DEV Community: indi-seo

The Ad Was Never the Problem

indi-seo — Fri, 25 Sep 2026 17:43:07 +0000

Picture two ads running side by side. One has a beautiful hero image, tight copy, a designer clearly spent hours on it. The other looks like it was made in five minutes — slightly off font, a stock photo that's seen better days. The ugly one wins. Not by a little. By three times the return.

This happens more often than anyone in advertising likes to admit, and once you understand why, you start seeing the real game being played underneath every campaign that succeeds or quietly bleeds money.

The ad is not the offer

An ad has one job: interrupt someone for a second and convince them the interruption was worth it. That's it. What happens after the click is a completely different contest, and it's the one that actually decides whether the money was well spent.

If the landing page asks for more commitment than the ad implied, if the price shows up as a surprise, if the whole experience feels like a bait and switch even a small one, the click was already wasted the moment it landed. No amount of clever copy upstream fixes a mismatch downstream. This is why "make a better ad" is often the wrong fix for a campaign that's underperforming. The ad did its job. Something after it didn't.

Nobody does the math before they panic

Here's a number almost nobody calculates before hitting launch: if a customer is worth $150 and each click costs $4, roughly 1 in 37 people need to convert just to break even. Not to profit. To break even.

Most campaigns go live without this number ever getting written down anywhere. Three days later, someone's staring at a dashboard convinced the campaign is failing, with no actual definition of what failing means for their specific business. Panic sets in, budgets get pulled, and the real lesson, that nobody set a target before spending money, never gets learned. The campaign gets blamed instead of the math that was never done.

This single missing step is probably responsible for more prematurely killed campaigns than bad creative, wrong platforms, and weak targeting combined.

Precision is not the same thing as relevance

Every ad platform on earth will let you target with terrifying accuracy. Job title, income bracket, interests, behaviors, down to a scarily specific level. It feels powerful. It's also frequently a trap.

Think about two people who share the exact same job title: Marketing Director. One runs marketing solo at a twelve-person startup, answering to a founder who wants results yesterday. The other manages a team of eight inside a four-thousand-person company, buried in approval chains and quarterly planning decks. Same title. Same targeting filter. Completely different Tuesday. An ad written for one will feel tone-deaf to the other, and most campaigns never account for the gap.

The sharper question isn't "who has this job title." It's "what does this specific person's actual day look like, and where does this offer fit into it." That question doesn't show up in any targeting menu. It has to be answered before anyone opens the ad platform at all.

The retargeting trap that quietly drains budgets

Retargeting people who already showed interest is smart. Showing them the identical ad they already scrolled past is not, and this mistake is sneaky because the campaign still looks alive. Impressions are climbing. Frequency is climbing. Spend is climbing. Something feels like it's working.

But if someone already clicked, looked around, and left without buying, that's data. They already said no to this exact pitch once. Showing it to them again isn't a second chance, it's the same rejected argument on repeat, and most people learn to tune it out fast. A sharper retargeting ad addresses the actual hesitation directly: price, trust, timing, whatever it likely was, rather than simply shouting the same message louder.

Why this matters beyond advertising

Strip away the platform-specific jargon and this stops being a story about ads at all. It's a story about what happens when people skip the unglamorous work of defining success before chasing it, when they mistake precision for understanding, and when they confuse repetition with persistence.

Those same three mistakes show up everywhere decisions get made under pressure and measured against a deadline: product launches, fundraising pitches, even hiring plans. The businesses that consistently avoid them aren't smarter or better funded. They just did three unglamorous things before anyone saw the flashy part: they defined what success actually needed to look like, they understood exactly who they were talking to, and they knew the difference between showing up again and saying something new.

Everything else is just decoration.

👋 Hey, I'm Indi — here to talk AI search visibility

indi-seo — Thu, 24 Sep 2026 21:11:16 +0000

I run Aechron, where I help small businesses and startups get found — not just by Google, but by ChatGPT, Perplexity, Gemini, and every other AI tool people are now using to find a business before they ever click a link.

A little about how I ended up here: I've spent 10+ years across front-end development, technical SEO, and brand/creative design. Honestly, I got into SEO for practical reasons, not because I loved it — it always felt like a mix of guesswork and chasing an algorithm nobody fully understands. But the AI-search shift is the first part of this field that's actually made it interesting again, because it's less about gaming a ranking system and more about a genuinely new problem: how do you get a machine to understand, trust, and cite your business?

That's mostly what I'll be posting about here:

Why AI-built sites (Lovable, v0, Bolt, and similar tools) often look polished but are functionally invisible to AI crawlers, even when they rank fine visually
What AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization) actually mean in practice — not the buzzword version, the "here's what to actually change on your site" version
Simple, five-minute checks anyone can run on their own site to see how an AI model reads it right now
Real examples from client work at Aechron — what's working, what's outdated advice that's still floating around, and what I got wrong before I figured it out

I'm not interested in recycling generic SEO advice that's been repeated a thousand times. Most of what I share will come from actually testing things — checking how ChatGPT or Perplexity responds to a real site, poking at schema markup, seeing what gets cited and what gets ignored.

If you want more of this between posts here, I write longer breakdowns on the Aechron blog — same topics, just more room to go deep.

If you're building sites fast with AI tools and want to know whether they're actually findable once they're live, if you're curious about "directed AI" workflows (human strategy steering AI execution, not replacing it), or if you just want to compare notes on what's genuinely working in SEO right now versus advice that's past its expiration date — say hey. Always happy to talk shop.

Why Your AI-Built Website Might Be Invisible to ChatGPT

indi-seo — Thu, 24 Sep 2026 20:53:45 +0000

I keep running into the same thing lately. Someone spins up a site in an afternoon using one of these AI builders — Lovable, v0, Bolt, whatever — and it looks genuinely great. Clean design, smooth animations, the whole thing. Then I check how it actually performs when an AI model tries to read it, and it's like the site doesn't exist.

Not broken. Just invisible.

For most of the last two decades, "getting found online" meant one thing: rank on Google. That's shifting fast. A lot of searches now never produce a click at all — Google just answers the question in the AI Overview, or someone asks ChatGPT or Perplexity directly and gets a recommendation without ever visiting a website. Which means your site isn't only competing to rank anymore. It's competing to get cited — to be the thing an AI model actually pulls from when it answers "best [whatever you do] in [your city]."

And here's the part that surprised me the first time I dug into it: most AI-generated sites are really bad at this, even the ones that look flawless.

Why "looks finished" doesn't mean "readable by a machine"

These builders are optimized to impress a human scrolling through a preview. That's it. They're not thinking about what a crawler — human-built or AI — needs in order to actually understand and trust a page. A few things I see constantly:

The page title and meta description are missing or generic. People assume that stuff is an old SEO relic. It's not — it's often the very first signal an AI system uses to figure out what a page is even about.

The content itself is thin. A hero section and three feature cards don't actually answer any question a customer would ask, so there's nothing for a model to cite even if it wanted to.

No schema markup anywhere. Structured data is basically how you hand a machine an unambiguous answer instead of making it guess from vague marketing copy.

Internal linking is shallow or broken. If a crawler can't move through the site logically, it can't build a full picture of what the business actually does.

Nothing on the page is shaped like an answer. A plain sentence that says how much something costs or how long it takes gets pulled into AI answers way more often than clever brand language does.

None of this shows up when you're just looking at the design. It only shows up once you check how an AI actually parses the page — which almost nobody does before launching.

The shift from SEO to AEO/GEO

Worth knowing even if you're not a marketer, because the vocabulary is genuinely shifting under everyone:

AEO — Answer Engine Optimization — structuring your content so it directly answers what people are asking, in a form a model can lift cleanly.

GEO — Generative Engine Optimization — earning the kind of trust signals (structured data, authority, clean citations) that make a model comfortable citing you inside a generated answer.

Traditional SEO isn't dead. It's still the foundation both of these sit on. But it's not enough by itself anymore if a real chunk of your traffic is going to come from someone who never clicks a blue link in the first place.

A quick way to check yourself

Takes about five minutes, and it's genuinely worth doing before assuming this is a someday-problem:

Ask ChatGPT or Perplexity a question your business directly answers. Something like "best [service] in [your city]." See if you show up. If you don't, that's data, not bad luck.

Look at your homepage's raw HTML. Is there an actual meta description and title tag that describe what you do — not just your brand name sitting there alone?

Search site:yourdomain.com and see whether your core pages are even indexed.

Ask an AI model to summarize what your business does using only your homepage. If the summary comes back vague or just wrong, that's exactly what a real AI search result would sound like too.

If most of that comes back weak, the site probably looks done and functions half-built — at least from a machine's point of view.

The part that actually matters

These fast-build tools solved something real. A lot of small businesses genuinely couldn't afford a custom site before this. But shipping fast and shipping findable aren't the same thing, and that gap is exactly where a lot of AI-built sites are quietly losing traffic they don't even realize they're missing.

If your site was built to impress a human scrolling through it, but nobody's ever checked how a machine reads it, that's worth five minutes of testing before it's worth a full redesign.

Ask an AI model to summarize what your business does using only your homepage. If the summary comes back vague or just wrong, that's exactly what a real AI search result would sound like too.

If most of that comes back weak, the site probably looks done and functions half-built — at least from a machine's point of view.

The part that actually matters

If your site was built to impress a human scrolling through it, but nobody's ever checked how a machine reads it, that's worth five minutes of testing before it's worth a full redesign.