DEV Community

muddasirriaz099
muddasirriaz099

Posted on

I built a free calculator site from scratch: 3 weeks of honest Search Console data

Three weeks ago I started an SEO experiment: build a set of free US work and pay calculators, host them on free GitHub Pages, and track everything in Google Search Console in public. No budget, no custom domain, no backlinks.

This is the honest report, including the parts that did not go well.

What I built

EverydayCalc is a static site with 10 calculators: work hours, time card, overtime, a California overtime calculator, lunch break, hours between two times, hourly to salary, salary to hourly, BMI and paycheck.

Each page has the calculator first, then the formula, worked examples, a visible FAQ and links to official sources (U.S. Department of Labor, California DIR). I wrote the FAQ schema to match the visible FAQ, and I added a "Quick answer" box at the top of each page so a search engine or AI answer can quote it.

The numbers (23 Oct to 12 Nov)

  • Impressions: about 930 in total, from 0 on the first day to a peak of 173 in one day (3 Nov)
  • Clicks: 4, and all four were me checking my own site
  • Average position: about 29
  • Queries Google showed the site for: over 100, almost all long-tail ("time calculator with lunch", "8 to 5 how many hours")
  • Best position for a single query: 13 ("hourly labor calculator")
  • Where the impressions came from: the United States had the most (position around 43, so page 4 or 5), then Brazil, Canada, Russia and Bangladesh

What went wrong

1. Impressions fell after the peak. After 3 Nov the daily number dropped from 173 to 91, then 14, then 0 on 10 and 11 Nov.

2. Some pages are "Crawled - currently not indexed". In the Page indexing report, 9 pages were indexed and 4 were not: the home page, the time card calculator, the overtime calculator and the contact page. When I inspected the work hours page (my best one) it also said "URL is not on Google", even though Googlebot had just fetched it successfully.

3. The sitemap says "Couldn't fetch". The XML file is valid and opens in the browser, but Search Console has not accepted it for almost two weeks.

What I think is going on (and what I am not sure about)

I am not sure, and I do not want to pretend I know. My best explanation is that Google tested a brand-new site on many queries, found no signals that anyone cared (no backlinks, no real clicks), and then stopped showing it. "Crawled - currently not indexed" usually means Google fetched the page and decided it was not worth keeping yet. The page itself is technically fine (fetch successful, no blocks, Google's chosen canonical is the page itself, so it is not seen as a duplicate).

What I am doing next

  1. Waiting a week before touching the pages that are being tested, so I can tell cause from effect
  2. Getting real mentions and links (this article is part of that)
  3. Buying a proper domain and connecting it after the test
  4. Reporting the numbers again in a week, including if they are bad

Lessons so far

  • Impressions come before rankings. A site can show up for 100 queries and still get zero clicks.
  • Position 30 earns nothing. Page 3 or 4 does not get clicks, so impressions alone are a vanity number.
  • Always reset your Search Console filters before reading the data. I once thought my traffic had collapsed. It was a filter I forgot to clear.
  • Do not change five things at once. If something moves, you will not know why.
  • New sites need signals from outside. Content is not enough on its own.

If you have run into "Crawled - currently not indexed" on a new site, what worked for you? I would like to hear it in the comments.

Top comments (6)

Collapse
 
md_khorshedalam_8e5decac profile image
Md Khorshed Alam • • Edited

Appreciate the honesty here, especially counting out your own 4 clicks so the data stays clean.

Just a quick note on the sitemap error: GSC almost always gets stuck on "Couldn't fetch" for free .github.io URLs even when the file is completely fine. Moving to a custom domain usually fixes it instantly. In the meantime, your internal links between the 10 calculators are more than enough for Googlebot to find everything.

Will be following along for next week's update. When you connect the custom domain, are you going to redirect the old GitHub Pages URLs or just treat this first month purely as a throwaway test?

Collapse
 
muddasirriaz099 profile image
muddasirriaz099 •

Thanks, that's really helpful, and good to know the sitemap issue is common on github.io. I'll report back on whether it clears after the move.

On your question: I'm not treating it as a throwaway. The plan is to connect a custom domain after my current week-long test finishes, let the old GitHub Pages URLs redirect to the new ones (GitHub does this when a custom domain is set, but I'll verify it before relying on it), update the canonical tags and sitemap to the new domain, and add the new property in Search Console. I'll write up how the migration goes in a follow-up post, including whether indexing improves.

Collapse
 
onelifevibe profile image
OneLifeVibe •

Thanks for sharing the real numbers instead of a success story. The part that stood out to me was the drop after 3 Nov, from 173 impressions down to zero, and your decision to wait a week before touching the pages so you can tell cause from effect. That takes discipline when the instinct is to fix everything at once. I also liked that you counted out your own 4 clicks so the data stays honest. Looking forward to next week's update, especially what happens after the custom domain goes live.

Collapse
 
muddasirriaz099 profile image
muddasirriaz099 •

Thanks, I appreciate that. The week of waiting is the hardest part, because every instinct says to fix something. I'll share what changed, or didn't, next week, and then what happens after the domain move.

Collapse
 
rezapratama_serpair profile image
Reza Pratama •

On the “Crawled - currently not indexed” URLs, I’d separate that from the sitemap fetch issue: a sitemap helps discovery, but it doesn’t make Google keep a URL. After your week of waiting, compare one indexed and one excluded calculator for genuinely distinct examples and content, internal links, and the canonical Google selected. That should help distinguish thin or overlapping pages from a temporary processing issue without changing several variables at once.

Collapse
 
seotto profile image
seotto.com •

Thanks for sharing the part that didn't go well, most experiment write-ups stop at the peak. A 173-then-zero curve on a brand-new site with no links looks a lot like Google trying the pages out and then backing off, so waiting a week before changing anything sounds right. The one thing I'd do sooner is the custom domain: if you move off github.io after Google has started forming an opinion, the new host starts from zero again. Has the "Couldn't fetch" sitemap cleared yet, or is Google finding pages through internal links only?