A tool I reviewed audited my own site and found a bug I'd missed

The "what the tool says vs what it actually does" gap is the most underrated SEO review angle. I've found the same thing auditing review tools against my own setup.
The pattern I keep seeing: the tool says the audit is X / B+ / Pass, but when I actually check the page they flagged, the issue is real but the fix is wrong (or fixable, but they're calling it unfixable). Frase in particular is good at finding the real issue but bad at prioritizing which to fix first.
One methodology tweak that helped me: instead of "did the tool find issues", track "did the tool find issues I wouldn't have caught manually." A tool that just restates your title tag is useless. A tool that finds something you genuinely missed is gold. That filter cuts the tool pile in half fast.

It’s great that you're diving into AI/SEO tools and sharing your insights from your reviews. This can help a lot of solo founders and content creators. I’ve been experimenting with different tools in my own journey with Scaleblogger, where I focus on optimizing content creation and managing SEO more effectively.

One thing I've found valuable is using platforms that integrate AI seamlessly to audit sites and identify issues. For instance, Scaleblogger’s built-in analytics can help highlight content performance and areas needing improvement without manually digging through data. This is especially helpful when you're dealing with large amounts of content saving significant time and effort.

If your reviews touch on how these tools can automate processes (like content creation and optimization), that could really resonate with those of us looking to streamline our workflows. Looking forward to seeing what insights you gather from the tools you review!

Rewriting the title in the same session as the GSC hit is better than a calendar of hypotheticals.

I’d only add a boring weekly glance at Search Console so “something surfaces” isn’t luck. Reactive is fine; blind is the failure mode.

Your numbers are almost exactly mine (eight tool pages, four indexed, ~18 impressions a day, zero clicks), so take this as a peer guess, not wisdom. What I've actually seen: new pages got indexed by recency of crawl, not age (the page I shipped last was indexed first while older ones waited) which tells me Google is rationing a link-poor domain, and neither more content nor patience fixes rationing. The only lever that changes it is a handful of real links from pages that are themselves crawled. So my answer is neither: the order is links first, content second and the content only starts paying once the crawl budget exists. The og:image catch is a good story precisely because you kept it in the review instead of quietly fixing it.

Great breakdown. I’m also learning SEO while building a small website, and I’ve been wondering about the same thing. From what I’ve seen so far, I think a combination of useful content and a few relevant backlinks makes more sense than focusing heavily on either one alone. Early GSC numbers can definitely be discouraging, but three months still feels quite early.

That sequencing makes sense — content first, then backlinks once there’s enough surface area to judge. I’d probably keep the two paths explicitly separated in the notes so the early baseline stays interpretable. I’m curious what signal you’ll use to decide there’s enough content to start the backlink experiment.

Content first, no contest — 19 impressions means you're still being indexed. Backlinks amplify pages that already rank; they don't create the first ranking. Write until GSC shows a cluster with clicks, then chase links on those exact pages. I ask my analytics AI which posts actually moved the needle: https://amami.dev

This is the measurement boundary problem in reverse - you built the site, you review it, and you still miss the bug. That's not a capability problem, it's a measurement system problem. Your own mental map of your own site has a blind spot, and only a tool that sees the site fresh (without your intentions narrated on top of it) catches it.

The uncomfortable part is how universal this is. A founder reviews their own pricing page and doesn't see the price inconsistency. A team reviews their own API docs and misses the undocumented parameter. Not because they're careless, but because their internal knowledge of what the page should say interferes with seeing what the page actually says.

External audits work because they measure against the artifact itself, not against the intended artifact. The tool doesn't know what you meant to build. That's the gap where bugs hide."

This is why I like running tools against my own sites. You catch the boring stuff you mentally marked as “done” weeks ago.

The og:image thing hit home — I had the same false "done" feeling about social previews until I actually scraped one of my own URLs. Good call putting the fix in the review instead of hiding it; that's the kind of transparency that builds trust with readers.

On content vs backlinks: for my first site, content alone got me indexed and ranking for long-tail terms, but backlinks were what pushed posts from page 3 to page 2. Content built the foundation; one decent link from a relevant blog moved more posts at once than any single article did.

How long did you wait before deciding a page was worth rewriting versus just moving on to the next piece?

The "actually use the tool before writing about it" rule is what makes reviews worth reading, and this post is the proof — a review that finds a real bug in your own site can't be written from a feature table. The og:image thing is a classic solo-founder failure mode: it's in the "done" column in your head, nothing ever errors, and no user reports it because they just don't click. I've started treating anything I marked done from memory rather than from a check as untested, and running a boring pre-ship pass on meta tags, canonical URLs and 404s before launching anything. Curious whether you'll keep running each reviewed tool against your own site as a standing test — that seems like a better review methodology than most affiliate sites have.

the 19 impressions and zero clicks make me wonder if the next useful test is not more content versus backlinks, but packaging. i would pick one page that already gets impressions, rewrite only the title and description around the exact query, then watch the click-through rate for two or three weeks. that gives you a cleaner signal than changing page count and links at the same time. the honest audit angle is strong, especially because you are showing the misses instead of only the wins.

This resonates — the scariest bugs are the ones your own tooling can't
see because you built the blind spot into the tool itself. Did you end
up changing your review process after finding it, or was it a one-off fix?

That “use it before writing about it” rule is a great differentiator—especially when the tool exposes something you thought was already fixed. The og:image example is a good reminder that checklists can create false confidence unless they’re validated from the outside. For an early content site, I’d be curious whether you’re planning to treat backlinks and additional content as separate experiments, so you can see which one moves impressions first. The low early numbers feel useful as a baseline rather than a verdict.

This is classic! Using your own tool and finding a bug. What was the bug you missed?

That’s probably where these tools are most useful — not replacing manual QA, but catching the small things you stop noticing after staring at your own product for too long. What kind of bug did it catch?

Love this level of transparency. It’s such a good reminder that we’re all blind to bugs on our own sites, even after we think we’ve checked everything. Curious, have you noticed any social preview improvements after adding the og:image tags? Also following your question — I’d lean toward more targeted content first in the super early phase.

The "use it on your own site before writing" rule is the whole review, honestly. The failure mode with solo content is writing from notes that were true a month ago — a checklist beats memory here: run it, record what broke, publish the ugly finding. Cheapest version I've found is one review slot a week where the only job is to re-verify claims I've already made. Are you keeping a running list of things you assumed were done, or catching them as tools surface them?

I’d focus on content first, especially at this stage. With 19 impressions, Google probably still doesn’t have enough pages or signals to really understand the site yet. A few relevant backlinks can help, but I’d rather have 30–50 genuinely useful, experience-based posts like these before spending much time chasing links. Also, including the failures you found while actually using the tools is probably your strongest angle. Most AI/SEO review sites just rewrite feature pages.

This is a really interesting use case for AI agents. The part I like most is the feedback loop: using the tool to discover something you wouldn’t necessarily notice yourself.

The bigger opportunity seems to be moving from “find a bug” toward continuous verification — especially when a system is changing frequently. That’s where observability and automated checks become really valuable.

the "I already marked this as done" trap is exactly what happened to me this week, just with a demo video instead of a bug. thought a trimmed, tightened cut would obviously perform better, posted it, got the same near-zero engagement as the first version. my own confidence that I'd "fixed it" wasn't evidence either, same as your og:image notes

on your GSC numbers: 19 impressions, 0 clicks after three months is rough, but not shocking for that stage, curious if the pages targeting a specific pain point rank differently than broader ones though, position 36 average could be hiding better and worse performers if it's an average across very different search intents

on backlinks vs more content, no direct experience with SEO myself, but the pattern I keep seeing elsewhere this week is that genuine community presence (answering real questions where people already are) compounds faster early on than either backlinks or content volume alone, might be worth testing alongside whichever of the two you pick

19 impressions, 0 clicks after 3 months — I feel this. I just launched my first project a few days ago and I'm still figuring out where the actual traffic comes from versus where I'm just shouting into the void.

On backlinks vs. content — from what I've seen so far, it seems like content only starts moving the needle once something links back to it from a place with actual traffic (a community, a newsletter, someone sharing it). Pure content without distribution feels like it just sits there, even if it's good.

Also respect for the "actually use the tool before reviewing it" rule — that Frase example (flagging your own bug) is a good example of why that matters. Most reviews I read clearly skip that step.

I learned this the annoying way on my own site rebuild: a checklist said pages were complete, but checking the rendered site exposed claims and states the source documents no longer supported. Since then I separate "more content" from "more verified content." At 19 impressions and an average position of 36, I'd publish more pages before chasing backlinks, but only if every page contains a finding you reproduced yourself. I'm not sure backlinks can rescue a site before Google has enough distinct, trustworthy pages to understand what it is. Which page has earned the most impressions so far?

Counter-datapoint on "the automated one, every time" — same failure you're describing, sign flipped.

An audit tool reported a batch of 5XX errors and broken links across our site. Every one was our own CDN rate-limiting the auditor's crawler. Nothing was broken for a human or for Googlebot. Separately, several "broken" outbound links were vendor hosts that return 403 to any non-browser user agent: fine in a browser, dead to the tool. And a code audit of ours produced three findings marked critical that were all non-bugs once I opened the URLs they named.

So automated audits have their own version of the Make problem: the tool reports confidently, and the report turns out to be about the tool rather than the system. Yours caught a true thing you'd marked done; mine invented work I'd have spent a day on. Same category — a status produced by something that never verified the downstream artifact.

The check I use now: reproduce the finding the way a real visitor arrives. Browser user agent, normal request rate, one URL at a time. If it doesn't reproduce that way, the finding is about the auditor. Two minutes per finding, and it has killed more of them than it confirmed.

On your traffic question, one thing worth ruling out early: those 5XX responses weren't harmless while they lasted. Search Console showed crawl volume dropping across the two weeks our CDN was throttling, and recovery needed a sustained clean run, not a fix and a re-request. At 19 impressions that won't show up in clicks, but crawl stats will show it. Worth checking before concluding it's a content-volume problem.

That’s a good point about using the tools you review on your own site. I’d probably prioritize more content first, especially in the early stage, but focus on low-competition topics rather than just increasing volume.

Once you have some pages getting impressions, backlinks could have a much clearer impact. Your GSC numbers are still very early, so I wouldn’t read too much into them yet.

That’s exactly why automated audits are useful—small technical bugs are easy to miss manually. Tools like Ahrefs and Semrush are great for SEO data, but I prefer SerpSpur when I want a broader site audit with technical, on-page, off-page, and local checks in one report.

https://serpspur.com

This captures something that silently kills shipping velocity: the gap between "I noted it as done" and "it measurably works in production."

The og:image thing is sharp because it's not a missing feature - it's a completed task on a checklist. You marked it done, moved on. But your measurement system was the person-who-wrote-the-note, not the automated verifier.

I notice teams that move fast maintain parallel measurement systems: human checklist (fast, forgetful) and automated audit (slower to set up, reliable, finds three more things). The audit doesn't feel like "more work" - it's the difference between "we think we solved it" and "we know we did."

And that gap compounds across 50 features. Your checklist says 50/50 done. The tool says 22/50. Which measurement system determines what you fix next?

The two findings that make this post work are both claimed-versus-actual gaps: your notes said social previews done while Frase found zero tags, and Make logged success while the email never sent. I would make that exact gap a fixed section of every review instead of a side lesson. 'What the tool says happened versus what we could verify' is harder for competitors to copy than opinion, and it gives readers a reusable question instead of another rating. On the traffic question: at position 36 with 19 impressions, I would publish more pages first, but only pages with a distinct verified finding; thin filler will not give you anything worth ranking.

This is why I paid for a Zarek audit before PH. Found 8 issues, 3 critical. The "expert eye" catches what the builder is blind to after 100 hours in the code.

Best $0 I spent (it was a free community audit). Worth 10x more than any ads.

The part about writing “social previews done” in your notes and then discovering they were completely broken is painfully relatable. 😅

I think that's also why actually using a tool before reviewing it is such a good rule. A feature can look perfect on paper while something completely different breaks once it touches a real workflow.

And honestly, 19 impressions / 0 clicks after only a couple of weeks doesn't sound particularly worrying to me. I'd probably prioritize publishing consistently first, then use GSC to see which pages/queries start getting impressions before going hard on backlinks.

Curious to see how the traffic changes once you've got a bigger sample size.

The "og:image done" → "actually zero tags" gap is pure measurement clarity in action. Your notes were a mental model. The tool was reality.

Most founders never run this kind of audit because it's easier to trust our own annotations. But what you're doing here is outsourcing verification to something that doesn't know how to lie - it just counts what exists.

That 19 impressions / 0 clicks is interesting because it actually tells you something specific: you're getting search visibility but the click-through is catastrophic. Most people would blame "position 36" and build more content. But your stack audit already found a concrete friction point (og:image, and presumably those "success" logging failures).

The measurement system determines what you can see. You built one that surfaces friction instead of just celebrating views.

The “I already marked this as done” part is the most interesting to me. It shows how easily our own notes become assumptions instead of evidence. Automated audits are valuable not just because they find bugs, but because they challenge what we think we've already verified. Curious whether you'll start treating these audits as a recurring pre-publish check.

The strongest part is the willingness to let the product audit the founder’s own assumptions. That makes the “review” model more credible because the process is producing findings, not just content about tools.