FILE 10 — OPEN LOG
We started link building on three sites in the same week. This page is a log, not a result.
Most link-building case studies are written after the links that worked have been picked out. This one was written before: the starting profiles, the strategy, the filtering rules and the numbers from week one. In December we add the results to this same page — including if they are bad.
- Scope
- avangardreklam.com, gofluentacademy.com, spaceagency.online
- Markets
- Türkiye, Argentina, United States
- Started
- 20 – 23 September 2026
- Status
- In progress — results December 2026
Week one, three sites combined
A pipeline, not a result
- 179
- Individual pitches
- three sites combined
- 1,449
- Prospects screened
- Go Fluent + Space Agency
- 16
- Linkable assets
- live before the first send
- 0
- Links bought
- 5 payment requests declined
None of these figures is an earned link. Sending began 20–23 September; a publication adding a link takes weeks.
Source: each site’s outreach logs, 28 September 2026
01 — Why now
All three sites had run out of on-page SEO
This programme came from a ceiling, not an opportunity. All three sites had largely finished what technical and content work could do; the one missing input was the external signal that makes Google trust the pages.
We measured the ceiling most clearly on spaceagency.online: four SEO waves multiplied impressions 5.6× while clicks stayed at four. The diagnosis was the referring-domain count. The same pattern held on our own site and on Go Fluent — less dramatic, same cause.
So we started three programmes in the same week, with the same method. The goal is not only to collect links; it is to see whether the method holds across different languages and markets.
02 — Where we started
Three profiles, three different problems
A referring-domain count on its own misleads. On all three sites the number looked “not bad”; looking inside showed the real starting point was close to zero.
On our own site, 12 of 31 referring domains were injected PBN links we had nothing to do with. The September move from 27 to 31 is therefore not earned, and we don’t count it. In June there were only 4 dofollow domains, and editorial Turkish media links still stand at 0.
On Go Fluent, 9 of 18 domains were nofollow and the profile carried a significant share of low-quality links — so the work started with cleanup, not growth. On Space Agency, 102 of 120 referring pages were nofollow and 100 were footer links from a single platform.
| Site | Referring domains | What was wrong with it | Measured |
|---|---|---|---|
| avangardreklam.com | 31 | 12 injected PBN links; 0 editorial .tr links | 20 September 2026 |
| gofluentacademy.com | 18 | Half nofollow; significant low-quality links | 14 September 2026 |
| spaceagency.online | 20 | 100 of 120 pages are one platform’s footer | 31 August 2026 |
Source: DataForSEO and SE Ranking; measurement dates in each row
Rule: we measure the number of domains Google would trust, not the total. The December update to this page is scored on that number.
03 — Strategy
One method, three markets
The same four steps ran on all three sites. What changes is not the method but who we approach in each market, and with what.
- 01
Competitor link gap
Domains that link to competitors but not to us, pulled from DataForSEO, Semrush and SE Ranking data. This is the spine of every prospect list.
- 02
A linkable asset
For each market, an original page worth citing: a data study, an index, a tool or a glossary. No email went out before the asset existed.
- 03
Market adaptation
In Türkiye, trade media and health tourism; in Argentina, language learning, education resources and local institutions; in the US, small-business resource pages and library guides.
- 04
One follow-up, then stop
At most one reminder per prospect. A third email to an editor who hasn’t replied does not earn a link; it spends the domain’s reputation.
04 — Linkable assets
16 pages first, email second
The weakest form of outreach is “would you link to our site”. Instead, on each site we built something an editor would want to show their own readers.
avangardreklam.com — Türkiye
- Turkish word-order search-volume study
795 keyword variants; natural order carries 26.7× the volume of the inverted order.
- Ad cost index
Cost-per-click by sector — a single page of data a journalist can quote.
- Hair transplant & dental price index
Sourced clinic prices, built for health-tourism publishers.
- International patient review study
A systematic analysis of clinic reviews.
gofluentacademy.com — Argentina
- Adaptive Spanish placement test (with shareable badges)
- Learning-time calculator (embeddable widget)
- Lunfardo / Argentinian slang dictionary — 114 terms
- English irregular verbs tool
- Course price index
- Wine glossary
spaceagency.online — United States
- Web design pricing report
- Website cost calculator (embeddable)
- AI search visibility checker
- ChatGPT vs Google local results (research)
- AI crawler access index (research)
- Free templates: SEO proposal and website RFP
None of the assets is a product page. Publications don’t link to product pages; they link to data their readers can use. The commercial pages take their authority from these assets through internal links.
05 — Filtering
On Go Fluent we rejected 86% of prospects
For Go Fluent, 1,254 prospects were reviewed and 1,074 were rejected. That ratio is not inefficiency; it is exactly where the quality of the programme is decided.
- 01
Topical fit
If the site doesn’t write about our asset’s subject, it’s out — whatever its authority.
- 02
Real traffic
“High DR” sites with no rankings and no organic traffic are out. Most of them exist to sell links.
- 03
Paid-post signals
Sites advertising “sponsored”, guest-post prices or link packages are out.
- 04
A reachable person
No named editor or page owner, no send. We don’t bulk-email info@ addresses.
06 — Week one
The pipeline, as of 28 September
179 individual emails have gone out across the three sites. Not enough time has passed to count an earned link; in December this table is rewritten with placements.
| Site | Market | Prospects screened | Sent | Replies | First send |
|---|---|---|---|---|---|
| avangardreklam.com | Türkiye · Turkish | — | 51 | — | 20 September 2026 |
| gofluentacademy.com | Argentina · Spanish / English | 1,254 | 33 | 0 | 21 September 2026 |
| spaceagency.online | United States · English | 195 | 95 | 7 | 23 September 2026 |
Source: each site’s outreach logs, 28 September 2026
On the Space Agency side, 4 of 95 emails bounced and 7 replied (7.7%). 2 were engaged; 5 asked for payment. Of our own directory submissions, 3 are live — 2 of them nofollow, so they carry brand visibility rather than authority.
07 — Limits
What we don’t do
A link-building programme is as credible as what it refuses. These rules are not negotiable.
- 01
We don’t buy links
All 5 replies asking for payment were declined. We found plenty of paid news placements on competitors’ profiles and kept them off our target lists.
- 02
We don’t run a link network
Our own sites link to each other only where it describes a real relationship, and mostly nofollow. Turning three sites into a link network would put all three at risk.
- 03
We don’t publish target lists
Which publications we pitched, who asked for money, and the email copy stay private. They are our relationship with editors and our working capital.
Paid links would improve this table faster in the short run. But if we break on our own site the method we recommend to clients, this page is worth nothing.
08 — December
How this page will be scored
The results update is due 21 December 2026. We are writing the success criteria now, before seeing the results.
- 01
Earned domains
Not the total: only dofollow, topically relevant domains won by the programme. Injected or paid links excluded.
- 02
The Space Agency threshold
Target: from 20 to the 45–60 band. If it is still under 30 in December, the problem is distribution, not content — and the strategy changes.
- 03
Links per asset
Which assets earned links and which didn’t. An asset type that earns nothing is not built again next round.
- 04
Search impact
Search Console position change for the pages that earned links and for the commercial pages they feed through internal links.
Free link gap snapshot
We’ll show you the 20 sites linking to your competitors but not to you.
Send your domain and two competitors through the contact form. We run the first step of this programme — the competitor link gap — on your site and email you the ranked list. No sales call.
Send my domainDecember results
Get the results when they land.
When we update this page with the 90-day measurement — what worked and what didn’t — we send one email.