Turkish is head-final: the modifier precedes the head noun. English is head-initial. Agencies and SEO tools built around English routinely generate Turkish keyword targets in English order — seo diş kliniği instead of diş kliniği seo — and the resulting pages target phrases nobody types.
We measured it. 795 keyword variants priced against Google Ads data in a single batch. Every figure below comes from one snapshot taken 2026-09-04, and replicates an independent pull from 2026-09-02 exactly.
The headline result
We built a systematic grid of 25 verticals × 9 services, generating all three Turkish orderings for each pair (225 combinations, 675 keywords), plus 10 cities × 6 services in both orders (120 keywords).
| Form | Example | With any volume | Total monthly volume |
|---|---|---|---|
| Natural order | diş kliniği seo | 31 / 225 | 2,670 |
| Inverted order | seo diş kliniği | 10 / 225 | 100 |
| With "için" (for) | diş kliniği için seo | 2 / 225 | 60 |
The natural order carries 26.7× the total search volume of the inverted order.
The geographic axis shows the same effect independently:
| Form | Example | With any volume | Total monthly volume |
|---|---|---|---|
| City first | izmir seo | 33 / 60 | 6,980 |
| Service first | seo izmir | 20 / 60 | 940 |
7.4× in favour of city-first, on a much larger volume base.
The pairs
Head to head, same two words, order reversed:
| Natural | MSV | Inverted | MSV |
|---|---|---|---|
| sağlık turizmi reklam ajansı | 1,000 | reklam ajansı sağlık turizmi | 0 |
| sağlık turizmi seo | 390 | seo sağlık turizmi | 0 |
| e-ticaret seo | 260 | seo e-ticaret | 10 |
| sağlık turizmi dijital pazarlama | 170 | dijital pazarlama sağlık turizmi | 0 |
| e-ticaret reklam ajansı | 140 | reklam ajansı e-ticaret | 0 |
| turizm pazarlama | 140 | pazarlama turizm | 0 |
| e-ticaret pazarlama | 90 | pazarlama e-ticaret | 0 |
| diş kliniği seo | 70 | seo diş kliniği | 0 |
| otel reklam ajansı | 70 | reklam ajansı otel | 0 |
| e-ticaret web tasarım | 50 | web tasarım e-ticaret | 10 |
| otel web tasarım | 40 | web tasarım otel | 0 |
| emlak pazarlama | 30 | pazarlama emlak | 0 |
Not one pair runs the other way.
We made this mistake ourselves
This study exists because we found the error on our own site. Pricing all 102 sector and location page titles on avangardreklam.com against Google Ads data:
| Status | Titles | Share |
|---|---|---|
| Target has search volume | 28 | 27% |
| Target has zero volume | 74 | 73% |
Total demand across all 102 titles is 11,160 MSV — but 8,100 of that sits on a single page. The other 101 pages share 3,060 MSV between them.
The zero-volume set was systematic, not random: every Meta Reklamları {vertical} page, every Google Ads {vertical} page, and every surgical sub-vertical (rinoplasti, tüp bebek, obezite cerrahisi, göz ameliyatı, dermatoloji, diş implant, estetik cerrahi) returned 0 MSV in every word order.
Some of our own titles used natural order (Diş Kliniği SEO, Obezite Cerrahisi SEO) and others were inverted (SEO Rinoplasti, SEO Dermatoloji, SEO Tüp Bebek) — inconsistent within the same site, which is the signature of template-generated pages rather than researched ones.
What this does and does not show
It shows that for these 795 phrases, Turkish word order is strongly associated with whether a phrase has measurable search volume at all, and that the natural head-final order dominates.
It does not show that reordering a title creates demand. Where a phrase has zero volume in both orders — which is most of the surgical sub-verticals above — the term simply is not searched, and no amount of reordering changes that. Word order matters when demand exists; it cannot manufacture demand.
Sampling. The grid was chosen by us and reflects a marketing-services vocabulary. It is not a random sample of Turkish search. Google Ads volumes are bucketed 12-month averages, so small numbers are approximate and a 0 means "below the reporting floor", not literally zero.
Replication. These figures were produced twice, on 2026-09-02 and again on a fresh pull on 2026-09-04, with identical results (31/225, 10/225, 2/225).
Method
- Build a grid of 25 verticals × 9 services, generating all three Turkish orderings for each pair, plus 10 cities × 6 services in both orders.
- Filter to Google Ads limits (≤80 characters, ≤10 words).
- Price every variant via
keywords_data/google_ads/search_volume/live,location_code: 2792,language_code: tr, in a single batch so all figures come from one snapshot. - Keep Turkish characters intact — never ASCII-fold. saglik and sağlık are different query strings with different volumes.
Total cost of the pull: 0.09 USD.
Practical takeaway
Before publishing a Turkish landing page, price the exact title phrase in both orders. If the inverted form is what you were about to publish, you are likely targeting a phrase with no searches behind it — and if both forms return zero, the page should probably not exist.
This holds across every vertical we measured. In health tourism digital marketing, for instance, the natural order carries 1,000 monthly searches while the inverted order returns exactly zero — with the same two words.
Want the raw data?
We have all 795 keywords, with all three orderings and their volumes. Write to us if you would like us to check your own vertical.
Request the Data