Google Was Translating 28% of Our Search Impressions: hreflang for English, Hindi and Hinglish in 2027

“Translated results” is a Search Console search-appearance label, not evidence that FRQ published a translated page. Google uses it when it machine-translates the title and snippet of an existing page into the searcher’s language. The page itself may still be entirely English.

For FRQ, this label identified searches where Google showed an English result with a translated preview. The export’s Search appearance report included a row named “Translated results,” allowing the team to compare its impressions and clicks with other appearances. The translated preview helped the result appear relevant, but it did not solve the underlying language mismatch.

That distinction matters because the label revealed missed opportunities rather than successful localization. Translated results made up 28% of FRQ’s impressions, yet their CTR was only 0.52%, compared with 1.53% for other results. FRQ is therefore shipping real Hinglish pages, and later Devanagari pages, instead of relying on Google’s automatic translation alone.

Translated results represented 28% of FRQ’s total search impressions during 26 September–27 October 2026. That is more than a small edge case: over a quarter of the site’s visibility came from people seeing a translated preview of an English page. The figure shows that language demand was already present in FRQ’s search data.

Their performance was much weaker. The translated-results row recorded a 0.52% click-through rate. All other search appearances combined recorded a 1.53% CTR. FRQ calculated the second figure by subtracting translated-result clicks and impressions from the totals in the Chart sheet, then dividing remaining clicks by remaining impressions.

The gap is the article’s central signal. Translated-result CTR was roughly one-third of the rate for other results, despite appearing at about the same position. FRQ treats this as an opportunity, not a completed success story. It plans to publish real Hinglish guides, then Devanagari versions, and measure the effect for the 2027 season.

Ranking alone did not remove the language barrier. Hindi and Hinglish searchers were typing in Devanagari or Latin-script Hindi, while FRQ offered only English pages. Google could place those pages around positions 6–9, but the result still presented English page content to audiences searching in another form of language.

The Queries sheet illustrates the mismatch. It included searches such as “karwa chauth ki shopping kab kare,” which sat around position 6–9 with zero clicks. The article also reports that almost none of the Hindi and Hinglish queries in the visible sample led to clicks on an English page. A translated title or snippet was not enough to make the page feel like the requested result.

The available query data is incomplete: Google showed 333 queries covering 2,340 of 6,886 impressions, while the rest were anonymized. The script classifier was also crude. Even so, its pattern was clear enough for FRQ to prioritize language-specific pages rather than assuming strong rankings meant strong engagement.

English is written using English vocabulary and the Latin alphabet. Hindi written in Devanagari uses Hindi vocabulary in the Devanagari script. Hinglish is Hindi expressed with Latin letters, often mixed with English words, as in “shopping kab kare.” These are distinct ways FRQ’s audience searches and reads content.

The article’s script-splitting code looks first for Devanagari characters. If it finds none, it checks a small list of Hindi function and shopping words typed in Latin letters, including “kab,” “kaise,” “kare,” “kapde,” and “dukan.” Queries that match neither category are labeled “english_or_brand.”

This classification is useful but imperfect. The word list can miss Hinglish and can occasionally catch an unusual English query. Still, the grouped data showed a real share of Hindi and Hinglish searches, with almost none clicking an English page. FRQ therefore plans real Hinglish versions first and Devanagari versions later, rather than treating every query as English.

Separate language URLs let FRQ publish complete versions of one guide for different audiences. Each version can contain the appropriate language and script, instead of leaving Google with only an English page whose title and snippet it must translate. Hreflang then tells Google that these URLs are alternate versions of the same content.

For example, FRQ’s English guide uses one URL, while its Hinglish version uses another. Each page lists the English, Hindi-language, and x-default alternatives with link tags. The same relationship can also be recorded in the sitemap with xhtml alternate links. Google can use those signals when selecting a version for a searcher.

The setup only works when the alternate set is consistent. Every page must list every version, including itself, and alternates must point back to one another. Each URL must also use a self-referencing canonical. FRQ is shipping Hinglish first and later Devanagari pages, but the article presents this as a measured plan, not a proven result.

Return links and self-referencing canonicals prevent conflicting signals. If page A names page B as an alternate but page B does not name page A, Google may ignore the pair. The relationship must be reciprocal, so every page in the set lists every other version, including itself.

Canonical tags serve a different purpose. A page’s canonical should point to its own URL. If the Hinglish page canonicals to the English page, it effectively tells Google to treat the Hinglish page as something to ignore. That undermines the purpose of publishing a separate, useful language version.

FRQ’s example therefore puts both signals on both pages: each has a self-canonical and alternate links for the English and Hinglish URLs, with x-default pointing to English. The sitemap can carry the same relationships. This consistency matters on a growing blog, where missing return links or copied canonical tags could prevent Google from recognizing the intended alternates.

FRQ should treat the hreflang rollout as a measured plan, not assume it worked. It can establish the existing 26 September–27 October 2026 period as a baseline, then compare later Search Console data after the new pages launch. The comparison should cover the English, Hinglish, and later Devanagari URLs separately.

The main checks are clicks, impressions, CTR, average position, and the share labeled “Translated results.” FRQ should also group visible queries by Devanagari, Hinglish, and English or brand using its script method, while remembering that anonymized queries are missing and the hand-made Hinglish list is imperfect. Queries such as “karwa chauth ki shopping kab kare” provide a concrete watchlist.

Improvement would mean more clicks and higher CTR from Hindi and Hinglish searches, alongside less reliance on translated results. FRQ should compare equivalent periods where possible, because the current data covers a defined window and the plan targets the 2027 season. The article does not report post-launch results yet.