Position 18 on the Discount Calculator. The Queries Were '130?' and 'Da'.

#seo #webdev #ai #claude

Forty days ago, the United States didn't appear in our top-20 countries by impression volume. That was one of the findings from the first audit — the site wasn't even showing up enough in US searches to land in the country breakdown. I noted it, flagged it as something to investigate, and moved on.

This week I pulled the 28-day GSC data and the US is now our single biggest market by impressions.

5,328 impressions. Two clicks. CTR: 0.038%.

That's 46% of every impression the site generates, coming from a country where we have almost no brand recognition, no backlinks worth mentioning, and an average ranking position of 56.4 — which is page five to six in Google.

The US went from invisible to dominant. And we have nothing to show for it.

The setup

The automation has run every day since June 30 without missing one. Forty consecutive commits, one calculator page per day — benchmarks updated, worked examples corrected, internal links wired up. The last seven days alone touched ROAS, Burn Rate, CPA, CTR, CPC, Valuation Multiple, and Book Value calculators. Before that, 33 more: real estate, startup finance, equity, valuations, compound interest.

Forty pages improved. What did it do to GSC? That's what I sat down to measure.

Total clicks in the last 28 days: 41. Total impressions: 14,675.

That's roughly 1.5 clicks per day, which is up from the baseline — the first audit found 45 total clicks across 90 days (0.5/day). Not nothing. But not the trajectory I was hoping for after 40 pages of content work, either.

The US country data is where the story gets interesting.

Finding #1: We're visible in America. At position 56.

Eight weeks ago, the US didn't crack the top 20 countries by impression volume. Now it's first.

Country	Impressions	Clicks	CTR	Avg Position
USA	5,328	2	0.038%	56.4
UK	1,887	2	0.106%	59.7
India	919	12	1.306%	16.7
Philippines	541	1	0.185%	27.5
Indonesia	511	2	0.391%	29.3

India still generates the most actual clicks — 12 out of 37 total from country-level data, at an average position of 16.7 and a CTR that's 34 times higher than the US. That pattern is familiar from the Week 1 data. What's new is that the US now generates more impressions than every other country combined.

How did that happen? I didn't change the hosting region. I didn't add US-specific content hooks. The automation was hitting finance calculator pages with broadly universal content.

I don't have a clean answer. But I have a clue. It's in the query breakdown of one specific page.

Finding #2: The discount calculator ranks at position 1 for "130?"

The discount calculator is an interesting case in the 28-day page data: 1,400 impressions, 1 click, average position 18.1.

Position 18. That's page 2 — which, for a competitive head term, is where the work lives. I expected the query breakdown to show something like the loan-payment-calculator from the first audit: dozens of tiny long-tail queries averaging to an aggregate that looked better than reality.

That's not what I found.

I pulled the query-level breakdown for the discount calculator page. Of 1,400 impressions, only 85 showed up in the visible rows — 6% of the total. The rest are below GSC's privacy threshold. Of those 85 visible impressions, here are some of the queries:

Query	Impressions	Position
discount calculator	40	43.9
check discount	19	44.9
20% off 70$	2	9.0
chain discount calculator	2	39.0
da	2	2.5
130?	1	1.0
1490	1	1.0
450 mil	1	1.0
4600	1	1.0

Let me explain what's happening with "da," "130?," "1490," and "4600."

These are not discount calculator searches. These are people using Google as a quick calculator — typing a number or a fragment and letting Google auto-complete or suggest results. Google is surfacing the discount calculator page at position 1 for these queries because it's a calculator that handles numbers. The user who typed "130?" was almost certainly not looking for a discount percentage tool. They were looking for a quick numeric answer.

The result: we appear first in their SERP, they don't click, and that impression goes into our GSC data with a position of 1.

This is different from the position-averaging trap I described in earlier posts — where many tiny real queries with relevant intent averaged out to make a page look stronger than it was. This is something else: queries with completely wrong intent ranking us at position 1, inflating our impression count while contributing nothing to clicks or engagement.

The actual "discount calculator" head term sits at position 43.9. That's where the real intent lives, and that's where we actually need to compete.

The 1,400 total impressions include some unknown volume of arithmetic queries below the privacy threshold — queries like "130?" but in different numbers, or other keyboard artifacts. They're there in the data, pulling the average position down (making it look more favorable), showing up as impressions, and then not converting because they can't. The user who types "1490" into Google doesn't want our calculator.

This is probably affecting more calculator pages than just the discount one. The ROI calculator has 2,277 impressions in 28 days and 1 click — at position 34.9. I'd want to see the query breakdown to know how many of those impressions are genuine calculator intent vs. arithmetic noise.

Finding #3: The og:title bug is still in index.html

In the first audit post, I found that every page had two og:title meta tags — one generic, baked into the static HTML, and one page-specific one injected by React Helmet at prerender time. I wrote that I'd strip the static tags out of index.html "tonight."

I curled the discount calculator page this week:

<meta property="og:title" content="Valuefy - Free Business Calculators & Financial Tools">
<meta property="og:title" content="Free Discount Calculator — Sale Price, Stacked & Reverse" data-rh="true">

Both tags are still there. I confirmed it on the homepage, the ROI calculator, the impression calculator, and the MSP blog post — same pattern, every page. The fix was to remove four lines from client/index.html. It has not happened.

The immediate consequence: social shares of every individual page show the generic site-wide title in the preview card. Whether this is affecting GSC click-through rates is harder to measure — but it's the kind of compounding technical debt that doesn't get cheaper over time.

What I'm going to do

Fix the og:title duplicate — this time actually. Strip lines 23–25 from client/index.html. The fix is four lines. It has been pending for eight weeks. There is no excuse for it to still be open.
Pull query breakdowns on the top five impression pages. The discount calculator analysis revealed arithmetic query contamination. I want to know how widespread this is — specifically for the ROI calculator (2,277 impressions, 1 click) and the impression calculator (2,101 impressions, 12 clicks). If the 12-click page is also catching irrelevant queries, that changes how I read its CTR.
Look at the US position distribution. The average US position is 56.4 across all pages. That's an aggregate of everything — some pages might be at 30 (page 3, not hopeless) and others at 90 (might as well not exist). I want to see which US-visible pages are closest to page 2 and whether the automation pages are in that subset.
Keep the automation running. Forty days of consistent updates is the minimum sample before drawing conclusions about which categories respond to content improvements. The query contamination issue doesn't change that — it changes how I read the impression data, not whether the content work matters.

The uncomfortable lesson

There are two ways to look at the US impression surge. One is optimistic: the site went from not appearing in US searches at all to generating 5,328 impressions in a single month. That's genuine progress on visibility.

The other way is more accurate: nearly half of those impressions come from queries where we have almost nothing to offer. The arithmetic query contamination I found on the discount calculator is probably a site-wide pattern — calculators that handle numbers will keep appearing for arithmetic queries because Google's intent classification isn't perfect. And at position 56 average, even the legitimately relevant queries aren't generating clicks.

The impression count looks better every month. What it hides is the distribution: some impressions from real intent at bad positions, some impressions from wrong intent at good positions, and almost no impressions from good intent at good positions.

I've spent 40 days improving content. I haven't spent a single day cleaning up the bug I promised to fix eight weeks ago. That tells me something about where my attention is going — and it's probably not where the leverage is.

I'll check back in three weeks: og:title fixed, ROI calculator query breakdown pulled, and a read on whether the US position distribution has any pages in striking distance of page 1. If the arithmetic contamination turns out to be the dominant story in that breakdown, I'll say that too.

I'm running these experiments on valuefy.app and writing the findings as I go. If you're working through the same programmatic SEO problems — position metrics that don't match what the query breakdown shows, automation that generates commits without generating clicks — drop a comment. I'd like to compare notes.

I also run AImiten, where we build AI tooling for companies. This side project is where the ideas get stress-tested.