Building an SEO Tool That Doesn't Lie to People

Objective

I’d built earlier versions of an SEO content-checking tool back when this site ran on Blogger. Going back through that old code honestly, it had real problems: a “GEO Score” presented like a validated metric but actually just an arbitrary point system, and an email capture form that said “Check your email! Your report is on the way” — while actually just storing the address in the browser’s local storage and doing nothing. That’s not a small style choice. That’s telling someone something happened when it didn’t.

I wanted to see if I could rebuild the useful part — a real content checker — without any of that.

Starting Point

The old tool checked title length, word count, first-paragraph length, keyword presence, and heading structure. The actual checks were reasonable. The presentation around them wasn’t.

Hypothesis

If I strip out the fabricated score, the fake “2,000+ bloggers” social proof, and the non-functional lead capture, what’s left should still be a genuinely useful tool — just a smaller, more honest one.

Process

Rebuilt from scratch as a real React component inside the Astro site (not a Blogger widget this time), with each check calculating something real and explaining, in plain language, what the number actually means and doesn’t mean. There’s still a single summary percentage, but it’s calculated transparently as “X of 5 checks passing” — not a mysterious index number pretending to measure something it can’t.

Results

Live and working at /tools/seo-checker.

Five real checks, all calculated client-side, no external calls, no data collection. Every threshold used (like the 50–60 character title guideline) is labeled as a common heuristic, not a scientific finding or proprietary research.

What Worked

Keeping the score transparent turned out to make the tool more trustworthy, not less — five separate, honestly-labeled checks, summarized as a plain pass-count percentage, are more useful than one opaque number that hides what’s actually being measured.

What I Learned

A fake success message is worse than no feature at all. If a tool can’t actually do something, saying so is better than faking it.
Removing a single misleading metric (the old “GEO Score”) didn’t make the tool feel weaker. It made it feel more trustworthy, which matters more.
This is a good template going forward: before I ship anything with a “score” or a “report,” ask whether it’s a real calculation or a number I made up to look impressive.

Conclusion

Yes — a genuinely useful version of this tool is possible without any of the fabricated parts. It does less than the old one claimed to do, but everything it does is true.