Programmatic SEO for SaaS: How We're Targeting 130 Keywords With AI

# Programmatic SEO for SaaS: How We're Targeting 130 Keywords With AI

Most SaaS founders treat SEO like a box to check. Write a few blog posts, sprinkle in some keywords, wait six months, wonder why nothing happened.

I decided to do it differently. Instead of guessing which keywords to target, I had Ari build a full keyword research system, map every relevant phrase in our space, and then systematically publish against them — one article per day, fully automated.

Two weeks in, we're 14 articles deep. Here's the entire methodology.

The Problem: Zero Indexed Content

When we launched SwipeBase, we had a landing page and a product. That's it. No blog, no indexed content, no organic traffic. Every visitor came from direct outreach or paid channels.

Meanwhile, competitors like Foreplay and MagicBrief had hundreds of indexed pages ranking for the exact terms our potential users were searching. Terms like "best ad spy tools," "how to build a swipe file," "foreplay alternative."

We were invisible to anyone searching for a solution we'd already built.

Step 1: Map the Entire Keyword Universe

I told Ari to find every keyword phrase someone searching for an ad creative research tool might use. Not just the obvious ones — I wanted the long-tail, the comparison queries, the "how to" searches, the industry-specific stuff.

The process was straightforward:

1. Start with seed terms. Swipe file, ad spy, ad creative, hook formulas, competitor research.
2. Expand through search suggestion mining. What does Google autocomplete when you type each seed? What "People Also Ask" questions show up?
3. Cross-reference with competitor content. What are Foreplay, MagicBrief, AdSpy, and BigSpy ranking for? Where are the gaps they haven't filled?
4. Validate with intent analysis. Is someone searching this phrase likely to want a tool like ours? Or are they just browsing?

We ended up with 130+ keyword phrases. Not a round number we aimed for — just where the research naturally landed.

Step 2: Cluster by Topic, Not Just Volume

Raw keyword lists are useless. You need structure. I had Ari group everything into 12 clusters based on topic similarity:

1. Swipe File / Ad Library — Core positioning keywords (12 phrases)
2. Hook Formulas / Scroll-Stoppers — High-volume traffic drivers (11 phrases)
3. Ad Breakdowns / Creative Analysis — Analysis content (11 phrases)
4. Funnel Teardowns / Landing Pages — Full-funnel content (9 phrases)
5. UGC / Creator Ads — Trending format content (10 phrases)
6. Platform-Specific — Meta, TikTok, YouTube tactics (12 phrases)
7. Industry-Specific — Ecommerce, SaaS, coaching verticals (10 phrases)
8. Competitor Tools — Direct comparison keywords (12 phrases)
9. Ad Copy / Scripts / Frameworks — Copywriting content (10 phrases)
10. Creative Testing / Scaling — Performance optimization (11 phrases)
11. Retargeting / Funnel Stages — TOFU/MOFU/BOFU content (10 phrases)
12. Seasonal / Trending — Calendar-driven content (12 phrases)

Clustering matters because it shows you where to build depth, not just breadth. One article about swipe files is a blog post. Twelve articles about swipe files, hooks, ad analysis, and creative research? That's topical authority.

Step 3: Prioritize by Intent Level

Not all keywords are equal. A comparison search like "Foreplay alternative" has way higher purchase intent than "what is a swipe file." Both matter, but the order you publish them in changes everything.

We used a three-tier priority system:

• Priority 1: High volume + low competition + strong tool intent. These get written first.
• Priority 2: Solid opportunity, maybe more competitive or lower volume.
• Priority 3: Niche but valuable — worth having for topical authority.

Within Priority 1, we further segmented:

• Tier 1 (Publish First): Core positioning + competitor comparison keywords. These are the people actively shopping for a tool like ours.
• Tier 2 (Publish Next): High-volume educational content that builds traffic and backlinks.
• Tier 3 (Calendar-Driven): Seasonal content timed to when people actually search for it. Black Friday content goes live in September, not November.

Step 4: Automate the Publishing Pipeline

Here's where it gets interesting. I didn't want to manually write 130 articles. I also didn't want garbage auto-generated content that reads like it was written by a robot.

The solution: I set up a daily cron job. Every morning at 10am, Ari picks up the next article from the queue, reads the keyword brief and any source material, writes the full article, generates a hero image, and deploys it to the live site.

The pipeline looks like this:

article-queue.md → AI writes article → hero image generated → deploy script → live on site

Each article targets 1-2 primary keywords plus related long-tail phrases. The queue file tracks what's published and what's next. No human bottleneck.

Does every article come out perfect? No. Some need tweaking. But the baseline quality is high enough that publishing daily beats publishing perfectly once a week.

Step 5: Start With Competitor Keywords

This was a deliberate choice. Our first batch of articles targeted competitor comparison terms:

• "Best Foreplay alternatives"
• "AdSpy vs BigSpy vs Foreplay"
• "Minea alternatives"
• "SwipeWell vs Foreplay"
• "Free ad spy tools"
• "Best ad spy tools 2026"

Why? Because someone searching "Foreplay alternative" is actively looking for a tool. They already understand the category. They already have budget intent. They just need to know SwipeBase exists.

These comparison articles convert at dramatically higher rates than educational content. One well-ranked "Foreplay alternative" article is worth more than ten "what is a swipe file" posts in terms of trial signups.

The Numbers So Far

After 14 days of daily publishing:

• 14 articles live across 3 clusters
• Tier 1 competitor keywords mostly covered
• Moving into Tier 2 traffic-driver content (hook formulas, ad breakdowns)
• Zero manual writing — every article generated and deployed by the pipeline

It's too early for meaningful organic traffic data — Google takes weeks to index and rank new content. But the foundation is laid. When these articles start ranking, they'll compound.

What I'd Do Differently

Start SEO on day one. We launched SwipeBase and then waited weeks before thinking about organic content. Every day without indexed content is a day of compounding you'll never get back.

Prioritize comparison keywords even more aggressively. If I could go back, the first 20 articles would all be competitor comparisons and "best tools" listicles. They're the highest-intent, most commercially valuable content you can publish.

Build internal linking from the start. We're retrofitting this now, but it should have been baked into the article template from article one.

The Takeaway

Programmatic SEO isn't about tricking Google. It's about systematically covering every way a potential customer might search for what you offer. 130 keywords sounds like a lot. But when you cluster them, prioritize by intent, and automate the publishing pipeline, you can cover them all in a few months.

The AI doesn't replace the strategy. I picked the keywords, defined the clusters, set the priority order, and decided which articles to write first. The AI just makes it possible to execute at a pace that would be insane for a solo founder to do manually.

That's the whole point of Machine Earned. Not AI replacing humans. Humans with AI doing what used to require a team of ten.

---

Building SwipeBase in public. Follow along at machineearned.com for daily updates on what's working, what's failing, and what we're building next.