Turkish Hook Bank: What Actually Stops a Scroll in Istanbul

Most foreign brands enter Türkiye with an English ad and a Turkish subtitle track. It performs badly, and the conclusion drawn is that Turkish audiences do not respond to video — when the reality is that they responded to a foreigner's sentence in the two seconds where the decision happens.

The hook is the part of an ad that cannot survive translation, and in Turkish the problem is structural rather than stylistic. English hooks front-load a benefit in three or four words. Turkish builds meaning with suffixes, so the same idea needs a longer unit, and a translated version lands either too late in the clip or too formal in register. By the time the viewer has parsed it, they have scrolled.

So write the hook twice: once as a Turkish writer would open, and once in English, and test which audience you actually have.

Two seconds, muted, in a dense feed

Turkish feeds are as saturated as any in Europe and the constraint is identical: the realistic window before a scroll is under two seconds, the phone is held vertically, and the sound is off.

Three conditions the hook must survive at once:

• Muted. It has to work as readable text. If it only lands as audio, most of the audience never receives it.
• No context. The viewer has not read the caption and does not know the brand.
• One idea. Not a benefit and a price and a deadline. One.

The read-aloud test catches most failures before spend: say the line to a Turkish speaker with no picture and no caption, and ask what is being sold and who it is for. If either takes more than a second, the hook is doing too much work.

The hook bank

Twenty openers that work in a Turkish feed, with the objection each one answers. Treat them as templates, not scripts — swap the product and the specific.

Two things to notice. Every line is short enough to read in a muted feed at arm's length. And the top of the list is shipping, not the product — in a market where kargo bedava is the phrase buyers scan for, a hook that ignores delivery is answering a question nobody asked yet.

The register that reads local

Turkish advertising is warm, direct and problem-led. It assumes the viewer has a specific irritation and names it faster than an English ad would, but it does so conversationally rather than aggressively.

Four calibration notes:

Second person, informal, plural. The natural register for consumer ads is the second-person plural — "bakın", "deneyin" — which is polite without being distant. Singular informal is reserved for youth brands and can read as presumptuous from an unfamiliar seller.

Problem before promise. "Bu fiyata bu kalite gerçekten var mı?" is a better opener than any benefit claim, because it is the question the buyer is already holding. English ad writing tends to name the benefit; Turkish writing often names the doubt.

Superlatives are suspicious, not persuasive. "En iyi" — the best — invites the question of who decided. Numerals are more convincing than adjectives: 4.7 stars, 4,800 reviews, 14 days, 24 saat.

Length discipline. Because Turkish is agglutinative, a caption that fits in English overflows. The fix is a shorter Turkish line, never a smaller font or a faster read. A hook that cannot be said in four words in Turkish should be replaced, not compressed.

The muted-first production rule

In a video ad, the hook exists in three layers at once: the spoken line, the on-screen text, and the first frame. All three have to work without the other two.

• First frame: product visible, in focus, on a surface a buyer recognises. No logo card, no animated opener.
• On-screen text: the hook, in the middle band of the frame, in a font carrying the full Turkish character set including İ, ı, ş, ğ, ç, ö and ü. A font subset that drops them produces broken script and instantly marks the ad as foreign.
• Audio: the same line, in a native voice. If you cannot get a native voice actor, keep it dialogue-free with music and text — a synthetic Turkish voice that is fluent but wrong in register is worse than no voice at all.

One specific software trap: Turkish case mapping. An uppercase "I" lowercases to "ı", not "i", and tools that ignore this mangle hashtags, headlines and captions. Check every automatically uppercased asset by eye before it ships.

The test grid, priced in lira

Ten variants, one product, three days, one variable at a time.

Judge on hook rate first and add-to-cart second. A small lira budget cannot produce a stable return figure, and pretending otherwise produces confident decisions from noisy data. The discipline is the same one set out in six hooks, one product, with the language dimension adding a variable that most foreign brands never test.

Where Turkish creative goes wrong

Five failures that repeat, and all of them are cheap to avoid:

1. Translated hooks. The single most common and most expensive mistake, because it is invisible in a performance report until you compare against a native version.

2. A generic delivery claim. "Fast delivery" without a named carrier reads as a reseller. Aras, Yurtiçi, MNG, PTT, Sürat — name one.

3. No shipping line at all. In a market where kargo bedava is the scanned phrase, omitting it is leaving the easiest conversion on the table.

4. A price format from another country. 1.299,00 TL is Turkish; 1,299.00 TL is a tell.

5. Cinematic production on a listing. Turkish marketplace buyers are comparing on a grid and reading ratings. Polish reads as concealment where the buyer wants to see the product behave.

None of these require a new shoot. They require a Turkish pass over the script, the caption layer and the brief before anything is generated — which is the cheapest step in the entire pipeline and the one most likely to be skipped when a batch is optimised for volume.

The caption layer that carries the ad

If the hook is muted-first, the caption layer is the ad. In a Turkish feed the on-screen text does more work than the voice, and it is also where the cheapest mistakes happen.

Five rules that hold up across categories:

One card, one idea. A card that says "Kargo bedava" and "14 gün iade" and "Bugüne özel" has said three things and communicated none. Split across three cards or cut two.

Six words maximum. Turkish agglutination means a literal translation of a five-word English card becomes a nine-word Turkish one. Write to the six-word limit in Turkish and the card will fit.

Middle band only. Platform chrome eats the top and bottom of the frame. Turkish captions placed low disappear behind the caption bar and the CTA button, which is the single most common layout error when an English template is reused unchanged.

Contrast over styling. White text with a solid shadow, or a dark translucent bar. Turkish feeds are dense and bright, and thin light text disappears in daylight.

Numbers in Turkish format. 1.299,00 TL, not 1,299.00 TL. A comma decimal with a dot thousands separator is the local convention, and getting it wrong signals a foreign seller faster than an accent does.

The caption layer is also where the case-mapping trap lives: any automatic uppercasing must be Turkish-aware, because a generic ASCII pipeline turns "istanbul" into "ISTANBUL" instead of "İSTANBUL" and a dotless "ı" into a dotted "i". Check every generated card by eye before it ships.

What to test this week

Take your best-performing English ad from the last quarter and produce two Turkish versions of the opening line only: one shipping-led, one doubt-led. Same body, same product footage, same end card. Run both at 500 TL a day for three days.

Whichever wins tells you what your Turkish audience is actually optimising for — and it costs less than a single studio hour. The market detail behind that test, including what a Turkish listing video has to do on Trendyol, is in AI video for Turkish e-commerce, and the pricing in lira is in what AI video costs in Türkiye.