daughter was born.

Before that? Over a year of watching my sites slowly die while telling myself it was temporary. It’s just another algorithm update. Things would bounce back.

They didn’t.

Maybe you run a blog too, where rankings keep slipping. Or your content site, which once drew steady traffic, barely registers now. Perhaps you feel frustration in a similar way to me.

But, hang on, I don’t create content, I hear you say. Well, stick around—today’s framework applies to any mess you’ve ignored, any system or strategy that stopped working while you hoped it would fix itself.

Here’s what I learned during that year of neglect, and the four months of complete absence: much of my current content was actively destroying my authority.

Google’s Helpful Content updates continued to roll out throughout that year. Each one confused me more. Traffic dropped after every change. Nothing made sense anymore, so stepping away felt easier than figuring out what to do.

It turns out that neglect plus algorithm chaos equals disaster. Also not great.

When I finally came back, I needed to see clearly before I could fix anything. So I built a simple AI-powered system to audit everything systematically (link below).

80% of my content needed to either improve dramatically or disappear altogether.

Yep! Quite a bit of work to do.

And if you’re in the same boat as me, this approach can help when you’re facing similar challenges—algorithm confusion, declining traffic, or paralysis about what to fix first. Keep reading 👇.

The AI Learning Guy newsletter 🤖 🧠💡

AI learning hacks and mega prompts delivered to your inbox.

The slow-motion disaster

Coming back felt like opening mail you’ve been avoiding for months.

My business stats and data didn’t look great.

And, the algorithms didn’t pause either. Google kept updating. Competitors kept publishing. The ecosystem shifted around my static, neglected sites.

But the damage started long before I left.

Those Helpful Content updates had been landing for over a year. Each one made me more uncertain about what Google actually wanted. I’d read many guides and followed the advice. Traffic dropped anyway.

The confusion created paralysis.

If I couldn’t figure out what Google wanted, how could I fix anything? So I published less, updated nothing, and hoped things would stabilize on their own.

Classic avoidance wearing a patience costume.

On top of that, the four-month absence just accelerated what was already happening.

What the data actually showed

Running a complete content audit revealed categories that hurt to see.

I had old affiliate reviews that I hadn’t touched in two years; half of the online courses were obsolete. Off-niche posts that initially made sense are now confusing Google about my actual goals, purposes, and expertise.

The content still working?

Maybe 20% of everything I’d ever published.
Those pieces had depth, a tight niche fit, answered real questions, and stayed current.

Everything else was kinda pulling my site down.

Actively damaging because Google was using all that thin, scattered, outdated content to decide my site lacked niche focus and expertise.

The algorithm identified my problem long before I acknowledged its existence.

An AI-powered system

Staring at hundreds of URLs in a spreadsheet, I needed a decision system simple enough to actually use.

I asked AI to help me build it.

Two axes solved everything: Performance on one side (traffic, rankings, revenue), Relevance on the other (niche fit, topical authority).

Every piece of content landed in one of four clear quadrants.

1. High Performance + High Relevance: Your winners. Protect them, optimize them aggressively, and build more content following their patterns.

2. High Performance + Low Relevance: Interesting anomalies. They work somehow, but pull your site’s focus. Worth careful evaluation.

3. Low Performance + High Relevance: Hidden opportunities sitting on your site. They fit your expertise perfectly, but aren’t getting found. Candidates for expansion.

4. Low Performance + Low Relevance: Dead weight. These dilute your authority, confuse Google about your expertise, and provide zero value. Time to prune.

Nothing sophisticated. But systematic thinking was exactly what broke through over a year of paralysis and confusion.

The risk management layer

Making changes scared me more than the declining traffic did.

What if I deleted something that I needed to recover? What if I killed a page that secretly drove conversions? What if I made things worse?

One critical constraint changed everything: protect what’s currently working, even when you don’t fully understand why it works.

Identify the low-risk pages first, or those that might benefit from updates or changes, so-called low-hanging fruit.

This constraint transformed fear into confidence.

I started with obviously dead content—pages showing zero traffic for six months, clearly off-niche, thoroughly outdated, and no valuable inbound links.

Migrate off-topic content that lacks traffic and monetization to different websites, allowing for a tighter niche focus. Redirect URLs for SEO signals. Archive old content. Remove irrelevant content.

Things remained stable. Phew!

That experience gave me the confidence to make the tough decisions.

The process that worked

What actually moved things forward looked completely unexciting.

Export every URL from my sitemap, combine its traffic data and ranking positions, and include any revenue information I have. Some data is generated with AI, while other data is added manually.

Score each piece for niche relevance by asking one question: Does this strengthen my authority in my core topic, or pull focus away?

Map everything into the performance versus relevance matrix using a simple spreadsheet AI helped me build.

Start with easy wins: prune obvious junk, redirect dead URLs, fix glaring technical errors that showed up immediately.

Move to harder decisions: expand content with clear potential, consolidate pieces covering the same ground, and update anything salvageable with current information.

Monitor everything weekly: watch what happens, track changes, and adjust based on actual results instead of assumptions.

Tip: Work in batches, one content cluster at a time, tracking impact before moving forward. Take your time.

While still working on the project, Google seems to have stopped penalizing my site for lack of focus. I can see some previous rankings coming in again.

The one question

Stop hoping individual pieces would magically start performing better. Stop assuming that more content automatically solves problems.

Start asking one question about every piece: Does this strengthen my authority or dilute it?

That question made everything obvious.

Half of my content diluted my authority by pulling my site in too many directions.

Removing or migrating that content to new websites initially felt terrifying and overwhelming. But keeping it felt worse once I understood the real cost.

Your starting point

You don’t need months to start figuring this out. You need a few hours and an honest assessment.

Question 1: Pull up your top 20 pages by traffic, then separately list your top 20 pages by relevance to your core niche. How much overlap exists? If there’s a significant gap, you’ve found your core problem.

Question 2: Identify three pieces of content you know are genuinely dead weight. Check the data carefully. If they’re truly dead—no traffic, no links, no relevance—remove them or redirect them. Watch what happens. Probably nothing catastrophic.

Question 3: Find one underperforming piece that fits your niche perfectly. Spend two focused hours making it genuinely better—add real depth, answer more questions, provide actual value. Track what happens over the next month.

Start small. Build confidence through results. Keep moving forward.

The AI assessment tool

I built a simple and free Content Health Assessment [LINK] that systematically walks through this exact process, using AI to speed up the analysis.

Performance versus Relevance matrix, decision trees for each quadrant, risk assessment guidelines you can actually follow, and implementation steps that work.

Add this AI-Powered Content Health Assessment PDF to your AI tool of choice.

You can use it to audit your content methodically. Determine what deserves to be kept, what needs fixing, and what should be eliminated.

If you’re staring at declining traffic, wondering where everything went sideways—this is your entry point.

Track what happens. Keep going.

The rest follows naturally from that first clear view of what you actually have.

Mark
The AI Learning Guy
👋⚡😎

The AI Learning Guy newsletter 🤖 🧠💡

AI learning hacks and mega prompts delivered to your inbox.

Interesting Sources

Note: No single website has all the answers. This list serves as a starting point for those who want to explore or satisfy their curiosity about AI.
Links: Links with * are affiliate links. See disclosure below.