Go back

Update 03 - Can SEO Alone Build an Audience in the Age of AI?

Published:  at  08:50 AM

I promised monthly updates, then disappeared for three months like a dad who said he was going out for cigarettes. (I don’t smoke, but you get the idea.)

July through September 2025 Site Progress Update 03

Intro post on why or how this all started: Can SEO Alone Build an Audience in the Age of AI? Let’s Find Out.

Previous updates:

Quick Summary

Table of contents

Open Table of contents

What This Update Covers

Welcome to the “I said I’d do monthly updates and then real life reminded me I’m not a full-time blogger” edition of my SEO experiment.

Here’s what happened: Kids started school in September, which means the entire summer was a chaotic sprint to squeeze in every last bit of family time before the routine kicked back in. Then Q3 rolled into Q4, and if you’ve ever worked in tech, you know Q4 is when everyone suddenly remembers all the things they promised to deliver this year. I don’t know if I can live like this, but here we are.

So instead of three separate monthly updates (which would’ve killed me to write), you’re getting one mega-update covering July through September. Think of it as a quarterly earnings report, except instead of disappointing shareholders, I’m disappointing the three people who actually read these posts. (Hi mom! Hi random person from Croatia!)

But honestly? This three-month gap gave me better data to work with. Monthly fluctuations can be noise, but three months shows actual trends. Silver lining, right?

Key Findings

The Good:

The Frustrating:

The Reality Check: Working full-time while running an SEO experiment is like trying to maintain a sourdough starter while training for a marathon—technically possible but you’re going to neglect something. Spoiler: it’s the blog.

The Numbers: July through September 2025

Google Search Console Performance

Let me hit you with the overall three-month totals first:

Combined July-September:

July-September 2025 Combined Google Search Console Performance

Combined July-September 2025 GSC Performance

Now here’s where it gets interesting. Watch what happens month by month:

July 2025:

July 2025 Google Search Console Performance

July 2025 GSC Performance - Starting from position 26.7

August 2025:

August 2025 Google Search Console Performance

August 2025 GSC Performance - Position improved to 21.4

September 2025:

September 2025 Google Search Console Performance

September 2025 GSC Performance - Cracked into position 15.3!

What This Actually Means:

My average position improved by 11.4 positions from July to September. That’s massive! I went from page 3 obscurity (position 26.7) to almost cracking page 2 (position 15.3). For a site that’s only six months old with zero backlink building strategy, that’s actually not terrible.

But here’s the weird part: my CTR is welded at 0.4%. It didn’t budge. Not even 0.1%. This suggests one of two things:

  1. My titles and meta descriptions are boring people to death
  2. Google’s showing me in SERP features that people aren’t clicking (hello, AI Overviews)

I’m betting on number 2, but number 1 is definitely possible given my tendency to overthink titles until they become committee-approved corporate nonsense.

Traffic & Analytics: The Great Analytics Hoarding Experiment

Remember how in Update 02 I was still trying to pick an analytics platform? Well, I solved that problem by… adding all of them. I’m like a data hoarder but instead of cats, I’m collecting analytics dashboards.

Current analytics stack:

Why three platforms? “Different platforms show different things and I’m curious” is my official excuse, but the real answer is I’m an analytics hoarder and I probably have a problem. Also, I wanted to fact-check Google any way I can, which is probably the pettiest reason to add analytics tools but here we are.

The irony: I now have so many analytics tools tracking my site that I’m genuinely worried they’re slowing it down. Remember those perfect Lighthouse scores I bragged about in Update 01? Yeah, those might be in danger.

Traffic by Month (Google Analytics 4):

July 2025:

August 2025:

September 2025:

Combined Totals:

Top Pages by Traffic - Google Analytics 4

Google Analytics 4: Top pages by sessions July-September 2025

The Pageview Mystery:

Did you catch that? Users went UP in August but pageviews went DOWN. That means people are visiting my site but reading less per session. Either my content is so concise they’re finding what they need immediately (optimistic interpretation), or they’re bouncing after realizing this isn’t what they wanted (realistic interpretation), or Google Analytics is drunk (most likely interpretation).

This is why I added PostHog—I need a tiebreaker to figure out what’s actually happening. Although adding a third analytics platform to solve analytics confusion might be the definition of insanity.

What Vercel Shows vs Google Analytics:

Vercel counts EVERYTHING that hits the server, including all the bots, crawlers, and whatever mysterious entities are constantly pinging websites. Google Analytics is supposedly smarter and filters out bot traffic.

The result? Vercel thinks I’m way more popular than I actually am. It’s like asking your mom how many people came to your birthday party versus asking the kid who actually counted.

Going forward, I’ll trust Google Analytics for user metrics and PostHog for session quality and events. Vercel can keep doing its thing in the corner, making me feel better about my traffic numbers.

Top Performing Pages: The Great Rank Swap

Here are the top 10 pages by clicks for July-September combined:

Top Performing Pages July-September 2025

Top 10 pages by clicks - htaccess takes the crown
  1. How To Use htaccess to Password Protect Your Website - 112 clicks, 9,743 impressions, 1.1% CTR, Position 19.0
  2. How to Be Cool: A Dad’s Guide - 82 clicks, 26,900 impressions, 0.3% CTR, Position 9.0
  3. Line Breaks in Markdown - 53 clicks, 15,446 impressions, 0.3% CTR, Position 12.2
  4. Building with AI Systems - 50 clicks, 4,271 impressions, 1.2% CTR, Position 18.9
  5. Learning To Fly A Helicopter - 40 clicks, 9,254 impressions, 0.4% CTR, Position 17.8
  6. Will AI Replace Software Engineers? - 19 clicks, 5,653 impressions, 0.3% CTR, Position 18.8
  7. Spending Quality Time With Family - 9 clicks, 10,030 impressions, 0.1% CTR, Position 12.8
  8. Why Does the Trailing-Slash Matter? - 7 clicks, 13,648 impressions, 0.1% CTR, Position 49.5
  9. Why AI Tools Are Making Software Engineers Less Capable - 4 clicks, 349 impressions, 1.1% CTR, Position 10.2
  10. Build Yourself to Fit Your Real Life - 3 clicks, 941 impressions, 0.3% CTR, Position 9.1

The Big Story: htaccess Overthrows “How to Be Cool”

In Update 02, “How to Be Cool” was my undisputed champion—the post that desperate teenagers and confused adults kept finding when googling existential questions about coolness. It was getting 82 clicks across three months and ranking at position 9.0.

But the king is dead. Long live the king.

The htaccess password protection post climbed to 112 clicks and claimed the throne. Why? Because I updated it in early October with current Apache syntax and better explanations. Google noticed, rankings improved, and suddenly everyone who needs to password protect a directory is finding my post instead of StackOverflow threads from 2012.

The lesson: Updating old content works. Who knew? (Everyone. Everyone knew this. But I proved it for myself, which counts.)

The Tragedy:

“Why Does the Trailing-Slash Matter?” has 13,648 impressions but only 7 clicks. That’s a 0.05% CTR. Google is showing this page to thousands of people who are apparently saying “nah, I’ll stay confused about trailing slashes, thanks.”

Either my title is boring, or people just want to complain about trailing slashes without actually learning about them. Probably both.

SEMrush Data: Keywords & SERP Features

Ranking Keywords Growth:

July 2025 SEMrush Organic Keywords

July 2025: 115 ranking keywords

August 2025 SEMrush Organic Keywords

August 2025: 138 ranking keywords (+20%)

September 2025 SEMrush Organic Keywords

September 2025: 153 ranking keywords (+10.9%)

Total growth: +33% keyword expansion over three months. That’s actually solid for a site this young.

The Domain Rank Filter Situation:

Full disclosure: some of my SEMrush screenshots have a “domain ranks” filter applied, which is why the numbers look different between images. I didn’t realize I had it on until I was pulling data for this update. The filtered numbers show fewer keywords (12 in July, 5 in August/September) but those are the ones where my domain specifically ranks for the term.

The larger numbers (115, 138, 153) are total keywords I’m appearing for in any capacity. Both matter, but I’m tracking the total keyword count going forward for consistency.

AI Overviews & People Also Asked:

MonthAI OverviewsPeople Also Asked
July1217
August57
September510

July 2025 AI Overviews

July 2025: 12 keywords appearing in AI Overviews

August 2025 AI Overviews

August 2025: 5 keywords in AI Overviews

September 2025 AI Overviews

September 2025: 5 keywords in AI Overviews

Translation: Google’s AI is stealing my content for overviews on 12-5 different queries, and “People Also Asked” is showing my content in 7-17 question boxes.

Is this good or bad? Honestly, I don’t know. On one hand, appearing in AI Overviews means Google thinks my content is authoritative enough to train their AI. On the other hand, people are getting their answers without clicking through to my site, which explains my CTR being stuck at 0.4%.

It’s like being quoted in a newspaper article but the article just uses your best lines without mentioning where readers can find more of your work. Flattering but ultimately not driving traffic.

Content Published: July through September

After the pathetic June showing where I barely published anything, July came through with actual content. Then August and September? Well, we don’t talk about those.

July 2025 (6 posts):

August 2025 (1 post):

September 2025 (0 posts):

Updated Posts (October):

Total: 7 posts across three months (6 in July, 1 in August, 0 in September) + 2 major updates in October

Let’s be real: July carried this entire quarter. I front-loaded all my content because I could feel Q3/Q4 work hell approaching like storm clouds on the horizon.

Content Strategy Observations:

July was my everything month—six posts covering AI topics, personal development, technical tutorials, and tech news roundups. These were meatier posts with real technical depth, not just hot takes. I was on a roll.

August managed one Tech Chaos post before work consumed my life.

September? September is what happens when you work full-time and have kids starting school. Content creation becomes a pleasant fantasy you remember from simpler times.

The Line Breaks in Markdown post is fascinating because it’s such a small, specific problem (how to create single line breaks instead of paragraph breaks), but apparently enough people search for this that it’s already in my top 3 pages by clicks. Sometimes the smallest problems have the most search volume.

Remember in Update 01 and 02 when I said “all I have is spam backlinks”? Well, something actually happened!

New Backlinks:

  1. inc.com (Domain Authority: 60)
  2. substack.com (Domain Authority: 82)

The inc.com Story:

Someone at inc.com wrote an article titled “Software Developers Are (Literally) Losing Their Minds to AI: Our most gifted technical minds are atrophying. This is how AI dystopia happens. Here’s how we prevent it.”

In the article, he quoted my post “Why AI Tools Are Making Software Engineers Less Capable” with this introduction:

“Or, I don’t even know this guy, but his cogent argument on Why AI Tools Are Making Software Engineers Less Capable.”

“I don’t even know this guy” might be the most backhanded introduction I’ve ever received, but YOU KNOW WHAT? I’ll take it. He called my argument “cogent” and linked to my site from a Domain Authority 60 publication. That’s a win in my book.

What This Means:

Someone found my content valuable enough to cite in their own work WITHOUT me reaching out, begging for links, or doing any outreach whatsoever. This is exactly what I’m testing—can SEO alone build an audience? Can you create content good enough that people naturally link to it?

Apparently yes, but they might introduce you as “I don’t even know this guy.”

The Substack Twist:

Plot twist: the Substack backlink is from the SAME GUY who wrote the inc.com article. He writes for inc.com and has his own Substack newsletter. So I didn’t get two separate people to notice my work—I got one person who liked my content enough to cite it twice.

I’m not sure if this counts as two backlinks or one very enthusiastic backlink, but I’ll take it either way.

This raises an interesting question: Am I missing something by not having a Substack? It’s got a Domain Authority of 82. If I cross-posted content there and used canonical tags pointing back to kentgigger.com as the original source, would Google index it faster? Would the Substack audience discover my main blog?

Maybe that’s something to test in a future update. What do you all think? Should I join the Substack club, or is that just adding another platform to manage when I already can barely keep up with one blog?

Link Building Progress:

So after three updates of saying “nothing to report here, just spam,” I can finally say: I have two legitimate backlinks from high-authority domains. It’s not a flood of links, but it’s proof that creating quality content can attract organic backlinks.

Will this move the needle on rankings? Maybe. Google claims backlinks are still a ranking factor, but who knows what they actually prioritize anymore. I’ll report back in Update 04.

The Great Google Indexing Saga Continues

Google’s still playing games with my site, and I’m starting to think it’s personal.

Current Indexing Issues:

1. Alternate page with proper canonical tag (15 pages affected)

This is my favorite type of Google nonsense. These pages aren’t indexed because Google claims they have “alternate pages with proper canonical tags”—but here’s the thing: I REDIRECTED ALL THE TRAILING SLASH VERSIONS TO THE NON-TRAILING SLASH VERSIONS.

Google sees:

My site redirects the trailing slash version to the non-trailing slash version. The canonical tag points to the non-trailing slash version. Everything is configured correctly.

But Google’s like “hmm, I see two versions of this page” and then refuses to index either one properly. Or indexes one and marks the other as “alternate.” It’s lazy. Google handled http to https just fine—why can’t we come to a consensus about trailing slashes?

This has been happening since April (first detected 4/19/25) and Google just… doesn’t care.

Alternate page with proper canonical tag error

Google claiming alternate canonical tag issues - even though everything is properly redirected

2. Excluded by ‘noindex’ tag (5 pages affected)

Five pages are being excluded because Google thinks they have noindex tags. Spoiler: they don’t. Or rather, they DID have redirects that maybe confused Google at some point, and now Google won’t re-crawl them to realize the issue is fixed.

Yes, I know you can request reindexing through Google Search Console. I’ve done it. Multiple times. It’s a back-and-forth chess game where I request reindexing, Google pretends to look at the page, and then says “nah, still has noindex” even though it demonstrably doesn’t. Google plays stupid, or at least that’s what I’m saying.

The affected pages include:

First detected May 6, 2025. Still not fixed. Google’s taking its sweet time.

Pages excluded by noindex tag

Google insisting pages have noindex tags that don't exist

3. Crawled - currently not indexed (4 pages affected)

These are the most frustrating. Google crawled these pages. Google looked at them. Google said “eh, maybe later” and walked away.

One example: AI Is the Death of Outbound

This post was published in August. It’s been crawled. The content is good (I think?). But Google’s just… not indexing it. Why? No explanation. No error. Just vibes-based indexing.

This is the “maybe later” pile I mentioned in Update 01. Except it’s been “maybe later” for months now, and I’m starting to think “maybe later” actually means “probably never.”

First detected: May 10, 2025. Still waiting.

Crawled but currently not indexed

The "maybe later" pile - Google crawled these but won't index them

What I’ve Learned:

Google conserves its energy for sites it considers valuable. If you’re a new site without much authority, Google will crawl you occasionally but won’t bother fully indexing everything unless it thinks there’s something worth indexing.

This is why building authority through backlinks, social signals, and consistent content matters. Google needs proof you’re worth the indexing effort.

The ironic part? The inc.com backlink might actually convince Google to re-crawl and index more of my content. We’ll see.

Technical Issues:

Part of this is also on Astro. The framework I’m using has some quirks with trailing slashes that I need to figure out. It’s on my never-ending list of things to fix when I have time (so, never).

The Analytics Hoarding Problem

Let me be honest about something: I now have three analytics platforms running on this site, and I’m not entirely sure why.

The Stack:

  1. Vercel Analytics (included free with hosting)
  2. Google Analytics 4 (added July 2 because everyone uses it)
  3. PostHog (added October 23 because I heard it’s good for session tracking)

My Official Excuse:

“Different platforms show different things and I’m curious about comparing data quality.”

The Real Reason:

I’m an analytics hoarder and I’m probably slowing down my own site with all these tracking scripts. Also, I don’t trust Google (see: everything in this post about Google’s indexing nonsense), so I wanted independent verification of traffic numbers.

What Each Platform Actually Shows:

Vercel: Total server hits including bots. Makes me feel popular. Probably lying.

Google Analytics: Filtered user data excluding bots. More realistic. Might also be lying but in a more sophisticated way.

PostHog: Session quality, events, user behavior. Just added it so I don’t have enough data yet, but it looks pretty and has A/B testing capabilities I’ll never use.

The Question:

Do I need three analytics platforms? Absolutely not.

Will I remove any of them? Also no.

Am I the problem? Probably.

Performance Impact:

Here’s my concern: each analytics platform adds JavaScript to my pages. More JavaScript = slower load times = worse Core Web Vitals = Google ranks me lower.

Remember those perfect Lighthouse scores from Update 01? I should probably check those again. But I won’t, because ignorance is bliss and I have enough problems already.

Going Forward:

I’ll report primary metrics from Google Analytics (since it’s the industry standard) and use PostHog for deeper session analysis once I have meaningful data. Vercel can keep doing its thing as my ego-boost platform.

If my Lighthouse scores tank, I’ll revisit this decision. Until then, I’m collecting analytics dashboards like Pokémon cards.

The LinkedIn Reality Check

In Update 02, I mentioned I might start posting on LinkedIn because relying entirely on SEO felt like betting everything on Google’s mood.

So I did it. I posted three “Tech Chaos Chronicles” style posts on LinkedIn—the same content that became my monthly tech roundups.

The Results:

People liked the posts. They got decent engagement. Comments, likes, shares—all the LinkedIn metrics looked good.

Traffic to my blog: Basically zero.

Maybe 2 people total clicked through to read the full posts on my site.

What I Learned:

Nobody leaves LinkedIn. People consume content where they find it. If your post is interesting, they’ll engage with it on LinkedIn. But asking them to click a link and read more on an external site? That’s apparently too much effort.

This isn’t unique to LinkedIn—it’s true for every platform. Twitter (sorry, X), Facebook, Reddit—users want to stay in the app. They don’t want to be redirected to some random blog.

The Implications for This Experiment:

My original hypothesis was “can SEO alone build an audience?” The answer is looking more like “SEO alone can drive traffic, but social platforms keep traffic trapped in their walled gardens.”

Even when I post on social media, it doesn’t meaningfully impact blog traffic. The traffic I’m getting is almost entirely from organic search.

What This Means Going Forward:

I’ll probably keep posting Tech Chaos roundups on LinkedIn because:

  1. People seem to enjoy them
  2. It takes minimal effort to repurpose content
  3. Maybe someday someone will click through (hope springs eternal)

But I’m not counting on social media to drive traffic. This experiment is still primarily about SEO.

The Honest Assessment:

If I wanted to build an audience quickly, I’d go all-in on LinkedIn or Twitter and forget about the blog. Post native content, engage with comments, build relationships.

But that’s not the experiment. The experiment is: can a blog survive on SEO alone in the AI era?

The answer so far: yes, but slowly. Very, very slowly.

Google’s AI Is Eating My Lunch (But At Least It’s Citing Sources?)

One more thing about those AI Overviews: Google’s showing my content in AI-generated summaries for 5-12 queries per month (depending on how SEMrush measures it).

What this looks like in practice:

Someone searches “how to create line breaks in markdown” and Google’s AI generates a summary that includes information from my post. The user gets their answer without clicking through to my site.

The good:

The bad:

The ugly truth:

This is the future of search. Google wants to answer questions directly in the SERP without users having to click anywhere. Publishers become free training data for AI while Google keeps all the traffic.

It’s like being quoted in a newspaper article where the article just reproduces your entire post and there’s no reason for readers to find you.

And if that wasn’t enough, OpenAI just launched their own browser called Atlas. Another search alternative that will pull content from sites like mine, generate answers, and keep users in their ecosystem. Between Google’s AI Overviews, ChatGPT search, and now Atlas, traditional search is slowly dying.

Will this blog even matter in the future? I honestly don’t know. Maybe I’m documenting the death of blogging in real-time while pretending it’s an SEO experiment. That would be very on-brand for me.

Can I opt out?

Technically yes—I could add meta tags to block AI scraping. But that would also hurt my chances of ranking in regular search, which defeats the entire purpose of this experiment.

So I’m stuck in this weird limbo where Google uses my content but doesn’t send me traffic. Fun!

The silver lining:

At least my CTR staying at 0.4% makes sense now. A significant portion of my impressions are in AI Overviews where clicking through isn’t even an option.

This isn’t a personal failure—it’s just the new reality of SEO in 2025.

Goals for Next Month (I Mean It This Time)

Let’s be realistic about what I can actually accomplish:

Content Goals:

Technical Goals:

Measurement Goals:

Realistic Goals:

What I’m NOT Promising:

I’m not promising to post on LinkedIn regularly. The engagement is nice but it doesn’t drive blog traffic, and I’d rather spend that time writing content for my own site.

I’m not promising to fix all the Google indexing issues. Some of this is just Google being Google, and I can’t force the algorithm to index pages it doesn’t want to index.

I’m not promising to reduce my analytics tools. I have a problem and I’m okay with it.

The Honest Assessment:

This experiment is working, but it’s slow. Really slow. Traffic is growing, rankings are improving, and I even got some legitimate backlinks. But “working” and “successful” are different things.

Can SEO alone build an audience? Yes.

Can it do it quickly? No.

Is it sustainable while working full-time and having a life? Barely.

Will I keep doing it? Yeah, probably. I’ve come this far.

The Bottom Line

After six months of this experiment, here’s what I know:

Google will index you, but slowly Quality content can attract organic backlinks Updating old content works Rankings can improve without backlink building Social media doesn’t drive blog traffic (not for me, anyway) AI Overviews kill CTR Balancing this with full-time work is hard

If you’re reading this and considering starting a blog in 2025, here’s my advice: lower your expectations, commit to the long game, and don’t expect results for at least six months. Maybe a year. Possibly longer.

SEO isn’t dead, but it’s definitely not the fast path to success it might have been in 2015.

Will I make it to Update 04? Ask me in December when I’m drowning in Q4 deliverables and questioning all my life choices.

Until then, thanks for following along on this weird experiment. If you found this useful (or at least entertaining), hit me up on LinkedIn and let me know. I promise I’ll see your message even if you don’t click through to the blog.

Intro post on why or how this all started: Can SEO Alone Build an Audience in the Age of AI? Let’s Find Out.

Previous updates:

Share this post on:
Share this post via WhatsApp Share this post on Facebook Share this post on X Share this post via Telegram Share this post on Pinterest Share this post via email
Previous Post
How to Be Cool: A Dad's Guide
Next Post
CMS Headless: Essential Guide for Modern Content Management