How I Ranked My AI Poem Generator on Google's First Page With 0 Backlinks

A while back I built a small AI poem generator as a side project. Nothing fancy. You type in a feeling or an occasion, and it spits out a poem you can actually use, for a card, a caption, a toast, whatever.

Like most indie hackers, I had zero marketing budget and zero patience for outreach emails that get ignored. So I never built a single backlink. No guest posts, no directory submissions, no link swaps. And yet the site ended up ranking on page one for several decent keywords in its niche.

Here's exactly what I did, in case it helps anyone else sitting on a small tool with no audience and no links.

The starting point was rough

When I launched, the site had nothing going for it. No domain history, no social following, no email list. Just a working product and a domain I'd registered a few months earlier.

I want to be upfront about something. Ranking with zero backlinks does not mean SEO doesn't matter. It means I put all my effort into the parts of SEO that don't require begging other websites for a link. That turned out to be enough, at least for this niche.

Step 1: I picked keywords a big site wouldn't bother with

Everyone wants to rank for "AI poem generator." That term is dominated by tools with years of authority behind them. I never had a shot there, and going after it would have wasted months.

Instead I looked for longer, more specific searches. Things like poems for a specific occasion, poems in a specific style, poems for a specific relationship. These searches get less volume individually, but there are hundreds of variations, and almost nobody was writing dedicated content for them.

Low competition plus real search intent is the whole game when you have no authority to lean on.

Step 2: I made the tool answer the query directly

This part matters more than people think. When someone searches for "poem for my mom's birthday," they don't want a blog post about poetry history. They want a poem for their mom's birthday, right now.

So instead of writing generic articles, I built landing pages where the tool was pre-configured for that exact search. Someone lands on the page, the context is already set, and they can generate a poem in one click. Google noticed that these pages matched search intent better than most of the competition, which was just blog content with no actual tool attached.

If you're building something similar, my honest suggestion is to try a poem generator style setup yourself before assuming you need a huge content library. The tool being useful on its own did more for rankings than any article I wrote.

Step 3: I wrote content that was actually useful, not padded

I'm not going to pretend content didn't matter, because it did. But I kept every page short and specific. No 3,000 word essays stuffed with fluff just to hit a word count. If a topic needed 500 words to be useful, I wrote 500 words.

I also answered the exact follow up questions people search for right after the main one. Things like how long a poem should be for a certain occasion, or what tone works best. These small additions picked up featured snippets a few times, which brought in extra clicks without needing to rank number one.

Step 4: I fixed the boring technical stuff early

This is the least exciting part of the story but probably the most important. Before I wrote a single page of content, I made sure:

• Pages loaded fast, since a slow tool kills conversions and rankings both

• Every page had a clear, unique title and description

• The site worked properly on mobile, because most of my traffic ended up being mobile anyway

• Internal links connected related pages so Google could actually crawl and understand the site structure

None of this is glamorous. But a fast, clean, well structured site with mediocre content will often beat a slow, messy site with great content. I learned that the hard way on an earlier project.

Step 5: I let user behavior do some of the work

Once a few pages started getting a trickle of traffic from long tail searches, something interesting happened. People stuck around. They generated a poem, copied it, sometimes generated a few more. Time on page was decent and bounce rate wasn't terrible.

I can't prove this directly influenced rankings, since Google doesn't confirm exactly how much weight it gives to engagement signals. But it's a pretty consistent pattern across my projects that pages people actually use tend to climb faster than pages people bounce off of immediately.

What I'd tell someone starting from zero

If you're building a small tool or site and have no links and no audience, here's what actually moved the needle for me:

1. Go after specific, low competition searches instead of the obvious head term

2. Make sure whatever you're offering solves the exact query, not a related topic

3. Keep content tight and useful instead of padded for length

4. Fix speed, mobile, and basic structure before worrying about anything else

5. Give people a reason to stick around once they land

Backlinks help, no doubt about it. I'm not saying skip them forever. But if you're stuck at zero and waiting for links before you start optimizing everything else, you're leaving a lot of easy progress on the table. I got a small tool onto page one without a single one, just by making sure it actually deserved to be there for the searches it was targeting.

Try my Poem Generator here: https://poem-generator.net/