Eight weeks in. Forty-three automation commits since May 30 — one calculator page improved per day, every day, without a gap. This week's batch: seven startup/VC pages, ARR through Cap Table. And while pulling the GSC data for this post, I landed on a single row in the ROI calculator breakdown that derailed the whole narrative I'd planned to write.
Last week I found three clicks I couldn't explain. This week I found one click I can explain too well.
1 click. From 1,145 impressions on the query "roi calculator." In 28 days.
The setup
Here's the current state of things:
- 43 automation commits since May 30, one page per day, every day
- 160+ calculator and generator pages across valuefy.app
- This week's vertical: startup/VC calculators (ARR, Churn Rate, DCF, Funding, Vesting, Dilution, Cap Table)
- GSC, last 7 days: 5 total clicks, 1,004 impressions across 30 pages
- GSC, last 28 days: 46 clicks, 10,330 impressions across 50 pages
The 28-day per-week average works out to about 11 or 12 clicks. This week delivered 5. The four weeks of automation commits — and specifically the startup cluster we ran this week — produced a down week.
What I expected vs what the data showed
The ROI calculator was the second page the automation ever ran on — May 31, day two. Six weeks later, here's where it stands in the 28-day GSC window:
| Metric | Week 1 (90-day average) | Week 8 (28-day) |
|---|---|---|
| Page-level avg position | 58 | 33.1 |
| Query "roi calculator" | not tracked separately | 24.7 |
| Impressions | 1,483 | 2,630 |
| Clicks | part of 45-click total | 1 |
Position moved. Clicks did not.
Finding #1: Page 3 is a number that sounds good and does nothing
Position 24.7 on "roi calculator" is page 3 of Google — the third page of results, technically above position 30. When I saw it in the data I thought: that's close. That's progress I can feel.
Then I looked at the CTR: 1 click from 1,145 impressions. 0.09%.
I curled the live page to check what Google actually renders:
<title>ROI Calculator: Return on Investment & Payback | Valuefy</title>
<meta name="description" content="Calculate ROI, annualized return, and payback period
for any investment. Compare options side by side with consistent metrics.">
The title is specific. The description is benefit-driven. Neither is the problem. The problem is that at position 24, most searchers never see it. "Roi calculator" is a head term with abundant choice in the top 5 — and nobody solving a real problem scrolls to page 3 when page 1 has four adequate answers.
Moving from position 58 to 33.1 (page-level) and 24.7 on the head query is measurable progress. It's also progress toward a cliff. The click rate on head terms doesn't improve gradually as position improves — it's nearly flat from position 60 down to somewhere around position 10, and then it steps up sharply when you cross into the top 5. Everything below that step is a different shade of "not found."
Finding #2: The DCF calculator has real queries and zero clicks
The DCF page was improved July 7 — five days ago. In the 28-day window (which includes the weeks before the improvement), it shows up like this:
| Query | Impressions | Position |
|---|---|---|
| discounted cash flow calculator | 195 | 66.6 |
| dcf calculator | 117 | 57.6 |
| discount cash flow calculator | 10 | 68.3 |
| dcf intrinsic value calculator | 6 | 47.5 |
521 total impressions on the page. One click from a query below GSC's privacy threshold. Every visible query: zero clicks.
"Discounted cash flow calculator" getting 195 impressions in 28 days at position 66.6 is actually useful — it tells me the query is real, the page is indexed, and Google is matching the intent correctly. It also puts the page at position 67, which is page 7 of Google. The content improvement happened five days ago. Whether that changes the position — and by how much — is a 30-day question, not a 5-day one.
The DCF calculator today looks exactly like the ROI calculator six weeks ago. Same pattern: real query volume, positions in the 50s and 60s, no clicks.
Finding #3: The ARR calculator is ranking for German queries I didn't write
The ARR calculator — improved July 5 — has one of the stranger query distributions I've seen in this data:
| Query | Impressions | Position |
|---|---|---|
| arr calculator | 45 | 58.3 |
| arr mrr berechnung | 30 | 89.1 |
| arr multiples | 23 | 87.5 |
| annual recurring revenue calculator | 8 | 70.0 |
"Arr mrr berechnung" is German — it means "ARR MRR calculation." The page is English-language:
<title>ARR Calculator: MRR to ARR, NRR & SaaS Multiples | Valuefy</title>
Google is matching the concept across languages and surfacing the page to German searchers at position 89. Which is page 9 — so nobody clicks it from Germany either. But the cross-language behavior is new to me in this dataset. We haven't done any localization. We haven't added hreflang. Google just decided the ARR concept is relevant to German searchers of "arr mrr berechnung" and started sending those impressions to an English page.
"Arr multiples" at position 87 is a different mismatch. That's a VC valuation lookups query — someone checking current ARR multiples in the market, not calculating their ARR. The page catches it at position 87 because the term overlaps, but the intent doesn't align. Which is probably why it's at position 87.
Finding #4: Six of the seven pages we improved this week aren't in the data yet
Of the seven startup/VC pages automated this week, only DCF shows up in the 28-day GSC top-50 pages. ARR, Churn Rate, Funding, Vesting, Dilution, Cap Table — not there.
This is expected. When a page sits at position 80 or 90, improving its content moves a signal that Google may eventually re-evaluate. It doesn't immediately surface the page in impression rankings. The feedback loop on content improvements runs on weeks, not days.
Meanwhile, the pages delivering what little 7-day traffic there is are the ones I didn't touch: the impression calculator (299 impressions, position 34.8, 1 click) and the P/E ratio calculator (157 impressions, position 24.8, 2 clicks). I wrote about this in week 4. It keeps being true.
What I'm going to do about it
- Track DCF's head query specifically — "discounted cash flow calculator" at position 66.6 is the number to watch over the next 30 days, not the page-level average.
- Rerun the ROI drill-down in a month — if position 24.7 holds, there's something limiting CTR beyond position. If it climbs toward 15, I'll report what happens to the one click per month.
- Stop using page 3 as a success metric — position 24 felt like almost-there when I saw it. The click data says it isn't. Page 3 belongs in the same category as page 6 for conversion purposes.
- Continue the startup cluster — none of the seven pages are showing traffic yet. That's what I expected. This is month-scale work, and the cluster logic (related pages linking to each other) takes time to compound.
The uncomfortable lesson
There's a mental trap in position metrics I keep falling into: the assumption that "better rank" means "closer to getting traffic." The ROI data breaks that assumption cleanly.
Moving from position 58 to 33.1 on the page level — and 24.7 on the head query — took 43 days of automation and one dedicated content pass in late May. That's real work on a real signal. And the return is 1 click from 1,145 impressions.
The click curve on head terms is not a slope. It's a cliff. From position 60 to position 11, you're essentially trading impressions for impressions, not impressions for clicks. The jump happens when you cross into the top 5 or 10 — and we're not there yet on anything.
I still think the work is worth doing. You can't cross the cliff without crossing the territory before it. But I'd be misleading myself if I counted position 24 as meaningful progress toward traffic today. It's meaningful progress toward a position where meaningful progress might start.
I'll check back in 30 days with the DCF and ROI numbers. If something actually moved, I'll show the data. If it didn't, I'll say that instead.
I'm running these experiments on valuefy.app and writing the findings as I go. If you're working through the same impression-without-clicks wall — or building programmatic SEO and wondering what 40+ days of daily automation actually produces — 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 I stress-test ideas before they touch client work.
Top comments (0)